Humidtropics: Integrated Systems for the Humid Tropics

Transcription

1 Humidtropics: Integrated Systems for the Humid Tropics 24 January The Consultative Group on International Agricultural Research (CGIAR) is undergoing a major change process, which sees its Center operations implemented through a number of CGIAR Research Programs (CRPs). These are aimed to better coordinate Research for Development (R4D) efforts, enhance efficiencies, and encourage cooperation and collaboration with a focus on effective partnerships to achieve more development-oriented impacts. CRP1.2 Humidtropics stands for R4D partnerships that impact people s lives in the humid and sub-humid tropics while enhancing the ecological integrity of natural resources.

2 CONTENTS 1 Executive Abtract Vision and Theory of Change Vision Theory of Change Objectives and Program Design Justification Why the Humid and Sub- humid Tropics? Why Focus on Production System Intensification? Rural Poverty Alleviation in the Context of System Intensification Scaling of Intensification Processes Sensitivity to Different Baseline Conditions Strategic Research Program Humidtropics Global and Working Hypotheses Strategic Research Themes Impact Zones, Action Areas, and Action Sites SRT1 Systems Analysis and Synthesis SRT2 Integrated Systems Improvement SRT2.1 Integrating Markets through System Intensification SRT2.2 Increasing System Productivity SRT2.3 Natural Resources Improvement SRT3: Scaling and Institutional Innovation Quantified Impact Pathway Humidtropics Impact Approach Stepwise Impact Pathways Quantifying Production System Impacts Quantifying Natural Resource Integrity Quantifying Market Impacts Assessing Scaling- up Options and Development Impacts Partnership Management and Capacity Development Partnerships Organizing and Mobilizing Partnerships Partnership Management Communication Capacity Development Gender Research Strategy Rationale Mainstreaming and Integrating Gender Strategic Gender Research Institutional Support and Collaboration Innovation Interactions with other CRPs Why Interaction with Other CRPs is Needed Realizing Effective Collaboration Between CRPs Management arrangements for implementation Overview Governance Advisory Committee Primary Partners

3 10.5 Executive Director Management Committee Action Areas and Sites Teams Timeframe Risks Partners Limited Delivery Competences Narrow Integration of Humidtropics in Wider Development Contexts Limited Communication Limited Links Among Research and Development Partners Business Planning Monitoring and Evaluation M&E Framework M&E Process M&E Unit Budget Bibliography Appendix 1: Budget Requests for 2011/12/

4 ACRONYMS AND ABBREVIATIONS Below is a limited list of acronyms and abbreviations to guide with reading this document. A comprehensive updated list with most commonly used acronyms and abbreviations in the CGIAR system is available at AGRA Alliance for the a Green Revolution in Africa AIS Agricultural Innovation System APAARI Asia Pacific Association of Agricultural Research Institutes ARI Advanced research institute AU African Union AVRDC World Vegetable Center CATIE Centro Agronómico Tropical de Investigación y Enseñanza CB CGIAR Consortium Board CEO Chief Executive Officer CGIAR Consultative Group on International Agricultural Research CIALCA Consortium for Improving Agriculture-based Livelihoods in Central Africa CIAT Centro Internacional de Agricultura Tropical CIMMYT Centro Internacional de Mejoramiento de Maíz y Trigo CIFOR Center for International Forestry Research CIP Centro Internacional de la Papa CIRAD Centre de coopération internationale en recherche agronomique pour le développement CONDESAN Consorcio para el Desarrollo Sostenible de la Ecoregión Andina CORAF/WECARD Conseil Ouest et Centre Africain pour la recherche et le développement agricoles CRP CGIAR Research Program CRP1 CGIAR Consortium Theme 1 on integrated agriculture systems for the poor and vulnerable CRP1.1 CGIAR Consortium Research Program on integrated agricultural production systems for the poor and vulnerable in dry areas CRP1.2 CGIAR Consortium Research Program on integrated systems for the humid tropics now called Humidtropics CRP1.3 CGIAR Consortium Research Program on harnessing the development potential of aquatic agricultural systems for the poor and vulnerable CRP2 CGIAR Consortium Research Program on policies, institutions, and markets to strengthen assets and agricultural incomes for the poor CRP3 CGIAR Consortium Theme 3 on sustainable production systems for ensuring food security (with one program each on dryland cereals, grain legumes, livestock and fish, maize, rice, wheat, and roots, tubers and banana) CRP4 CGIAR Consortium Research Program on agriculture for improved nutrition and health CRP5 CGIAR Consortium Research Program on water, land and ecosystems: improved natural resources management for food security and livelihoods CRP6 CGIAR Consortium Research Program on forests and trees: livelihoods, landscapes and governance CRP7 CGIAR Consortium Research Program on climate change, agriculture and food security CSO Civil Society Organization DG Director General DDG Deputy Director General ECA East and Central Africa EIARD European initiative on international agricultural research for development EMBRAPA Empresa Brasileira de Pesquisa Agropecuária FAO Food and Agriculture Organization of the United Nations FARA Forum for Agricultural Research in Africa

5 FDI FORAGRO GCARD GIS GMS Humidtropics ICARDA ICIPE ICRAF ICRISAT ICT IICA IITA ILRI IPG IRRI IWMI M&E MDG NARS NGO NGS NRM PMEIA PNG R&D R4D RIU SLO SRF SRT SRT1 SRT2 SRT2.1 SRT2.2 SRT2.3 SRT3 WAGR WUR Foreign Direct Investment Foro de las Américas para la Investigación y Desarrollo Tecnológico Agropecuario Global Conferences on Agricultural Research for Development Geographic Information Systems Greater Mekong Sub-region CGIAR Consortium Research Program on Integrated Systems for the Humid Tropics (CRP1.2) International Center for Agricultural Research in the Dry Areas International Centre of Insect Physiology and Ecology World Agroforestry Centre International Crops Research Institute for the Semi-arid Tropics Information and Communication Technology Inter American Institute for Collaboration in Agriculture International Institute of Tropical Agriculture International Livestock Research Institute International Public Good International Rice Research Institute International Water Management Institute Monitoring and Evaluation Millennium Development Goal National Agricultural Research Systems Non-governmental Organization Northern Guinea Savanna Natural Resources Management Participatory Monitoring, Evaluation and Impact Assessment Approach Papua New Guinea Research and Development Research-for-development Research Into Use (DfID) System Level Outcome CGIAR Strategy and Results Framework Strategic Research Theme Systems analysis and synthesis Integrated systems improvement Integrating markets through system intensification Increasing system productivity Natural resources improvement Scaling and institutional innovation West African Guinea Rainforest Wageningen University and Research Center

6 1 EXECUTIVE ABTRACT This proposal describes CRP1.2: The CGIAR Consortium Research Program on Integrated Systems for the Humid Tropics, referred to as Humidtropics. This program seeks to transform the lives of rural poor in the humid lowlands, moist savannas and tropical highlands in three major Impact Zones of Sub-Saharan Africa and tropical America and Asia, presently containing a population of 2.9 billion persons, mostly poor smallholder farmers. Humidtropics research is guided by a Global Hypothesis A stepwise series of preferred livelihood strategies exist within the humid tropics where poverty reduction, balanced household nutrition, system productivity and natural resource integrity are most effectively achieved and contribute best to human welfare and several related Specific Hypotheses and Research Questions addressing integrated production system interventions, expanded livelihood options, and the abilities of institutions to take these interventions to scale and target rural poverty and gender equity outcomes. A dynamic program structure is built around three complementary Strategic Research Themes; Systems Analysis and Synthesis, Integrated Systems Improvement, and Scaling and Institutional Innovations. Together these SRTs will conduct a baseline Situation Analysis leading to identified entry points for integrated production systems research; design and implement an M&E Framework; assemble, test and refine systems interventions through participatory processes; champion new farm opportunities through Research for Development Platforms as pathways to assess fuller impacts and adoptability of the most promising opportunities; link these platforms to partner development institutions; and then advance the effectiveness of these institutions to scale up these interventions, with a particular focus on poor households and gender equity. These activities will be conducted within 11 Action Areas that were selected on the basis of being representative and capturing diversity of their larger Impact Zones, the urgent need for large-scale impacts on poverty and natural resource integrity and their potential to advance earlier investments and potential partnerships. Some of these Action Areas are arranged as transects from humid lowlands to highlands while others are nested into large concentrations of poverty and resource degradation. Together these Action Areas contain 11% of the population on 9% of the lands in the program s Impact Zones, leaving great potential for direct impacts through targeted interventions addressing rural poverty, food security, improving nutrition and health, and sustainable management of natural resources. Each Action Area contains nine to 24 Action Sites where integrative research on systems productivity, natural resource management and market development are conducted. These Action Sites will serve as proving grounds for integrating tree, crop and livestock enterprises, advancing the status of women and linking them to technical innovations, fostering and expanding agro-biodiversity, developing new products for rural industries, and linking farm management to better managing community natural resources and mitigating climate change. Humidtropics will be implemented through a governance and management framework balancing inclusiveness and responsibility. Its success is based upon partnership including raising the effectiveness of local and national partners and providing opportunities to test outputs developed by other CRPs. In many ways, the Action Areas operate as independent units but then combine their lessons learned, research outputs and developmental outcomes into a larger Global Synthesis focused upon the production and dissemination of International Public Goods. In its first three years, Humidtropics will reorganize diverse and narrowly focused research efforts in these Action Sites and transform them into a larger, impact-oriented thrust in their respective Action Areas in accordance with planned cross-center collaborative reform within the CGIAR. Over the next 15 years, Humidtropics will advance System Level Outcomes within the eleven Action Areas by increasing staple food yields by 60%, increasing average farm income by 50%, lifting 25% of poor households above the poverty line, reducing the number of malnourished children by 30% and restoring 40% of these farms to sustainable resource management. In this way, Humidtropics will serve as a model to other agencies seeking to link agricultural systems research to developmental impact. 1

7 2 VISION AND THEORY OF CHANGE 2.1 Vision Agriculture in the humid and sub-humid tropics is fundamental to rural livelihoods and offers the greatest potential to increase world food supplies. At the same time it risks significant loss of global biodiversity and negative impacts on natural resources. To address these interconnected socioeconomic opportunities and environmental threats, an innovative Research for Development (R4D) program known as the Humidtropics Program will work in diverse, selected agro-ecologies, market conditions, and farming systems to develop innovations integrating increased productivity, improved market performance, and sustainable management of the natural resource base. This will lead to intensification of livelihood systems based on integrated agricultural production in rain-fed farming areas while, at the same time, decreasing rural poverty, improving gender equity, restoring degrading farmlands and reducing pressure on natural ecosystems. The vision of the CGIAR is to reduce poverty and hunger, improve human health and nutrition, and enhance ecosystem resilience through high-quality international agricultural research, partnership and leadership. The Humidtropics supports this vision, which in turn contributes to the following CGIAR System Level Outcomes (SLOs): 1. Reducing rural poverty. Agricultural growth through improved productivity, market development and income generation has shown to be a particularly effective contributor to reducing poverty, especially in the initial stages of economic development. 2. Increasing food security. Access to affordable food is a problem for millions of poor in urban and rural communities and requires increasing global supply of key staples and reducing potential price increases and price volatility. 3. Improving nutrition and health. Poor populations spend most of their income on food and suffer from diets that are insufficient in proteins, vitamins and minerals affecting health and development, particularly among women and children. 4. Sustainable management of natural resources. Agriculture has a substantial impact on natural resources that must be better managed to supply sustainable ecosystem services, particularly in light of climate change. 2.2 Theory of Change Change in rainfed, smallholder farming systems in the tropics is gradual, adaptive, and stepwise; responding primarily to changes in market conditions, farmer-available resources, and increasingly, to changes of climate. Humidtropics seeks first to improve understanding of these processes in terms of alternative intensification pathways and critical points of intervention and then to design superior combinations of crops, livestock, fallows and trees; soil and water management practices; investment strategies for sustainable management of the natural resource base; and market innovations that direct intensification toward desired outcomes. These interventions will increase overall farm and system productivity and income while improving the integrity of the natural resource base, particularly soil and water quality. Humidtropics seeks to move beyond a mere focus on current principal staples but rather it strategically selects critical entry points that foster cumulative integration of new and more diverse system components. Humidtropics thus will attempt to generate a more equitable and integrative agricultural growth process, one that moves beyond commodity booms, diversifies risks, sustainably manages the natural resource base, and effectively reduces rural poverty. While Humidtropics delivers an innovative research program focused on integrated farming systems in the humid and sub-humid tropics, equally important is the design of 2

8 programmatic and institutional innovations that takes this research to scale. This research integrates market-level innovation with production system intervention. Furthermore, scaling up by itself, generates research questions that are critical in addressing how production system intensification impacts upon the CGIAR s SLOs. Research on scaling up will be located across eleven strategically selected Action Areas, within which are embedded representative Action Sites where research on markets, productivity, and natural resource management (NRM) is conducted. While going to scale with production system intensification is a direct pathway to achieving the food security and sustainable natural resource outcomes, other methodological and institutional innovations are required to ensure impacts on rural poverty and improved nutrition of the most vulnerable. These may involve increasing access of poor households to technology and markets through farmer associations, ensuring participation of poor households in farmer groups organized around extension, microcredit or marketing, using women groups to introduce marketing of women s food crops and value added processing, or ensuring poorer households participation in innovation platforms or value chain development. Strategies for scaling up programs to achieve these SLOs will build on past experiences and be directed by market opportunities and resource availability within the different Actions Sites. In this way, greater understanding on how to better design and adjust future rural development programs is generated. The overall program will be monitored and evaluated in relation to achieving targets related to CGIAR System Level Outcomes in the program s Action Areas. While the details on these targets is addressed later in this document, overall targets of the Humidtropics Program include increasing crop yields by 60% by optimizing input costs, crop and livestock combinations and land management practices; reducing poverty for 25 million people below the poverty line by increasing their incomes by at least 15%; and restoring productivity to 30 million ha of degrading cropland. The time frame for these achievements is estimated to be 15 years. What is presented in this document is a very new research agenda for the CGIAR, and the first three years of project implementation will focus on establishing the Action Areas, building the research teams, and testing basic protocols and data collection methods. Over these three years, a diverse array of existing, weakly related agricultural research thrusts, transition into a comprehensive, coordinated Research for Development (R4D) portfolio answerable to a wider array of partners and stakeholders. The following 15-year targets apply to the farming population in the Action Areas: a. Increase staple food yields by 60% b. Increase average farm income by 50% c. Lift 25% of poor households above the poverty line d. Reduce the number of malnourished children by 30% e. Nutrient depletion on 40% of farms reversed to sustainable nutrient flow 2.3 Objectives and Program Design An integrative approach to research is necessary to sustainably increase productivity, reduce poverty and ensure ecosystem services in the challenging ecosystems of the humid and sub-humid tropics. Research on farming systems rather than technological components is essential to meeting both productivity and resource management goals and farmer investment choices are enhanced by integrally linking farming systems to markets. These dimensions determine the following overall goal for Humidtropics: To strengthen research and stimulate institutional innovation that increases economic and social returns among rural households adopting enhanced and sustained agricultural production and marketing strategies, while improving the biological and ecological integrity of their natural resource base. 3

9 Meeting this goal is based on a program design organized around the following four specific objectives: 1. To develop an integrated research program on sustainable intensification of rainfed smallholder farming systems in the humid and sub-humid tropics. 2. To undertake research on the scaling up of this research to achieve impact on the four critical CGIAR System Level Objectives. 3. To develop a network of Action Areas systematically comprising the variability of ecologies, agricultural potentials, levels of market development, population pressures and exposure to climate change that maximizes learning and raises potential for scalingup and out. 4. To establish a platform for global change research in response to agricultural challenges and associated impacts on tropical ecosystems and resources. The program will be organized into three Strategic Research Themes (SRTs). SRT1 will involve a more centralized capacity to undertake characterization of Action Areas and their component Action Sites, to coordinate the development of surveys to monitor agricultural system change, and to provide analytical support in research synthesis at Action Area and program levels. SRT2 is an integrated research program organized around market development, system productivity, and natural resource management. SRT3 conducts research on scaling up and impacts on rural poverty and gender equity in a manner that raises institutional effectiveness of program partners. The research framework is a matrix of location-based research in the Action Areas drawing on critical capacities in thematic research groups arrayed across Centers and partner organizations. The articulation of research outputs between the three SRTs is essential to achieving the different research outcomes and in turn in translating these into impacts on system level objectives. The rest of the document essentially provides the detail on how these different elements will be developed and then linked to each other. 4

10 3 JUSTIFICATION 3.1 Why the Humid and Sub-humid Tropics? Agricultural systems span the humid and sub-humid tropics from the highly productive and densely populated irrigated rice systems of Asia, through the sparsely populated slash-andburn systems of the Congo Basin, and to the highly vulnerable areas of the humid Andes in South America. The humid and sub-humid tropics are at the same time critical to global supplies of basic foodstuffs, central to the maintenance of global biodiversity and the mitigation of greenhouse gases, and yet offer the largest potential for meeting world food demand over the next several decades. In addition, the bulk of the rural poor reside in the humid and sub-humid tropics, primarily in Sub-Saharan Africa and South Asia, areas also associated with poor household nutrition and soil fertility depletion. Indeed, sustainable intensification of farming systems in the humid and sub-humid tropics offers in many respects the greatest potential contribution to all four of the CGIAR s System Level Outcomes yet this opportunity is matched by the challenges to working in such a large and diverse suite of agro-ecologies where the pre-conditions for sustainable intensification are so varied and in many circumstances do not provide appropriate incentives. Agricultural growth trajectories in the humid and sub-humid tropics can be broadly characterized as either expansion of extensively cropped land or land use intensification, although these exist as poles along a continuum of farmer investment in sustainable land management. This continuum usually covers a transition from conversion of natural forest or savanna, a period of rapid degradation of the land resource base and then investment and improvement in land productivity (Scherr and Hazell, 1994; Lambin et al., 2001). The degree and scope of degradation and the point at which the land resource starts to increase in value causing investment are broadly determined by population density and market conditions. In areas farthest from markets, the cost of transportation reduces this value and restricts availability of farm inputs. At the other end of the intensification spectrum are the intensively managed irrigated agricultural regions of the Asian river deltas, where low costs of inputs, inefficient water use, and poor understanding of environmental externalities results in soil and water degradation and excessive reliance upon external inputs, particularly fertilizer and pesticides. Between these two extremes lie the bulk of agricultural lands in the humid and sub-humid tropics, where production systems are essentially rainfed, dominated by smallholders, and intensification faces a number of unique constraints. This is where CRP1.2 will focus, as the extensive margin in the humid forest is central to CRP6 and the more intensive margin is the focus of CRP1.3 and GRISP (Global Rice Science Partnership). Agricultural development in the humid and sub-humid tropics has been uneven, and is best reflected in what the World Development Report (World Bank, 2007) has called the three worlds of agriculture. In the first agricultural world, primarily in Sub-Saharan Africa, economies are dependent on agriculture but growth remains sluggish, and poverty predominates in rural areas. Generating agricultural growth is crucial for poverty reduction, and this must take place both where land availability is highly constrained, such as Kenya, Rwanda, Liberia and Malawi, and where land is relatively abundant, such as Mozambique, Zambia, DR Congo, Tanzania, Nigeria and Ghana. The incentives for production system intensification vary significantly across this spectrum, and yet the need to increase productivity and ensure sustainable management of the natural resource base is a constant. The transforming economies of the second agricultural world are primarily found in tropical Asia. These countries are characterized by rapid economic growth, where agriculture is a declining contributor to the overall GDP (although livestock as a proportion of agricultural GDP increases), but where the number of poor is still very large and concentrated in lagging 5

11 areas that have not participated in the agricultural growth process. These areas include Laos and Cambodia, eastern India and the mountainous highlands of Vietnam and Thailand. These lagging areas often have constrained access to markets, are dependent on rainfed farming systems, have trade-offs in the sustainable management of the natural resource base, and yet rely on agriculture as the means to escape poverty. Urbanized economies are primarily found in Latin America and in these third agricultural world countries agriculture is highly commercialized, poverty is concentrated in urban areas, and sustainable management of natural resources drives policy formulation in relation to smallholder agriculture. This region has seen the largest increase in new land brought into agricultural production, primarily in the savanna ecosystems in the cerrados of Brazil and the llanos of Colombia and Venezuela where large-scale agriculture dominates. Apart from the drier areas of the Northeast of Brazil, most smallholder agriculture in Tropical America is concentrated in the highland systems of the Andes and Central America. It is the highland ecosystem in Latin America that is most congruent with the CGIAR s SLOs. Humidtropics will cover all three agricultural worlds spanning humid forests, tropical highlands, and moist savannas. All three are important for Sub-Saharan Africa, while in Asia the focus will be on the highland regions and the highly populated, humid forest regions of Indonesia, while in the Tropical Americas highland systems predominate. Choice of these ecosystems across the three continents integrates the growth dynamics of the economy with CGIAR System Level Objectives. At the same time, Humidtropics will focus on production system intensification in these agro-ecosystems outside of both the immediate forest margins and the intensive, irrigated areas of Asia. Also Humidtropics will focus on livelihoods based upon integrated agricultural production rather than those dominated by a single commodity covered within other CRPs. The program will therefore, identify and focus on farming systems where the integration of components is more important than the production of any one. This focuses the work of Humidtropics directly on rainfed smallholder farming systems and their opportunities for sustainable intensification. To fully justify the focus of the Humidtropics CRP, however, we must address why the focus on integrated production systems, what is the best approach to alleviating rural poverty in the selected agro-ecosystems, and how to scale complex technologies in order to achieve desired SLOs. 3.2 Why Focus on Production System Intensification? Sustainable intensification of production systems in the humid and sub-humid tropics offers the potential to address simultaneously a number of pressing development objectives, particularly unlocking the agricultural potential of large parts of Sub-Saharan Africa, adapting production systems to climate change, sustainably managing land and water resources in upland regions of Southeast Asia, Eastern Africa, and Tropical Americas, and reducing rural poverty in regions that have been most intractable to lifting the poor above the poverty line. The approach that is developed in this program extends the work of the CGIAR in several different ways. Firstly, it locates technical change within a dynamic framework of production system evolution, where farming systems are conditioned by relative prices of land and labor and by the access and distance to market. Secondly, it pursues intensification beyond a singular focus on increasing staple food productivity; although important, there may be more profitable productivity increases derived from other cash crops or livestock enterprises. Thirdly, it pursues intensification through multiple means, not just increases in yields of single commodities but also by exploiting complementarities between system components, particularly tree-crop-livestock interactions. Fourthly, our increasing understanding of agro-ecological process moves beyond mere increase in off-farm inputs to the combined response in productivity resulting from more efficient harnessing of ecological services and farmer-available resources. Finally, it provides new areas of partnership by integrating investment in the natural 6

12 resource base with the returns from improved productivity and marketing through a phased process of system intensification. Humidtropics contextualizes this multi-prong research within a stage theory of system change and intensification. In the initial phase at very low population densities and high cost of participation in markets (what was termed above the extensive margin), the conditions for farmer investment in any type of intensification process are limited to population increase and settlement by colonists. Humidtropics has chosen not to work in such contexts. This critical point is reached in the humid forest zone at population densities of 15 persons per square kilometer where long-term shifting cultivation no longer suffices and at 30 persons per square kilometer in the moist savannas, where communal grazing systems collapse as population rises, improving the viability of combined tree-crop-livestock enterprises. Above these population densities, particularly where markets are not well developed, land managers continued to focus on subsistence production primarily by exploiting natural soil capital, causing food security to be sought at the expense of natural resource degradation. This need not be the case, however, when integrated approaches toward production, marketing and resource management systems are offered to farming households through reliable agricultural services, policy and infrastructure. There is no longer a single entry point for farm intervention, but rather a suite of complementary livelihood options that empower rural households to reject subsistence strategies while gaining proficiency in crop and livestock production. Increasing market integration in turn leads to a process of crop diversification and commercialization of marketable surpluses, where farming becomes a business. The farming system may nonetheless continue to devote some resources to the production of preferred staples, as the farm gate difference in selling and buying price is still often large. Improved road networks in the highlands of Southeast Asia and a little over a decade of market liberalization in Sub-Saharan Africa are leading to smallholder diversification into higher value activities such as horticulture, dairy, coffee, tea, cocoa, and soybeans. This process is driven by growing urban demand and by the recent trend of global food price increases. Diversification offers the potential to explore complementarities in tree, crop and livestock enterprises that further increase overall system productivity. In this phase, the rising value of land warrants further investment in the natural resource base. Where markets are highly developed and farmers can meet household needs through purchases, farmers objectives shift primarily to income generation. Often the tendency is to move to specialization within the production system based on Sub-regional comparative advantage within the country. Few smallholders are at this stage of farm development apart from irrigated and peri-urban areas. However, pressures toward such intensification through specialization are evident in the highlands of Northern Thailand, central Kenya, and parts of the Andes. In such market contexts, the question arises how best can intensification through diversification of production offer equally sustainable and economically viable pathways for development given farmers expectations of improved lives? This interplay of market conditions, farmer choice of production activities, and farmer investment in the natural resource base allows Humidtropics to explore intensification pathways across gradients of market, population density, and agro-ecological conditions. Humidtropics is a new research direction for the CGIAR and at the same time it builds on a number of system-wide programs and a range of potentially integrated sub-systems research. Much of the impetus for more integrated approaches resulted from a series of meetings organized around integrated natural resource management (INRM). Consensus grew that attempted to develop a framework for better integration of productivity research of the CGIAR with the expanding NRM work (Science Council, 2003). This integration of productivity and NRM at farm scale was not made operational within the CGIAR, as much of the NRM research moved to broader scales such as river basins or landscapes, as reflected in the Challenge Program for Water and Food and the Alternatives to Slash and Burn, and 7

13 East African Highlands system-wide programs. The System-wide Livestock Program also evolved from focused work on livestock feed to broader systems level research, for example on biomass trade-offs in smallholder systems. In response, INRM research tended to focus on sub-systems, such as smallholder dairy, agroforestry, conservation agriculture, integrated soil fertility management, and the development of multi-purpose crop varieties balancing food, feed, and soil management. Such research provides a strong base on which Humidtropics will build but such research still has not been framed within the context of farming systems at very different stages of intensification. This has led to a lack of nuanced approaches to scaling up such technologies, as reflected in the critique of conservation agriculture by Giller et al. (2009). Additionally, different Centers have often taken a different approach to system-level problems compatible with their mandates. A good example is Striga control where different centers each pursue a single, non-integrative line of research to the problem (De Groote et al., 2010; Vanlauwe et al., 2008), or the continuing issue of how to better integrate legumes into smallholder systems. Finally, other critical research areas to system intensification have received little attention, particularly increasing labor productivity in smallholder systems, a critical issue in system intensification. Humidtropics builds on this research base and lessons learned by integrating the ongoing research of the CGIAR at the level of the farming system (this issue will be discussed in more detail in Section 8). 3.3 Rural Poverty Alleviation in the Context of System Intensification Large-scale reductions in rural poverty are generally linked to economic growth and structural transformation of the economy, and are tied to both increasing farm incomes and moving population into higher-paid urban employment. As might be expected, there are larger effects from agricultural growth on poverty reduction in Sub-Saharan Africa and South Asia, but larger poverty-reducing effects of growth originating in other sectors in East Asia and Latin America (World Bank, 2007). Rural poverty rates across the humid and subhumid tropics largely vary depending on economic structure, economic opportunities outside the agricultural sector and growth rate in the rural population. Areas with high densities of rural poor correspond to the distribution of the program s three agro-ecological zones except for South Asia, where the rural poor tend to be concentrated in the semi-arid regions and river deltas. In Sub-Saharan Africa, while poverty rates tend to be higher in the semi-arid regions, by far the largest number of rural poor live in the humid highlands, with very high population densities practising rainfed agriculture, and in the higher population density areas of the humid forest. In Southeast Asia and Tropical America, rural poverty tends to be concentrated in the highlands, and this could also be associated with pressure on the natural resource base. Humidtropics offers a framework to better understand rural poverty dynamics under contrasting economic structures, market conditions, and rural population densities, the very factors that also govern intensification pathways. A key hypothesis governing impacts on rural poverty from investment in Humidtropics is that rural poverty reduction will come from direct income benefits at the early stages of agricultural intensification, but will shift to indirect income benefits in the latter stages of intensification. Indirect benefits come from increased employment in both the farm economy and the non-farm rural economy and expanded small business opportunities, all generated by increasing agricultural productivity. In what can be called the subsistence and diversification stages of intensification, the poor tend to have sufficient access to land and labor resources but are limited by market constraints, disinvestment in natural capital, and limited technological options. This is often compounded by limited education opportunities. The approach here is broadly to ensure participation of the rural poor in both market and technological opportunities, with an overriding issue of how to do this at relevant scale within what are being termed Action Areas. Impact on poverty in the early stages of intensification thus becomes a research and methodological question within the scaling process, an issue outlined later in this section. 8

14 The greater challenge is defining an impact pathway for poverty reduction in more densely populated tropical highlands. Poverty reduction in such circumstances is much more complex because of the limited and often declining farm sizes and the impact on land and labor markets from increasing market participation. Such conditions often result in poverty traps, where a degrading resource base, limited assets and inability to participate in markets results in few options for households to move out of poverty. This is compounded by some economies of scale in the intensification process, at least across the range of farm sizes in contexts such as the highlands of Ethiopia, Kenya and Rwanda. In such situations the majority of farms are often less than one hectare in size, and yet effective participation in dairy, horticulture, or even coffee requires farm sizes considerably larger as productivity significantly increases within the 2 5 hectare range. A certain minimum critical level of resources is needed to move effectively onto an intensification trajectory, but population pressures seldom allow increasing farm size, and alternatives must be explored for the poorest households falling beneath these thresholds. Livestock enterprise offers a partial solution such as dairying in the highlands of East Africa, small ruminants in the moist savannas or pigs in smallholder systems in Southeast Asia and parts of East Africa. At the same time, farm level growth in production based on increasing productivity inevitably stimulates the non-farm rural economy, which provides markets, processing opportunities, inputs, and transport within a growing agricultural economy. All of these provide employment and income possibilities for poor households. Such poverty dynamics are not well understood in terms of which phased interventions lead to the most significant impact on poverty and conducting research in a number of contrasting Action Sites will generate significant understanding, allowing more equitable growth under more densely populated conditions. 3.4 Scaling of Intensification Processes A critical and not widely explored research question is how to best move research on production system intensification to scale in order to achieve meaningful development outcomes. To date, scalability is largely associated with component technologies, reflecting the experience of the high yielding rice and wheat varieties that sparked the Green Revolution. This approach, however, represents a very narrow perspective. The situation within rainfed, smallholder agriculture is far more complex and has changed significantly since the Green Revolution. More recent impact studies demonstrate the narrow range within which technologies are adopted at scale, as Maredia and Raitzer (2006) show in their evaluation of CGIAR research in Sub-Saharan Africa. Biocontrol of cassava mealybug alone accounted for 80% of overall benefits, with varietal improvement accounting for much of the rest. Notably, mealybug biocontrol required no adoption decision on the part of farmers, relying only on the integration of the biocontrol agent into the agro-ecosystem. A wide range of project work attests to the fact that local adaptation is a requirement for successful technology adoption, and even this must cope within significant intra-farm variability such as nutrient gradients, striga infestation, or topography. Unrealistic recommendations concerning input use complicate this situation, as characterized by the early work of Global Working with multiple components linked to market demand offers a higher probability of profitable increases in productivity. This has led to a rethinking of institutional arrangements around the concept of innovation systems and moving beyond the oversimplified framework of a division between research and extension (Hall et al., 2006; Klerkx et al., 2009; Hall and Clark, 2010). The Humidtropics will build on the experience of the Sub-Saharan Africa Challenge Program (SSA-CP) and innovation systems theory, especially its development of innovation platforms (FARA, 2009; Klerkx et al., 2009). The SSA-CP was designed to integrate market development, increased productivity, and natural resource management through joint planning by principal stakeholders facilitated through an innovation platform. The SSA-CP was limited, however, to only technologies currently on the shelf (Lynam et al., 2010). 9

15 Although this will be explained in more detail in Section 4, Humidtropics will add an analytical framework to better identify critical entry points to intensify farming systems and to identify various lines of component integration and system intensification. Through the work in the Action Sites, a pipeline of component technologies will feed into this innovation platform. Methods for working across different farm typologies, especially with a focus on those below the poverty line, have not been well developed and this will be a particular focus of the methods development. Also, another critical element that goes beyond the work of the SSA-CP is to define an optimum scale at which the innovation platform will be organized and around which it will organize its activities. Scaling is central to realizing development outcomes and this in itself leads to a research agenda around institutional innovations, improved market efficiency and the reduction in the transaction costs faced by smallholders. Interaction between market access, technology options, and farmer investment strategies is central in technology development and selection of target farming systems. This research will rely on developing a well-designed farm and market agent survey capacity, with the expectation that this will be done in a panel survey design (Triomph et al., 2012). A coordinating capacity is needed to define which components of the monitoring system will be standardized across Action Areas and which will be adapted to the particular characteristics of individual sites. This system, when functioning properly, will provide the capacity to understand agricultural system change across much of the humid and sub-humid tropics, a unique international public good in itself. 3.5 Sensitivity to Different Baseline Conditions Figure 1 presents a structure for addressing different baseline conditions though systems interventions. It consists of three simplified initial situations (A, B and C) within a 2 2 population density and resource integrity matrix and a common preferred outcome. Situation A warrants balanced investment in infrastructure, markets, social cohesion, institutional strengthening, and policy. Situation B requires better infrastructure, non-agricultural employment, and strengthened supply chains. Poorer and more vulnerable people in these areas need support and the danger is that if degradation continues livelihoods will suffer, and the situation can degrade rapidly. Situation C results from relatively less stress on resources with high endemic poverty and vulnerability and limited livelihood options requiring investment in markets and livelihood alternatives that maintain resource condition while increasing incomes. Figure 1. A framework for designing R4D systems interventions. 10

16 Box 1. Lessons learned from the Sub-Saharan Africa Challenge Program (SSA-CP). The SSA-CP aimed at generating impact through the Integrated Agricultural Research for Development (IAR4D) process, which has its roots in the innovation system approach for development. The IAR4D approach was implemented through a structure called Innovation Platforms (IP), of which 36 are operational. The IAR4D approach has the following characteristics: A functional linkage point for all the stakeholders along the value chain of the specific agricultural commodity and system of production. The linkages bring together farmers, researchers, extension agents, private sector (input dealers, financial institutions, transporters etc.), policymakers and commodity end users (supermarkets, bulk buyers, institutional buyers, companies). Integration of productivity; NRM; markets; policy; product development and nutrition; and gender into the research agenda An efficient modality for organizing stakeholders for interaction and output delivery. An effective mechanism for knowledge generation and transfer to farmers and other stakeholders for over successive generations of innovations. Action research oriented toward problem solving and impact, and; Bottom up organizational development and scaling up of innovations. Although the SSA-CP is yet to conclude the proof of the effectiveness of IAR4D concept against the traditional linear approach; a recent CGIAR commissioned review of the SSA-CP program (Lynam et al., 2010) found sufficient indications that the program was on the right track. The attributes of the IAR4D concept are equally relevant to the Humidtropics and the objectives of the program partly overlap in the East and Central Africa Actiona Area with the SSA-CP. Several lessons learnt that would contribute to Humidtropics include: Market integration as process for internal as well as external quality control and adaptive management of the IPs. Increased access to institutional capitals (social, human, financial-credit) and exploring nonformal sources of credit (e.g., self help groups for community banking to leverage credit from commercial banks) requires specific facilitation, and; ICT tools are important for knowledge extension and human skill development. Three tasks for integrating the SSA-CP into the Humidtropics program are proposed: Integration of the SSA-CP pilot learning sites through selected IPs in DRC, Rwanda and Uganda as learning sites for partners in the Humidtropics. Use of SSA-CP learning sites to improve collaboration among researchers in the CG centers to address specific research issues raised on the IPs, Define scaling up process in new sites based on the experiences under the SSA-CP. Initial Action Sites in the East and Central African Action Area will include some of the most effective IPs, established through the SSA-CP. This is fully in line with a recommendation from the external review that there is an existing research infrastructure into which several million dollars has been invested that could be built upon as part of one of the CRPs (Lynam et al., 2010). The Humidtropics is an ideal match where the SSA-CP IAR4D approach and the innovation platforms operational frame concept could be used to foster quick delivery of impact along productivity, NRM, market, policy, gender and development of new products. Innovation systems approach, particularly IAR4D that has been used in the SSA-CP allows for setting up a comprehensive framework to utilize current solutions and develop future research agendas in each and all of these thematic research areas. This would be the case for Humidtropics where feedback is integrated in a broader scale and where lessons from the SSA-CP are so important. This framework serves several purposes. Since constraints are inherent within the framework, they can be better translated into activities and solutions. It provides focus to all the subsequent steps in planning and carrying out the proposed activities and facilitates a relatively complex set of objectives and questions within clear conceptual boundaries tailored to the various Action Areas in the humid tropics. It provides focus to the procedural 11

17 planning and choices from initial design selection, through instrument development or adoption, to the organization, analysis and interpretation of data at Action Site, Action Area and Global levels without being prescriptive in detail. Most important of all, is the impact of the explicit conceptual structure on subsequent inquiry in the same area. The research activities no longer stand alone but can develop into a line of inquiry that other researchers can check, replicate or build upon while reducing the effects of confounded and uncontrolled variables. In this way, Humidtropics addresses many of the thematic research priorities for Tropical Asia, Americas and Africa. It complements the priorities identified by the Asia Pacific Association of Agricultural Research Institutes (APAARI) and the Global Forum on Agricultural Research (GFAR), Its plans are in line with the priorities established by the Forum of American Agricultural Research and Technological Development (FORAGRO), particularly to promote collaboration and information exchange among government, nongovernment, farmer and private sector organizations in Central and South America. The aims of Humidtropics also relate well to those of The Comprehensive Africa Agriculture Development Program (CAADP), which is the agricultural program of the New Partnership for Africa s Development (NEPAD), which in turn is a program of the African Union (AU). In particular, Humidtropics works with the Forum for Agricultural Research in Africa (FARA) as one of the primary partners (Box 1). These relationships capitalize on South-South learning opportunities and offer immediate clients for the program s International Public Goods.. 12

18 4 STRATEGIC RESEARCH PROGRAM 4.1 Humidtropics Global and Working Hypotheses The Humidtropics Program will identify, examine, improve and promote clusters of desirable livelihood options based upon sustainable agricultural production and marketing systems within representative Action Areas in Sub-Saharan Africa, tropical Asia and tropical Latin America. These livelihood clusters represent stepwise targets for rural communities based upon farm resource endowment and household skill sets and their effects upon crop and livestock enterprises, farm input and crop production levels, conservation of soil, water and beneficial organisms, opportunities for gender equity and more accessible and fairer marketing. A Global Hypothesis (H 0 ) to this effect and related Specific Hypotheses (H 1 to H 9 ) relating to poverty alleviation, resource conservation and stakeholder equity follow: H 0 : A stepwise series of preferred livelihood clusters exist within the humid tropics where poverty reduction, balanced household nutrition, system productivity and natural resource integrity are most effectively achieved and contribute best to human welfare. Advancing these livelihood options through production systems integration, enterprise diversification and market development will have a profound effect upon food and nutritional security, household incomes and natural resource integrity where: H 1 : Intensification within the humid forest zone necessarily involves farm diversification and specialization, trends that must be weighed against natural resource integrity. Farm diversity contributes more toward food and nutritional security than does enterprise specialization, although the latter may offer greater potential for income generation. Within livelihood clusters, farm resource endowments determine how well diversification, specialization or both may be accommodated. H 2 : Innovation and policy support at farm level and above are essential to the establishment and maintenance of preferred livelihood clusters that, in the lower range require supportive incentives and in the upper range also require protection of farm viability (alternatively stated, rural landscapes are strengthened through policies and innovation that avoid large-scale resource degradation and peri-urban poverty cycles). H 3: Providing adequate water through rainwater harvesting, storage and supplemental irrigation will reduce the climatic risks of farming in rainfed areas of the sub-humid tropics, thus motivating farmers to increase crop yields through larger investment in farm inputs. Small-scale irrigation systems based on low-cost technologies lead to positive improvements in the livelihoods by enabling households to better respond to market opportunities on a year-round basis. H 4 : Livestock enterprises become more integrative and profitable in more intensively managed situations but also present increasing environmental risk and equity challenges. H 5 : Targeted livelihood clusters sequester carbon and reduce greenhouse gas emissions through the increase in carbon stocks in soils and biomass and tighter organic matter cycling. Mechanisms exist to quantify these gains and reward responsible land managers. H 6 : Biodiversity is most effectively fostered within targeted livelihood clusters below which it is exploited and threatened and beyond which it is contained or lost. H 7 : Women have greatest opportunity for high-value crop production and value-added processing within targeted livelihood clusters, below which there is insufficient demand for their high-value crops and cottage industry products; and beyond which time is more profitably spent in off-farm employment and trading. H 8 : Mobilizing the energies of skilled youth as service providers to the larger farming community is essential to upgrading livelihood clusters, below which their efforts lead to 13

19 natural resource exploitation and beyond which they are forced into idleness or migration. Examples of these services include land preparation, strategic transfer of bulky organic resources, pest and disease management and commodity processing and marketing, often using tools not readily available to most smallholder farmers. H 9 : The aged and vulnerable contribute positively to households operating at intermediate intensification gradients, below which they are exposed to numerous hazards and beyond which they risk becoming marginalized within larger society. This suite of Specific Hypotheses is intended to guide the design and analyses of integrative research activities but may have varying relevance to different agro-ecologies and communities within the program s Action Sites. These hypotheses often require expansion or re-expression into practical research questions before they may be tested. Many of these research questions are raised within later sections of this proposal. 4.2 Strategic Research Themes A dynamic program structure is built around three Strategic Research Themes; SRT1 Systems Analysis and Synthesis; SRT2 Integrated Systems Improvement and SRT3 Scaling and Institutional Innovations (Figure 2). SRT1 provides an analytical framework covering institutions and stakeholders for identifying alternative trajectories, bottlenecks and opportunities to address SLO targets. Its activities begin through conducting a Situation Analysis, including established baselines of the identified Action Areas. This Situation Analysis provides the basis for assembling the R4D innovation network, identifying specific needs, and prioritizing technological interventions, which are expected to vary among areas and sites, and identify specific entry points and actions under SRT2. Figure 2. Humidtropics program structure based on three strategic research themes and linked through M&E processes. 14

20 Activities within the Situation Analysis will also become formalized into a Monitoring and Evaluation (M&E) Framework (Section13), which establishes baseline conditions within the Action Sites and later describes adaptive integration within SRT2 and household and community benefits accruing from SRT3. Within three years, SRT1 will lead to a Global Synthesis that directs the strategies undertaken in the next phase of the program. SRT2 comprises the essential integrative field and socio-economic research aspects of the program. It is built around an R4D triangle where stepwise improvement in agricultural systems production (SRT2.2) is closely integrated with Natural Resource and Biodiversity Management (SRT2.3) and marketing of both needed farm inputs and farm surpluses, including value-added products (SRT2.1). Improved rural nutrition arises as a benefit from the R4D triangle as a whole through systems diversification, increased agro-biodiversity and market access to a wider variety of foods, but it is achieved via improved diets at the household level, often involving cottage industry. These elements will be targeted differentially based on household and community typologies in terms agro-ecological intensification trajectories. Systems productivity is viewed in the most holistic sense to include annual and perennial crops, trees and livestock and their relationship with one another and to the integrity of available farm resources, including on- and off-farm natural areas. Natural Resource Management (NRM, SRT2.3) is considered both within the context of resource conservation and ecosystem services. Renewable and conserved farm resources contribute to systems productivity through improved nutrient use efficiency and recycling, protection of soil quality and greater water use efficiency. Efficient management of ecosystem services particularly relates to those offered by trees and forests to stabilizing soil, mitigating climate change and preserving biodiversity. Successful NRM efforts prevent soil erosion, restore degraded lands and foster beneficial organisms, key considerations in systems sustainability. Community and landscape perspectives are required to assess ecosystem services including products extracted from natural common areas, carbon sequestration, status of biodiversity, access to clean water and flood control. Our focus to resource conservation is primarily at the farm and community scales as it is at these levels where individual and collective actions are best undertaken. Market integration (SRT2.1) is also considered in a holistic context to include greater availability of proven technologies and farm inputs, fuller access to fairer commodity markets and the social implications of stronger market integration. Part of the proposed research will involve comparing different marketing models, and analyzing their impacts on key household types, including women, skilled youth and the aged and vulnerable. SRT2.1 also examines value-added processing as a cottage industry and small enterprise. The overall goal of SRT2 is to identify specific farm technologies and enterprises linked in alternative rural trajectories to offer different household groups opportunities for more sustainable and profitable farming, and market progress is inherent to this process, in turn offering better lives within rural communities. It is important that findings from different Action Sites be shared within and between Action Areas. As new opportunities emerge from SRT2, they become formalized within participatory platforms that test and promote the first increments of farmer adoption, and link to the larger developmental research within SRT3. SRT3 advances the development outcomes within Action Areas (and indeed the entire Humidtropics Program). It inherits and expands the R4D learning alliances and innovation networks built upon successes of SRT2 within the context of researchable options and pathways of technology dissemination, adoption and adaptation. It necessarily involves expanding institutional capacities and effectiveness; formalized opportunities among women, youth and vulnerable; adaptation and optimization of proven technology transfer pathways; and the supportive infrastructure and policies necessary for gains developed within specific Action Sites to become realized within the larger Action Areas. Furthermore, 15

21 it is important that development outcomes be exchanged between Action Areas, by first cross-testing new technologies within additional Action Sites, and then adjusting them for wider application across the three program s agro-ecological zones. A standardized but dynamic M&E process continuously monitors the performance of SRTs. The M&E Framework first emerges from the Situation Analysis (SRT1) to produce a baseline of production, natural resource status, institutional landscape, markets and sociodevelopmental conditions within the Action Areas as a means of capturing initial livelihood strategies; but then focuses upon the Action Sites to document the emergence of new farm opportunities and unfolding integration within the R4D triangle. The focus of M&E actions within the Action Sites allows early assessment of the four CGIAR Strategic Results System Level Outcomes (SLOs) to reduce poverty and hunger, improve human health and nutrition, and enhance ecosystem resilience. The realized SLOs are better documented, however, when M&E is conducted among the program s Development Outcomes, particularly in terms of benefits realized among a prescribed number of households. The program s M&E Unit will document these gains through conducting short, standardized panel surveys reinforced by community discussions and strategic case studies for synthesis. Ultimately, the Humidtropics Program will produce International Public Goods (IPGs) that provide means to achieve these SLOs within and beyond the Action Areas. At first, these IPGs will arise from the ongoing Global Synthesis (SRT1) and be primarily methodological, particularly the extension of its M&E Framework for use by others and the publication of guidelines on integrated research within the R4D triangle. Within three years, however, these IPGs will include farm- and community-based approaches, and supporting technologies and trainers, that increase food security, reduce rural poverty, and improve household nutrition in a manner that ensures sustainable management of natural resources in the major eco-zones of the humid and sub-humid tropics. In terms of improving food security within the Action Areas over the next 15 years, the program seeks to increase farm yields by 60% while reducing poverty by 25% within its Action Areas. Increased crop diversification will lead to fully balanced and better diets, including among women and children. Sustainable management of natural resources will be practiced by 40% of farm households, protecting millions of hectares that would have otherwise degraded (see SRT2.3). This sustainable management will better direct and provide greater opportunities to skilled youth, women and more vulnerable members of the community. These are challenging goals, but ones that are achievable by a talented network operating within the robust program structure proposed by Humidtropics. 4.3 Impact Zones, Action Areas, and Action Sites This section describes the geographical logic of the Humidtropics in line with the overall research framework and developed around Impact Zones, Action Areas, and Action Sites. Specific selection and sampling strategies are in line with the Specific Hypotheses and related research questions in a manner that stimulates IPGs and facilitates both scaling up of needed technologies and zooming in on underlying processes, both of which are necessary to create and refine an environment for large-scale impact Impact Zones The Humidtropics is focusing its activities in three Impact Zones, based on agro-ecological conditions and minimum population densities. The targeted humid lowlands have growing periods of at least 270 days per year, altitudes less than 1000 m, and population densities of at least 15 people km -2. The targeted sub-humid and moist savannas have growing periods between 180 to 270 days per year, occur at altitudes less than 1000 m, and support populations of at least 30 people km -2. Targeted highlands have growing periods of greater than 180 days per year, altitudes above 1000 m, and population densities greater than 30 16

22 people km -2. This selection process is largely covered within the Justification (see Section 3). Overall, these areas cover approximately 3 billion ha and contain a total population of 2.9 billion people (Table 1). Table 1. Selected characteristics of the Humidtropics Impact Zones. Impact zone Population (million people) a Land area (million ha) b Average length of growing period (days) c Average altitude (masl) d Humid Lowlands 825 1, Africa Tropical Americas Asia Moist savannas 1,591 1, Africa Tropical Americas Asia 1, Tropical highlands ,557 Africa ,431 Tropical Americas ,830 Asia ,556 Total/Average 2,866 2, Sources: a Center for International Earth Science Information Network (CIESIN), Columbia University; International Center for Research on Tropical Agriculture (CIAT), Gridded Population of the World (GPW), Version 3 (beta). Palisades, NY: CIESIN, Columbia University. Available at b Notenbaert A, author's calculations, using ArcGIS. c Jones P G, Thornton P K, Global length-of-growing-period surfaces for a suite of Fourth Assessment climate models. Digital data set, ILRI, Nairobi, Kenya. d: global 30 arc second elevation data. These Impact Zones define the geographical focus of the Humidtropics in broad terms but have no further functional role in its research strategy, which is based on the representative Action Sites in key Action Areas within these zones (Figure 3) Action Area selection and characteristics To realize tangible synergies between Centers and partners, Humidtropics opts for a common approach towards R4D in welldefined Action Areas across Tropical Africa, America, and Asia. Action Areas are strategically selected parts of the target Impact Zones where intensification is a must, thus excluding areas below a minimal population density (e.g., forest margins which are targeted by CRP6) and the areas with intensive lowland rice production (which are targeted by other CRPs, see Section 9). All Action Areas occupy varying positions along the poverty NRM continuum, based on aggregate data presented in tables 4, 6 and Figure 3. Approximate location of the Action Areas along the poverty NRM status axes. 17

23 8 (Figure 3), allowing better comparison and wider learning across more broadly applicable farming conditions and community settings. In this way, the choice of Action Areas within Humidtropics shapes its relevance because this determines the range of interventions that the program invests in and the applicability of the public goods it generates. Three major criteria were applied for identifying Action Areas (Figure 3) within Humidtropics: 1. They are both representative and capture diversity. Since the Humidtropics is targeting humid and sub-humid environments, the length of growing period is at least 180 days per year and often much greater. Within these environments, Action Areas contain widespread poverty and poor natural resource integrity, justifying the investments towards the System Level Outcomes, targeted by the program 2. Urgent need exists for large-scale impacts. The Action Areas have relatively high population densities and poverty levels where intensification, production systems diversification and integration, and rural transformation are urgently needed. In most cases, natural resource degradation and vulnerability to climate change pose great risks to the poor and complicate progress toward rural development. This criterion includes areas with high absolute population densities, and therefore significant potential for increased market participation (e.g., the African and Andean highlands). 3. They advance earlier investments and existing and potential partnerships. Planned interventions are based on collaboration to generate new synergies between existing expertise of the Centers and their partners. At the same time, many earlier research investments in the Action Areas have not achieved their full impact. These earlier initiatives form the basis for creating more effective and efficient synergies between Centers and partners towards achieving unrealized impact and better working together in the future. Figure 4. Strategic selection of Action Areas and Action Sites. SRT1 provides a strategy for selecting Action Sites within Action Areas, SRT2 develops best-fit interventions on productivity, markets, and NRM at the Action Site level and SRT3 advances an institutional framework for scaling up within the Action Areas and beyond (Figure 4). SRT1 will also implement global synthesis activities across all Action Areas. 18

24 Table 2. Selected characteristics of each Action Area. Action Area Tropical Africa Western humid lowlands Central humid lowlands Southern humid lowlands East and Central highlands Western moist savannas Southern moist savannas Tropical Americas Central American humid lowlandhighland transect Andes humid lowland-highland transect Tropical Asia Greater Mekong sub-humid lowland to highland transect Indonesian humid lowlands South Asian subhumid lowlands Population (million people) a Land area million ha) b Average rainfall (mm) c [LGP (days)] d ,727 [256] ,427 [243] ,942 [276] ,347 [240] ,198 [191] ,356 [202] ,044 [270] ,658 [309] ,834 [231] ,645 [346] ,865 [253] Average altitude (masl) e Major farming systems following the Dixon classification (% of the land area) f 325 Tree crops (68%) 612 Forest based (45%); Cereal-root crop mixed (33%) 670 Rice-tree crop (73%) 1,566 Highland perennial (56%); Maize mixed (23%); Highland temperate mixed (16%) 293 Cereal-root crop mixed (45%) / Root crop (43%) 718 Cereal-root crop mixed (35%); Root crop (26%); Maize mixed (22%) 370 Coastal plantation & mixed (63%); Maize-beans (35%) 1,335 Intensive highland mixed (41%); Coastal plantation & mixed (29%); Highland altitude mixed (23%) 605 Upland intensive mixed (39%); Highland extensive mixed (29%); Lowland rice (26%) 259 Mixed tree crops mixed (59%) 444 Highland mixed (56%); Rice (31%) Sources: a Center for International Earth Science Information Network (CIESIN), Columbia University; International Center for Research on Tropical Agriculture (CIAT), Gridded Population of the World (GPW), Version 3 (beta). Palisades, NY: CIESIN, Columbia University. Available at b Notenbaert A, author's calculations, using ArcGIS. c Hijmans R.J., Cameron S.E., Parra J.L., Jones P.G., Jarvis A, (2005). Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25, d Jones P G, Thornton P K, Global length-of-growing-period surfaces for a suite of Fourth Assessment climate models. Digital data set, ILRI, Nairobi, Kenya. e : global 30 arc second elevation data. f Dixon, J., Gulliver, A. & Gibbon, D. (2001). Farming systems and poverty: improving farmers'livelihoods in a changing world. Rome, Italy: Food and Agriculture Organization of the United Nations (FAO); Washington, DC:World Bank. f Dixon, J., A. Gulliver, D. Gibbon, and M. Hall (ed.), 2001: Farming systems and poverty: Improving farmers' livelihoods in a changing world. FAO and World Bank, Rome and Washington, D.C., 412 pp. [available online from Based on above considerations, the Humidtropics will work in a total of 11 Action Areas of which 6 are in Tropical Africa, 2 in the Tropical Americas, and 3 in Tropical Asia (Table 2). These Action Areas cover 11% of the population in the humid and sub-humid tropics and 9% of the land area (Table 3). Within the current set of Action Areas, the level of activity and cross-center cooperation is variable from effective and operational with various Centers cooperating within a broad set of themes and through functional partner networks (Tier 1 Action Areas); to embryonic with a single Centre leading a set of activities of a limited scope (Tier 3 Action Areas). Tier 2 Action Areas are intermediate between both Tier 1 and 3. The 19

25 long-term vision of CRP1.2, beyond the initial 3-year period is to have active R4D platforms in all Action Areas within the logic of the overall Humidtropics research framework. Some required actions within these Tiers follow: Tier 1 Action Areas. Tier 1 Action Areas typically contain on-going R4D activities focusing on various components within specific farming systems and through cooperation between various Centers and national research systems, backstopped by Advanced Research Institutes. Partnerships usually include value chain and dissemination partners and substantial investments are made in the promotion and dissemination of best-fit technologies, supported by an effective M&E framework and partner capacity building. In Tier 1 Action Areas, priority interventions needed for transforming current partnerships into R4D Platforms include: 1) broadening of partnerships to cover all value chain components, 2) consolidating R4D innovation networks with clear goals, roles, and responsibilities, and effective interactions, 3) widening available expertise to include an innovation network perspective thereby enabling interventions for all system components, 4) instilling a systems approach towards intensification and natural resource integrity and biodiversity, 5) strengthening the dissemination tools and approaches to enable integration of complex knowledge, and 6) strengthening capacity for learning, evaluation, and scaling up. Tier 2 Action Areas. Tier 2 Action Areas typically have on-going R4D activities usually focused on one or few system components, promoted by a single Centre. Market linkage and value chain investments are usually minimal and dissemination partners are commonly promoting single component technologies developed by that Center. Systems research is not practiced but R4D is understood. Besides those detailed above, the following interventions will also be required to initiate R4D innovation networks in Tier 2 Action Areas: 1) incorporating other partner Centers in Action Area activities, sometime through linkage with other CRPs, 2) promoting the concept of innovation within the research agendas and development activities operating in that Action Area, and 3) identifying and establishing value chain and value-addition partnerships. Tier 3 Action Areas. Tier 3 Action Areas commonly have a very limited range of activities focused on specific, often isolated technologies without clear linkage to markets and with weak and disorganized dissemination channels. Tier 3 Action Areas also include parts of Action Areas lacking Center-lead activities but with potential for systems research and impact (e.g., Haiti in the Central American Action Area). All interventions mentioned under the Tier 1 and 2 Action Areas will be equally required for Tier 3 Areas. In addition to reduced capacities in technology development and outreach, the baseline conditions are often poorly documented and M&E skills lacking. This situation offers one advantage, however, because here all Centers may be introduced to national and local cooperators as an established team rather than a new partner. Table 3. Population and land area in the Action Areas. Population and land area characteristics Africa Americas Asia Overall total Total population (million people) a ,908 2,866 Population of Action Areas (million people) a Proportion of the population in Action Areas (%) Total land area (million ha) b 772 1, ,964 Land area of Action Areas (million ha) b Proportion of the land area in Action Areas (%) Sources: a Center for International Earth Science Information Network (CIESIN), Columbia University; International Center for Research on Tropical Agriculture (CIAT), Gridded Population of the World (GPW), Version 3 (beta). Palisades, NY: CIESIN, Columbia University. Available at b Notenbaert A, author's calculations, using ArcGIS. 20

26 4.3.3 Action Areas in Tropical Africa Six Action Areas, one in each sub-region are proposed for Tropical Africa: 1) the Western humid lowlands, 2) the Central humid lowlands, 3) the Southern humid lowlands, 4) the East and Central African highlands, 5) the Western moist savannas, and 6) the Southern moist savannas (Table 4, Figure 5). Population densities and poverty levels are relatively high as 28 to 71% of people earn less than US $1.25 per day with land degradation, a proxy for natural resource integrity status, averaging 40 to 90% of the land area (Table 4). Numerous partnerships exist within the Tropical Africa Action Areas that can form the basis of initiating Humidtropics activities (Table 5) Action Areas in Tropical America Two Action Areas are proposed in Tropical America: 1) Low to mid-altitude humid and subhumid Central America (and the Caribbean), and 2) the mid-altitude humid Andes (Table 6, Figure 6). Population densities and poverty levels are relatively low in these areas as only 5 to 20% of people earn less than US $1.25 per day but land degradation is widespread, averaging 46 to 72% of the land area (Table 6). Several partnerships exist within the Tropical Americas Action Areas that can form the basis of initiating Humidtropics activities (Table 7) Action Areas in Tropical Asia Three Action Areas have been identified in Asia: 1) the Greater Mekong Sub-region, 2) Indonesia, and 3) South Asia (Table 8, Figure 7). Population densities and poverty levels are relatively high, with 20 to 29% of people earn less than US $1.25 per day with land degradation averaging 31 to 44% of the land area (Table 8). Several partnerships exist within the Tropical Asia Action Areas that can form the basis of initiating Humidtropics activities (Table 9). 21

27 Table 4. Selected characteristics of Action Areas in sub-saharan Africa. Action Area Political boundaries Western humid lowlands Central humid lowlands Southern humid lowlands East and Central highlands Western moist savannas Southern moist savannas Ivory Coast: Agnéby, Bas-Sassandra, Dix-Huit Montagnes, Fromager, Haut-Sassandra, Lacs, Lagunes, Marahoué, Moyen-Cavally, Moyen-Comoé, N'zi-Comoé, Sud-Bandama, Sud-Comoé regions; Ghana: Western, Central, Ashanti, Eastern regions; Nigeria: Benue, Anambra, Enugu, Ebonyi, Imo, Rivers, Abia, AkwaIbom, Cross River states; Cameroon: Northwest, Southwest, West, Littoral, Centre provinces Angola: Uíge, Zaire provinces; DR Congo: Lukaya, Cataractes, Bas-fleuve districts of Bas-Congo province and Kwileu, Kwango districts of Bandundu province Madagascar: Diana, Sava, Alaotra Mangoro, Atsinanana, Analanjirofo, Vatovaavy-Fitovinany, Atsimo-Astinanana regions Rwanda: Northern, Eastern, and Southern provinces; Burundi: Cibitoke, Bubanza, Bujumbura, Mwaro, Gitega, Muramvya, Kayanza, Ngozi, Karuzi, Ruyigi, Muyinga, Kirundo, Cankuzo provinces; DRCongo:, Kalehe, Mwenga, Kabare, Rutshuru, and Masisi Terriories in Sud- and Nord-Kivu Provinces; Uganda: Kabale, Kisoro, Rukungiri, Bushenyi, Mbarara, Rakai, Masaka, Mpigi, Mubende, Kiboga, Luwero, Kamuli, Jinja, Iganga, Torora, Pallisa, Mbale, Kapchorwa districts; Kenya: Nyanza and Western Provinces; Ethiopia: Kamashi, East Wellega, West Wellega, West Shewa, Gurage, Hadiya, Yem, Illubabor zones Ghana: Northern, Upper West, Upper East regions; Togo: Kara, Centrale regions; Benin: Borgou, Donga departments; Nigeria: Niger, Kaduna, Kwara states Zambia: Northern Province: Kaputa, Mpulungu, Mporokoso, Luwingi districts of Northern province and Chiengi, Nchelenge, Kawambwa, Mwense, Mansa districts of Luapula province; Malawi: Central, Northern regions; Tanzania: Dodoma region; Population density (km -2 ) a Poverty (% of the population living on < Average market access (hours) c US$1.25 day -1 ) b Degraded area (% of total) d Mozambique: Zambezi province Sources: a Center for International Earth Science Information Network (CIESIN), Columbia University; International Center for Research on Tropical Agriculture (CIAT), Gridded Population of the World (GPW), Version 3 (beta). Palisades, NY: CIESIN, Columbia University. Available at and Notenbaert A, author's calculations, using ArcGIS. b Wood, S., G.Hyman, U. Deichmann, E. Barona, R., Z. Guo, S. Castaño, O.Rivera, E. Diaz, J. Marin. Additional data and technical contributions from: H. Coulombe and Q. Wodon, M. Muniz, S. Benin, T. Benson, J. Berdegue, J. Gonzalez, A. Barufi, P.Lanjouw and K.Simler, B. Blankespoor and S. Murray $1 and $2 day poverty maps of the developing countries. Input to the CGIAR Strategy and Results Framework. NzBjY2Q3Y2Iw c Nelson A Travel time to major cities: A global map of Accessibility. Global Environment Monitoring Unit - Joint Research Centre of the European Commission, Ispra Italy. Available at d Bai ZG, Dent DL, Olsson L and Schaepman ME Global Assessment of Land Degradation and Improvement. 1 Identification by remote sensing. Report 2008/01(GLADA Report 5), ISRIC - World Soil Information, Wageningen, 70p. 22

28 Figure 5. Selected characteristics of Action Areas in Tropical Africa. Data source: see references in Table 2. 23

29 Table 5. Existing partnerships within the Action Areas in Tropical Africa. Action Area Western humid lowlands Central humid lowlands East and Central highlands Partnership Sustainable Tree Crops Program (STCP) Innovate Plantain CIALCA [legume-based systems] CIALCA [banana and legume-based systems] Sub-Saharan Africa Challenge program [value chains] Challenge Program on Water and Food Partners CGIAR/ IARCS IITA, ICRAF, CIFOR Bioversity, IITA, CIRAD CIAT-TSBF IITA, Bioversity, CIAT-TSBF CIAT, IITA ILRI, ICRAF, IWMI NARES/Universities NGOs Private sector Cocoa Research Institute of Ghana, Institut de Recherche pour le Développement (IRAD Cameroon), Cocoa Research Institute of Nigeria, Danish Center for Forest, Landscape and Planning and Royal Veterinary and Agricultural University, US Department of Agriculture (USDA/ARS), The Ministry of Planning, Programming and Regional Development, Cameroon CARBAP (Cameroun), NIHORT (Nigeria), Crops Research Institute (Ghana), CNRA (Cote d Ivoire), INERA (DRC) INERA, PRONANUT, Kinshasa University INERA, ISAR, ISABU, IRAZ, UCB, etc INERA, ISAR, NARO Makerere (Uganda), National University of Rwanda and ISAE (Rwanda). University of Goma, DRC EIAR, NMA (National Meteorology Agency), ARARI (Ethiopia), ORARI (Ethiopia), MoA, Bahir Dar University, Ambo University, Wollega University, Addis Ababa University, Wollo University, Cornell university ANADER Côte d Ivoire, Société de Coopération pour le Développement International (Canada), TechnoServe, IFDC Aprodec, BDD, etc. -- Diobass, CRS, Action contre la Faim, etc. Diobass (DRC), Action de Sud (Rwanda) and Konaplikan (DRC) ODI, SEI, CRS, Ethiopian Economic Association, Ethiopian Rainwater Association World Cocoa Foundation, MARS, Nestle, ADM, Hershey, other cocoa companies, Yaro, and other Fertilizer Companies MEA Limited, Agrobiotech, etc Urwibutso (Rwanda), Huntex (uganda); Gap Pharmakina (DRC) -- Western moist savannas Sub-Saharan Africa Challenge program Challenge Program on Water and Food Agricultural Water Management Solutions IFDC, IWMI, ILRI IWMI Institute for Agricultural Research (IAR), Nigeria; National Animal Production Research Institute (NAPRI), Nigeria University for Development Studies (UDS), Tamale, Ghana; Savanna Research Institute (SARI), Tamale, Ghana, Water Resources Commission (WRC), Ghana University for Development Studies (UDS), Tamale, Ghana; Savanna Research Institute (SARI), Tamale, Ghana, Water Research Institute (WRI)-CSIR Farm and Infrastructure; Foundation; Sasakawa Global 2000; Information Resources and Network Services International Development Enterprise (IDE) International Development Enterprise (IDE) Fertilizer Producers and Suppliers; Association of Nigeria (FEPSAN), Manufacturers Association of Nigeria (MAN), Agrochemicals Group of Manufacturers Association of Nigeria (AGROMAN)

30 Action Area Western moist savannas Southern moist savannas Partnership Sub-Saharan Africa Challenge program Platform Mozambique Partners CGIAR/ IARCS IFDC IITA NARES/Universities NGOs Private sector Institute for Agricultural Research (IAR), Nigeria; National Animal Production Research Institute (NAPRI), Nigeria Instituto Investigação da Agrária de Moçambique (IIAM) Mozambique, Ministry of Agriculture Mozambique, Catholic University of Cuambo, EMBRAPA Farm and Infrastructure; Foundation; Sasakawa Global 2000; Information Resources and Network Services IKURU, CLUSA, World Vision Fertilizer Producers and Suppliers; Association of Nigeria (FEPSAN), Manufacturers Association of Nigeria (MAN), Agrochemicals Group of Manufacturers Association of Nigeria (AGROMAN) EMD Crop Biosciences, SoyGro East and Central highlands; Southern moist savannas Southern moist savannas; Southern humid lowlands Malawi Agroforestry Food Security Program (AFSP) Sub-Saharan Africa Challenge program BARNESA Small-scale cassava processing and vertical integration of the cassava sub-sector in Southern and Eastern Africa- Phase II ICRAF Bioversity, AVRDC Bioversity, IITA IITA ZARI (Zambia), DARTS (Malawi), LRCD (Malawi), IIAM (Mozambique) IIAM (Mozambique), Bvumbwe Agricultural Research Station (Malawi). Bunda Agricultural College, (Malawi) NARO (Uganda), ISABU (Burundi), ISAR (Rwanda), INERA (DRC), ARC (Sudan), ARDI (Tanzania), Ministry of Agriculture (Malawi), KARI (Kenya), IIAM (Mozambique) Tanzania Food and Nutrition Center, Centre National de la Recherche Applique au Developpment Rural (FOFIFA), Zambia Agricultural Research Institute (ZARI) Total Land Care, Catholic Relief Services CONCERN, World Vision International, Plan International World Vision (Malawi) National Farmers Association of Malawi, Conservation Farming Unit of the Zambia National Farmers Union Semente Perfeita Seed Company (Mozambique), Dengo Seeds (Malawi) 25

31 Table 6. Selected characteristics of Action Areas in Tropical Americas Action Area Political boundaries Population density (km -2 ) a Central American humid lowlandhighland transect Andes humid lowlandhighland transect Costa Rica: Guanacaste department; El Salvador: La Unión department; Guatemala: Chimaltenango, Chiquimula, Guatemala, Jalapa, Jutiapa, Petén departments; Honduras: Copán, Choluteca, El Paraíso, Intibucá, Lempira, Valle, Cortés, Atlantida, Yoro departments; Nicaragua: Chinandega, Estelí, Nueva Segovia, Madriz, León, Matagalpa, Rivas, RAAN, RAAS departments; Haiti: Nord, Nord-Est, Artibonite, Centre, Grand Anse, Sud dept. Peru: Cajamarca, Huanuco, Junin departments; Ecuador: Carchi, Imbabura, Pichincha, Bolivar, Tungurahua, Chimborazo provinces; Colombia: Antioquia, Choco, Risaralda, Caldas, Tolima, Valle del Cauca, Poverty (% of the populatio n living on Average market access (hours) c < US$1.25 day -1 ) b Degrade d area (% of total) d Huila, Cauca, Nariño departments Sources: a Center for International Earth Science Information Network (CIESIN), Columbia University; International Center for Research on Tropical Agriculture (CIAT).,2004. Gridded Population of the World (GPW), Version 3 (beta). Palisades, NY: CIESIN, Columbia University b Wood, S., G.Hyman, U. Deichmann, E. Barona, R., Z. Guo, S. Castaño, O.Rivera, E. Diaz, J. Marin. Additional data and technical contributions from: H. Coulombe and Q. Wodon, M. Muniz, S. Benin, T. Benson, J. Berdegue, J. Gonzalez, A. Barufi, P.Lanjouw and K.Simler, B. Blankespoor and S. Murray $1 and $2 day poverty maps of the developing countries. Input to the CGIAR Strategy and Results Framework. c Nelson A Travel time to major cities: A global map of Accessibility. Global Environment Monitoring Unit - Joint Research Centre of the European Commission, Ispra Italy. Available at. d Bai ZG, Dent DL, Olsson L and Schaepman ME Global Assessment of Land Degradation and Improvement. 1 Identification by remote sensing. Report 2008/01(GLADA Report 5), ISRIC - World Soil Information, Wageningen, 70p Table 7. Existing partnerships within the Action Areas in the Tropical Americas. Action Area Central American humid lowlandhighland transect Andes humid lowlandhighland transect Central American and Andes humid lowlandhighland transect Partnership Eco-efficient crop-livestock systems Genetic resources conservation NRM in the Andes Papa Andina Initiative Eco-efficient crop-livestock systems MUSALAC Partners CGIAR/IARCS NARES/Universities NGOs Private sector CIAT, CIMMYT, CATIE CIAT, Bioversity, CATIE, CIMMYT CIAT CIP CIAT, CIMMYT, CPWF Bioversity, CIAT, CIRAD ICTA (Guatemala), DICTA (Honduras), INTA (Nicaragua), Red SICTA (?) CENTA (El Salvador) INTA Nicaragua, DICTA, SICTA CRS,OXFAM, SNV, Swiss Contact, GTZ CARE, CRS, Heifer. Local NGOs Zamorano University (Honduras) Seed enterprises Corpoica (Colombia), INIAP (Ecuador), INIA (Peru), INIAF (Bolivia) CONDESAN -- INIAP (Ecuador), PROINPA FOVIDA (Peru) -- (Bolivia), Corpoica CONDESAN Fundallanura; Fedarroz UNAN Leon (Nicaragua), CORBANA (Costa Rica), IDIAP (Panama), INIA (Peru), CORPOICA (Colombia), IDIAF (Dominican Republic), MANRN (Haiti), INIFAP (Mexico), INIAP (Ecuador), EMBRAPA (Brazil) Asociacion de Plataneros (Guatemala) FHIA (Honduras) 26

32 Figure 6. Selected characteristics of Action Areas in the Tropical Americas. Data source: see references in Table 2. 27

33 Table 8. Selected characteristics of Action Areas in Tropical Asia. Action Area Greater Mekong sub-humid lowland to highland transect Political boundaries Laos: Phongsali, Luang Namtha, Bokeo, Oudomxay, Luang Prabang, Houaphanh, Xieng Khouang, Sayaboury, Borikhamxay, Khammuane, Savannakhet, Saravane, Sekong, Attapeu provinces, Cambodia: Rattanak Kiri, Mondul Kiri, Kratie, and Strung Treng Provinces; Vietnam: Bắc Giang, Bắc Kạn, Cao Bang, Ha Giang, Lang Son, Lao Cai, Phu Tho, Quang Ninh, Thai Nguyen, Tuyen Quang, Yen Bai, Dien Bien, Lai Chau, Son La, Hoa Binh, Dak Lak, Dak Nong, Gia Lai, Kon Tum, Lam Dong provinces; Population density (km -2 ) a Poverty (% of the population living on < US$1.25 day -1 ) b Average market access (hours) c Degraded area (% of total) d Indonesian humid lowlands South Asian subhumid Indonesia: West Sumatra, Jambi, South Sumatra, and Lampung provinces India: Kerala, Meghalaya, Tripura, and Nagaland states lowlands Sources: a Center for International Earth Science Information Network (CIESIN), Columbia University; International Center for Research on Tropical Agriculture (CIAT), Gridded Population of the World (GPW), Version 3 (beta). Palisades, NY: CIESIN, Columbia University. Available at and Notenbaert A, author's calculations, using ArcGIS. b Wood, S., G.Hyman, U. Deichmann, E. Barona, R., Z. Guo, S. Castaño, O.Rivera, E. Diaz, J. Marin. Additional data and technical contributions from: H. Coulombe and Q. Wodon, M. Muniz, S. Benin, T. Benson, J. Berdegue, J. Gonzalez, A. Barufi, P.Lanjouw and K.Simler, B. Blankespoor and S. Murray $1 and $2 day poverty maps of the developing countries. Input to the CGIAR Strategy and Results Framework. NzBjY2Q3Y2Iw c Nelson A Travel time to major cities: A global map of Accessibility. Global Environment Monitoring Unit - Joint Research Centre of the European Commission, Ispra Italy. Available at. d Bai ZG, Dent DL, Olsson L and Schaepman ME Global Assessment of Land Degradation and Improvement. 1 Identification by remote sensing. Report 2008/01(GLADA Report 5), ISRIC - World Soil Information, Wageningen, 70p 28

34 Figure 7. Selected characteristics of Action Areas in Tropical Asia. Data source: see references in Table 2 Table 9. Existing partnerships within the Action Areas in Tropical Asia. Action Area Greater Mekong Indonesia Partnership CPWF- Mekong; MRC Partners CGIAR/ IARCS CIAT; IWMI; ILRI; WorldFish; ICRAF; IRRI NARES/Universities NGOs Private sector NAFRI (Lao PDR); WREA (Lao PDR); SFRI (Vietnam); VAAS (Vietnam); MOWRAM BAPNET Bioversity Indonesian Tropical Fruit Research Institute (ITFRI); Indonesian Center for Horticultural Research and Development (ICHORD); Philippine Council for Agriculture, Natural Resources Research and Development (PCARRD); Bureau of Plant Industry-DA (Philippines); National Agricultural Research Institute (NARI, PNG); Bangladesh Agricultural Research Institute (BARI); Cambodian Agricultural Research and Development Institute (CARDI); China - Guangdong Academy of Agricultural Sciences (GAAS); India - National Research Centre on Banana (ICAR); Malaysian Agriculture Research and Development Institute (MARDI); Myanmar - Ministry of Agriculture and Irrigation (MAI); Sri Lanka - Horticultural Crop Research and Development Institute (HORDI); Thailand - Horticulture Research Institute (HRI); Vietnam Agricultural Science Institute (VASI) Vegetables for AVRDC IVEGRI/Udayana University/BTPT/DINAS/ FIELD -- Indonesia 29

35 4.3.6 Action Site Selection Within each Action Area, Action Sites will be identified in line with opportunities to address CRP1.2 research hypotheses and questions within which R4D platforms will be operationalized (Figure 8). The choice and number of Action Sites is a crucial component in the research strategy of the Humidtropics and will be driven by the main research questions towards the development of IPGs within and between Action Areas. Decisions concerning Action Sites and their farming systems of interest essentially define the research framework on sustainable system intensification within the Humidtropics program. The unit of analysis in the Humidtropics, particularly within SRT2 where technical innovation originates, is the farming system. The intent is to understand change in those farming systems, firstly as influenced by the main drivers of change, particularly population density and market development (see Section 3), and secondly by system interventions coming from research, including a wide range of technical, post-harvest and market innovations. Action Sites will thus be covering representative farming systems within the existing variation in population density and market access nested into an Action Area. At the highest level, an Action Site is defined within a priority value chain, chosen in accordance with the intensification trajectory within a certain Action Area, and includes all the partners operating within that value chain, including farmer associations, private business, development agents, and local policymakers. An Action Site is in effect a group of farming associations located within a common (set of) small watershed(s) and working together within a value chain through R4D platforms. Politically, an Action Site can be equated to the administrative sub-units that are best in line with the above watershed, often operating under different designations such as a Location in Kenya, a County in Uganda, or a District in Rwanda. This will allow access to geo-referenced indicators that are crucial for measuring impact of the Humidtropics. The number of Action Sites in each Action Areas will depend on several factors: Population and market access gradients: The Action Sites need to cover existing population market access gradients within each Action Area (Figure 8). Variation exists in population density and market access along different ranges for the different Action Areas (Table 10). Decision on the number of Action Sites to be identified along these gradients will be based upon the Situation Analysis under SRT1 (see Section 4.5). Figure 8 shows in situation (a), population or market access should not be considered as a variable between Action Sites (refer to the vertical arrow) while in situation (b) and (c), Action Sites will need to cover existing gradients in population or market access (2 sites in situation (b) and 3 sites in situation (c)). Figure 8. Hypothetical cumulative frequency distributions of population and market access within a single Action Area. 30

36 Table 10. Variation in population density and market access for the different Action Areas. Action Area Population density (% of the area) a Average market access (% of the area) b <= 15 <= 1 hr 1-2 hr 2-4 hr > 4 hr km km km -2 > 60 km 2 Tropical Africa Western humid lowlands Central humid lowlands Southern humid lowlands East/Central highlands Western moist savannas Southern moist savannas Tropical Americas Central American humid lowlandhighland transect Andes humid lowland-highland transect Tropical Asia Greater Mekong sub-humid lowland to highland transect Indonesian humid lowlands South Asian subhumid lowlands Sources: a Center for International Earth Science Information Network (CIESIN), Columbia University; International Center for Research on Tropical Agriculture (CIAT), Gridded Population of the World (GPW), Version 3 (beta). Palisades, NY: CIESIN, Columbia University. Available at and Notenbaert A, author's calculations, using ArcGIS. c Nelson A Travel time to major cities: A global map of Accessibility. Global Environment Monitoring Unit - Joint Research Centre of the European Commission, Ispra Italy. Available at. Model for integration of activities: Under SRT3 (see Section 4.9 below), two models for cooperation are proposed, one model whereby R4D platforms drive activities towards the identification and implementation of a comprehensive research agenda based upon entry points and innovation, and another model whereby platforms are merely facilitating value chain partners to cooperate, more in line with the innovation platforms of the SSA-CP. While it is certainly not the intention to duplicate all Action Sites identified in using both models, a limited number of Action Sites with similar population market access characteristics will be developed using each of these two models to answer specific research questions developed under SRT3. Replication of Action Sites: Action Sites with similar characteristics must be replicated within an Action Area to allow meaningful analysis of results. The exact number of replicates will be determined through SRT1 and its M&E Unit. In some cases, similar communities within a larger Action Site may serve as replicates, or different communities may serve to contrast interventions, again based upon results of the Situation Analysis. In conclusion, the exact number of Action Sites within an Action Area will likely vary between nine (e.g., = 2 population market access levels 3 replicates + 1 alternative model 3 replicates) and 24 (6 population market access levels 3 replicates + 2 alternative model 3 replicates), with the possibility of downward adjustment based upon 31

37 the Situation Analysis. The number of Action Sites may also be adjusted to reflect improvement of Tier status as partnership consolidates. Indeed, Action Sites will be periodically revised in terms of number, location, and types of activities in response to achievements, needs, and opportunities. Action Site exit strategies will likely result from successful transition from SRT2 to SRT3 (focus from Action Sites to Action Areas) but this is not anticipated over the first three years of the program. 4.4 SRT1 Systems Analysis and Synthesis People s lives and livelihoods in the humid and sub-humid tropics are complex, diverse, dynamic, vulnerable and greatly influenced by the environmental context in which they operate. One of the primary actions of the Humidtropics will be to understand the status of and variation in the current agricultural, social, economic, natural, institutional, and policy situation, including gender, within and between Action Areas. This understanding will form the basis for putting in place a correct framework for priority setting and identifying the appropriate blend of productivity improvement, natural resource management, market development and institutional strategies. In addition, these rural communities, and stakeholder groups within them, are affected by population growth, climate change, weak rural infrastructure and difficult commodity markets. An understanding of these drivers and their potential impact on people s lives is crucial information towards priority setting or improved integrated systems performance, and Humidtropics will provide this understanding. For the program to deliver on its goals, effective, complementary and efficient partnerships with an innovation perspective that cover the entire Humidtropics Program Structure are crucial. SRT1 serves to identify Action Sites and the most promising interventions through balanced priority setting for increased production, poverty alleviation, market development, natural resource integrity, and biodiversity enhancement. Indeed, the Action Sites within Action Areas that are currently operational will serve as a starting point but further guidance from SRT1 will provide a basis for transition to a more systematic set of Action Sites and activities within the project s first three years Research Questions and Methodology The assessment and analysis of the existing systems will take place at two levels. Firstly, the innovation system and institution level will address the array of potential stakeholders operating in the Action Areas, their capacities and their linkages for addressing the SLOs, essentially for both on-the-ground, policy and institutional change. This perspective will also take into account value chains and farmer and community organizations and local governments. Secondly, the needs of the poor and vulnerable within the Action Areas will be characterized using Sustainable Livelihoods guidance sheets, developed by DFID (2000), and using recently available information generated by stakeholder institutions at each Action Area. This will form the basis to develop livelihood typologies based largely on the analysis of Titonnell et al. (2010) that in turn determine the Action Sites and entry points for production increase, poverty reduction, market development and more sustainable natural resource management. This process is supported through the expression of key SRT1 Research Questions, including: 1. How many Action Sites are required to provide a basis for generating reliable and robust technologies scalable to the larger Action Areas? How many communities, households, farm types and (by extension) target number of beneficiaries should an Action Site contain? 2. What are alternative configurations of key stakeholders that facilitate learning alliances and innovation networks addressing the multiple objectives linked to improving the SLOs in the Action Areas? How can these be identified? 3. Which baseline information drawn across the Action Areas is required to prioritize entry points for integrative research? Who are major partners in these Areas? How is this 32

38 information best compiled, shared and analyzed among partners? How may comparable activities be framed within different Action Sites and Eco-zones to permit useful cross-site analyses? 4. How can drivers of change be identified which capture issues related to opportunities and bottlenecks in institutions and innovation systems, markets, natural resource status and others? 5. How can Action Sites be selected to complement sites chosen by other CRPs in the same Action Area in such a way as to lead to joint research activities that will result in comprehensive and tangible outputs and impacts? 6. Through which M&E, knowledge management and learning processes is analysis and priority setting best transitioned into a Global Synthesis leading to useful International Public Goods? 7. How is systematic improvement in current farm enterprises further enhanced by different livelihood strategies, natural resource investments and/or higher value for farm products? What opportunities and challenges does new farm enterprise bring for poverty alleviation, improved nutrition, natural resource integrity and market development? 8. What processes are required to apply our understanding of farmers circumstances and needs for wider benefit? How long does this process require before these goods become formalized and extended throughout and beyond the program s Action Areas and how may this process be accelerated? Several activities within SRT1 are required to address these research questions, some of which are strongly site-specific to the Action Areas. These activities include, but are not limited to: Compiling existing information on the Action Site systems status, gender, and policy/institutional information, Constructing household, community, innovation system and Action Sites typologies across all Action Areas as a basis for identifying opportunities towards enhanced system performance, Identifying common denominators underlying the facilitating role of policy and institutions towards increased wealth and natural resource integrity, Analyzing necessary conditions that make partnerships effective and efficient, Compiling the drivers of change affecting people s livelihood options across all Action Areas and ex-ante assessment of their potential relationship with the Action Site system s status within the timeframe of the program, Developing and utilizing priority setting tools and procedures towards the identification of Action Site-specific opportunities for enhancing system performance, and Merging existing and new partnerships towards implementing, evaluating, and disseminating priority opportunities. The role of policy and institutions affecting poverty and natural resource integrity will be assessed in both the Action Sites (SRT2 Research Outputs) and Action Areas (SRT3 Developmental Outcomes) through the Situation Analysis and the resulting M&E Framework and be considered within the Global Synthesis (Figure 2). Specific attention will be given to policies and institutions affecting collective action towards marketing and natural resource management, input and output value chains, agricultural service provision and capacity enhancement. The identification of major drivers affecting livelihoods at the Action Area level will be mapped using geo-referenced baseline data, available secondary information from IFPRI, national ministries, and development agencies and other available georeferenced databases including the Millennium Ecosystem Assessment ( Four partners, CIAT, ICRAF, IITA and ILRI will take the lead in these areas. Bioversity will strengthen the Situation Analysis of the status of agrobiodiversity and its role in production, nutrition, system risk and resilience. 33

39 Partnership analysis will be based on the evaluation of earlier and on-going initiatives, since existing partnership is a criterion for the identification of Action Areas, and will include lessons learned in terms of effectiveness and efficiency, gaps and need analysis, determination of eligibility criteria (including interest in co-support for Humidtropics activities), and identification of specific roles and responsibilities. More importantly they will be subject to partnership and communication research activities themselves, exploring types and purpose of partnership and important concepts of trust and transparency that may help define the investments needed in starting, sustaining and ending partnerships along the R4D pathway. IITA will take the lead in partnership analysis and designing necessary capacity building response. Based on a baseline understanding of farming systems, gender relations, and current policy and institutional support, opportunities towards enhanced system performance will be identified for all Action Sites. This course will be guided by the socio-ecological niche concept of Ojiem et al. (2006). The required partnership arrangements will be developed around these opportunities to ensure that both technical issues and a more enabling environment are considered. These must arise as early developments in the Situation Analysis so that activities under SRT2 may proceed. The full Situation Analysis, leading to a Global Synthesis, is led by all primary partners and intended to analyze issues beyond their site-specific dimensions (see Section 6). This synthesis contributes to the coherence of the program in terms of approach and studies aimed at enhancing the common vision within the Humidtropics. It will rely upon both qualitative and quantitative analytical approaches. The qualitative analysis will focus on the perspective, understanding and satisfaction of beneficiaries and systems actors. It will include participatory identification and evaluation of options and institutional arrangement for improving effectiveness of systems interventions. This will also include analysis of the communication roles and methods used for supporting change at the individual, network and innovation system level. Participatory evaluations will be aggregated for defining models for scaling up systems interventions. For example, the World Bank conducted a study at the global level on understanding poverty from the perspective of the poor themselves including their priorities, role of gender and how poor persons interact with institutions (Narayan, et al., 1999), we intend to undertake a similar approach. Meta-analysis requires that data collected in the Action Sites and Areas measure the effects of different interventions (see Section 13). Both econometric and systems modeling will be employed to further understand changes and factors that affect those changes. The relative magnitude and size of those effects will determine the most important ones that can be the focus for further policy and practice recommendations. Advances in the field of econometric modeling aimed specifically at evaluating treatment effects (Heckman, 2005) allow for very robust analysis of the effects of systems interventions including principal component analysis. Documenting and tracking adoption of the technologies introduced and tested in each site leads to better adjustment and replication of impact pathways and communication roles. Moreover, Global Synthesis expresses lessons useful to future projects and contributes to research tools and knowledge in the other CRPs (see Section 9). This synthesis will result in methods and models for assessing interventions on systems productivity and diversity, NR integrity, biodiversity and institutional effectiveness, including gender equity. This leads to a framework for defining extrapolation domains and targeting systems interventions beyond Action Sites; comparison of development trajectories, and identifying innovations and institutional policies supporting system s changes. The Global Synthesis will also define key system indicators for NR improvement and poverty reduction and develop models for assessing system interventions based upon these indicators. The roles of traditional and new methods of communication as extension and social media must also be considered. Special attention will be paid to institutional policies and practices that 34

40 support changes evaluated across Action Areas and strengthening capacities for scaling up complex interventions Research timeframe and outputs 1. Humidtropics Analytical Framework developed and tested for institutional and innovation stakeholders and for sustainable livelihoods (3 months, addressing SRT1 Research Questions 2 and 4). 2. Humidtropics Analytical Framework will identify Action Sites within all Action Areas of the program (4 months, addresses SRT1 Research Questions 1 and 5). 3. A combined baseline study, involving collection of gender disaggregated data and priority setting exercise will identify critical entry points for increasing systems productivity, NRM, and gender equity research by household typologies in these Action Sites (12 months) based on a preliminary identification of innovation network partners (addressing SRT 1Research Question 3). 4. Opportunities for developing self-sustaining input delivery and product marketing platforms that facilitate value addition and wealth creation at farmer level will be identified (9 months) and initiated shortly thereafter (12 months, addressing Research Question 7). Clusters of localized agri-businesses (service hubs) owned and managed by self-reliant and resilient small-scale farmers groups including women and young people emerge in Action Sites and lessons are learned on the organizational requirements for replicating such hubs across wider action areas and utilization (supporting SRT1 Research Question 3). 5. Institutions influencing poverty and the natural resource and agro- and biodiversity status will be assessed across all Action Areas and directed toward specific roles within Action Sites (15 months, addressing SRT1 Research Question 8). 6. Drivers of change affecting SLOs will be identified and their potential impact compared across Action Areas and the three Eco-Zones (18 months, addressing SRT1 Research Question 7 part 2). 7. Partnership arrangements to address integrated systems challenges that bridge Research Outputs and Developmental Outcomes, including strengthening of existing Innovation Platforms (where available) and establishment of new ones, will be formalized (24 months, SRT1 Research Questions 7 & 8 part 2). 8. The ongoing Situation Analysis, guided by continuous M&E will develop a Global Synthesis that identifies specific IPGs for the Phase 2 of the program (30 months). These IPGs will include detailed pathways directing systems integration across all Action Areas (addressing SRT1 Research Questions 6 and 8 part 2). 9. A detailed Phase 2 Humidtropics proposal will be developed from the Global Synthesis that includes a strategy to transition from partner-based to humid tropics system based operation beyond Year 3 (36 months) Realizing International Public Goods The CGIAR has set a goal of improving livelihood and reducing poverty through the interventions of Centers and their partners. At a global level, it is of interest to understand how R4D contributes to achieving these objectives. CRP1.2 will intervene through its achievements across the humid tropics of Africa, Asia and Americas. In all three regions, diverse, and in some cases common, issues related to poverty, ecosystem integrity, systems productivity, institutional effectiveness, natural resources management, and gender are affecting people s lives and livelihoods. Understanding these issues at the global level with a humid tropics perspective is therefore necessary to providing solutions that will not only apply to Action Sites but for scaling up and out program results to Action Areas and the humid tropics regions of the world. 35

41 The diversity in these research sites coupled with the differences in access to productivity enhancing technologies and natural resource endowments and degradation will provide a platform for understanding trajectories of changes. The M&E Framework that captures and classifies these trajectories will constitute the program s first IPG. Humidtropics seeks to redirect these trajectories (or their drivers) by prototyping flexible combinations of proven solutions, first tested in the Action Areas. Alternative approaches by household and community typologies that emerge from the Global Synthesis will produce the next generation of IPGs. The scene for such synthesis of IPGs is based to a larger extent on cross-site analyses of Action Areas strengthened by strategic case studies. In a systems-level intervention, tradeoffs among the different components is inevitable. It is essential to clearly attribute outcome to system components either individually or in groups. Untangling the effects of different system components is best performed at a higher level through meta-analysis across time and space. In addressing natural resource integrity, the Global Synthesis and IPGs resulting from it will guide diagnostics of natural resource degradation and provide site-specific responses to it. These alternative approaches will not be trivial or narrow in focus. Clearly, the design of the IPG packages must be built upon not only technological knowhow and its narrowly focused promotion to farmers, but rather upon proven approaches within different production systems and along the entire agricultural value chain at systems level. In this way, IPGs range from practical; guidelines directing land mangers response to production and natural resource management challenges to complex survey and analysis intended to partition and redirect drivers of change. 4.5 SRT2 Integrated Systems Improvement The overall goal of SRT2 is to identify technologies with a proven ability to offer different types of farm household opportunities for more sustainable and profitable farming, ultimately resulting in better lives in vibrant rural communities. In essence, SRT2 is going to work from the opportunities and critical entry points towards system intensification, identified in SRT1, through a R4D triangle comprising increases in system productivity, market integration, and improvement in the natural resource status differentiated by agro-ecological setting, household types and their respective trajectories. The unit of operation of SRT2 is the Action Site, allowing assessment of specific planned interventions within a well-described baseline farming condition. These interventions are based upon innovation within the R4D triangle (Figure 2) and involve interactions between production systems, natural resources and market availability. Action Sites consist of different sets of farms at a landscape scale, selected within a well-defined typology of resource and income levels, often compiled at a local administrative level. These farms are used for monitoring, on-farm research, and participatory testing. The Action Site also contain relatively controlled experimental sites, which may either be a representative farm or spread across an entire farming community according to local conditions and innovations to be tested. Action Sites are selected based on their varying positions on the poverty and NRM integrity axes of the Humidtropics Conceptual Model (see Sections 3 and 4.3) and intensification strategies identified within specific population and market conditions encompassing the variation within its constituent Action Area. Specific entry points will be directed towards the missing links to move different farm household types along an intensification continuum targeting productivity, market linkages, and natural resource interventions or most often a combination of the three. Two important aspects of SRT2 are related to comparisons across Action Sites. First, insights are reached by crosscutting its three component sub-srts in relation to most successful interventions across existing gradients and variability within farming systems. Some successful interventions will prove to be very site-specific while others are more 36

42 robust and adaptable. Another aspect is related to tradeoff analysis to ensure that systems are intensified based on principles of agro-ecological sustainability and in a way that s equitable to various farm typologies and stakeholder groups. The following suite of SRT2 Research Questions will be addressed through SRT2 activities within the overall research context of the program. Not all questions can be tested at every Action Area but planning will guarantee that all are addressed in a comprehensive fashion. These questions may also assist in structuring the SRT1 Situation Analysis as it relates to activities within Action Sites (see Section 4.4). These research questions are overarching as they relate to all three components of the R4D research triangle; production systems, markets and natural resource integrity, and their key interactions. SRT2 Research Question 1. How robust are critical entry points towards intensification in the context of existing variability in household resource endowment, agro-biodiversity use, soil fertility status and market access, and how do these entry points vary with gender roles? Rationale. Densely populated rural areas are very heterogeneous, both in terms of access to production factors for farming families within a community (with female-headed households most often belonging to the relatively poor typologies) and in terms of status of fields within farms and farms within landscapes. Lower resource endowment has been associated with a reduced ability to take risk in a manner that attenuates yield potential. Entry points related to intensification require investments of farmer s labor, cash, and land, and the outcomes of investment vary with natural resource conditions. Consequently, the effectiveness of these entry points is likely to vary across farmer typologies and natural resource gradients and this needs to be understood in order to identify robust interventions that embody least risk and optimal benefits. SRT2 Research Question 2. What is the impact of population density, land availability, initial natural resource condition, agro-biodiversity status and market access on the management of trees, crops, livestock and their interactions in mixed systems? How may system productivity, natural resource integrity, farmer income and household wellbeing best be enhanced simultaneously and how do these interventions differ by gender Rationale. Here this research question is discussed from the perspective of integrating livestock management but within the program s similar exploration of the impacts from perennial and annual crops. Livestock management systems can range from free-ranging to zero-grazing systems. In zero-grazing, feeds are produced as cut fodder and crop residues, and shortfalls must be purchased. Trees offer an important alternative feed source and provide shade for livestock. These sub-systems result in varying impacts on livestock productivity and natural resource integrity while access to markets determines their incomegeneration potential that in turn influences farmers to re-invest in trees and livestock. Beyond a certain population density, where land resources become too scarce, producing sufficient fodder within farm becomes less realistic, as in Rwanda, notwithstanding its equitably-based policy of one family one cow. In more sparsely populated areas, freegrazing leads to degradation as stocking rates exceed a critical limit. Appropriate conditions for intensifying tree-livestock systems need to be better understood to ensure that productivity and profit remains aligned with natural resource integrity. An expansion of this research question is How, and under what circumstances, can tree intensification along agro-ecological principles contribute to the productivity and sustainability of livestock production sub-systems?. SRT 2 Research Question Set 3. What is the role of trees and forest remnants in agroecological intensification? To what extent can and should intensification of tree-crops be 37

43 promoted among smallholders in response to climate change? What are the short- and longer-term consequences of tree-crop intensification for farm livelihoods? Rationale. Incentives for investment in trees by smallholders vary. In West Africa, restoration of cocoa production demonstrated potential to rehabilitate exhausted soils. At the same time, economic returns from rubber production are also increasing, driven by international demand outstripping supply. Farmers have little guidance as to which is more advantageous, the rejuvenation of existing stands or diversification into new tree-crops, and how does this impact household benefits and resource management over time? Can agroecological principles be applied in agroforestry systems to sustain more intensive production without negative consequences for the environment or leading to premature yield declines? Indeed, what are the comparative opportunities and household benefits for intensification and diversification of tree-crops through intercropping? In the East African Highlands, suitability of land for coffee, and the spread of pest and disease are both very sensitive to climate change. To what extent can management of coffee agroforestry buffer impacts of climate change and maintain key ecosystem services including pollination and pest and disease control? Can appropriate social capital be developed to allow smallholder coffee producers to reap benefits from higher value and certified market prices, and how will this affect carbon stocks? Clearly, the area of tree-crops, their relation to other farm subsystems and their role in resource management open many promising, and practical, avenues for integrative research. SRT2 Research Question 4. Are productivity and market-related interventions and their expected contributions to income and food and nutritional security necessary to allow land users to invest in measures improving their agricultural natural resource base? Rationale. Basically, will farmers invest in resource conservation directly, or is this best arranged as an indirect benefit of farm enterprise? Should interventions necessarily be crafted in a manner that provide immediate benefits to farmers but also assure that less tangible value in the resource base also be accrued over time? Early answers are mixed. Farmer interest in dual purpose soybean appears governed by the value of grains to the household for consumption and sale, not the supply of additional residual N as farmers prefer faster maturing, larger seeded varieties over those with greater symbiotic potential. On mid-altitude hillsides, the installation of soil and water conservation structures is widely recommended and farmers clearly recognize the threat of soil erosion but investments in terrace construction or reintroduction of trees on slopes has been minimal, and more often driven by subsidized, community-based programs. Fodder hedges that reduce soil erosion appear attractive to smallholders because they supply high quality livestock feed but are not necessarily positioned on the more vulnerable slopes. These perspectives suggest that immediate benefits to farmers are a near prerequisite for investments in NRM to be effectively scaled and integrative means to do so are required. SRT2 Research Question 5. How does the process of production systems diversification respond to market conditions, particularly the price differential between sale and purchase of foods, and what interventions are necessary to assure benefits resulting from stronger market orientation are fairly distributed across households and their members? Rationale. Markets are a key driver of farming system intensification and markets have correspondingly become more dynamic. Past policies relating to structural adjustment and market liberalization, together with more recent globalization of food markets and prices, have stimulated these changes as well. Nevertheless, much of the rainfed agricultural regions of the humid and sub-humid tropics face a range of constraints in accessing free and fair markets. A central understanding within the program is that market conditions determine the options and constraints on sustainable intensification, particularly the ability to balance subsistence and income objectives. This consideration in turn influences the 38

44 options for investment in farm inputs and the natural resource base, and the potential for plant and animal diversification within the farming system. More critically, changing market conditions create more options for poverty alleviation but at the same time lead to over exploitation of the natural resource base or increased risk that poorer households are less able to respond to improved market conditions, perpetuating cycles of poverty. SRT2 Research Question 6. How can improved household nutrition be directly stimulated within production system diversification and value added processing, and be accommodated within interventions targeting expanded rural markets and renewed resource conservation. Rationale. Pioneers into the humid tropics general enjoyed healthy diets based on foraged and cultivated plants and hunted and domesticated animals. Agro-ecological succession, where larger numbers of persons farmed smaller parcels of degraded lands, however, resulted in progressively poorer, carbohydrate-rich diets, particularly in terms of protein, mineral and vitamin availability. Mechanisms to improve diet through farm diversification include the production of traditional and market vegetables, greater reliance upon grain legumes, and more regularly consuming forest and animal products. Means of raising the mineral content of foods through blended fertilizers containing micronutrients are also available for use in areas with infertile soils. But improved nutrition also has a strong educational and institutional component because opportunity alone does not guarantee that families will consume better balanced diets. Even monitoring diets requires much closer contact with households than does characterizing and guiding their farming practice. Within the program, food processing must not only address markets but also household needs in a manner that may influence gender roles and differences. Ideally, improved nutrition will be embedded into interventions involving production systems and markets, but realistically it may prove necessary to commission specific interventions targeting food preparation practices and consumption patterns among the program s poorest households Smallholder farming dimensions Within each Action Site, several dimensions of smallholder farming systems are considered: Smallholder systems are diverse. Within a certain location, households differ in their resource endowment, production and consumption decisions, livelihood strategies, and long-term aspirations. Even in areas where a large majority of households can be considered to be resource-poor, differences in livelihood strategies between households may be decisive for their ability to adopt improved technologies and practices. The diversity of rural livelihood systems may be captured through typologies, which may be structural when clustering farm types that differ in their level of resource endowment or functional when they explain the determinants of each observed type, which in turn will be essential to understand the potential for an intervention s extrapolation and scaling-up through modeling. Smallholder systems are spatially heterogeneous. Differences in soil quality within a single farm may be as wide as between agro-ecological zones. Next to inherent variability of soil types in the landscape, often based on landscape position, management decisions on the allocation of (scarce) resources generates gradients of soil fertility within individual farms. Often animal manure and composted crop residues are added to the fields near the homestead, creating zones of higher carbon and nutrient concentrations within the farms. In undulating landscapes, the fields that are farther from the homestead are often located on steeper slopes. Due to soil heterogeneity, the performance of agricultural technologies fluctuate from success to failure across the fields of a single farm, and this needs to be better understood to improve technology adoption and innovation systems as a whole. To have a significant impact on natural resource management, it is necessary to have a suite of 39

45 interventions that collectively accumulate at the landscape level to improve environmental services, while at farm-level they will contribute to improving household income and livelihoods. Smallholder systems are dynamic. Farming systems experience changes in their configuration and functioning with time and their capacity of adaptation through human activity differentiates them from natural systems. To understand the dynamics of a system, it is necessary to consider the dynamics of the supra- and subsystems, that is, its context and internal components, by examining processes operating at immediately higher and lower scales, in this case the field, the farm and the landscape. The sustainability of farming systems depends largely on their capacity to adapt to changes at all scales. The contribution of agricultural technologies to the sustainability of smallholder systems, and their feasibility, should be evaluated in the long-term considering dynamic aspects of farming systems and their context. Moreover, rural families undergo different phases along a farm developmental cycle (Figure 9), which includes establishment, maturity and dissolution, along which household resources and objectives vary accordingly, often in accordance to changing market and economic opportunities. Resources Maturity Expanding family & resources Establishment and growth Maintaining & reproducing resources; production may exceed consumption Decline and dissolution Sub-dividing land Time (life cycle) Figure 9. Farm developmental cycle (Forbes, cited by Crowley, 1997 The impact on productivity, profitability, and natural resource integrity of specific technologies and interventions will be affected by Action Site variation. System modifications and successful interventions are those that deliver agro-ecological intensification outcomes for specific and well-understood biophysical, social, and economic conditions. Ojiem et al. (2006) proposed using biophysical and socio-economic constraints and opportunities to define socio-ecological niches within which compatible can be defined. SRT2 aims at identifying such alternative interventions along specific intensification stages, as affected by population and market conditions, local innovation networks and policy and institutional configurations Methodology SRT1 will define critical entry points based on a detailed characterization of the Action Areas and Action Sites (where opportunities will be defined) and these will be translated into specific interventions, implemented through the R4D Innovation Networks operating at the Action Site level (Figure 9). These interventions will be implemented across ranges of farmer resource endowments, within-farm soil fertility gradients, soil types within 40

46 landscapes, and other determinants existing within specific Action Sites. The performance of these interventions will be monitored in terms of productivity, income, household utilization and natural resource condition through the M&E framework, between farming units within communities and landscapes, and across various stakeholder groups (see Section 13). This approach provides a context that delivers the overall Humidtropics outcomes and simultaneously stream interventions that can then be integrated in the next generation of activities within the R4D Innovation networks. Subsequent refined interventions will deliver greater impacts at the farm and Action Site level as innovation skills, both by scientists and farmers, improve. As new opportunities emerge from SRT2, they become formalized within R4D Innovation Networks that test and promote the first increments of farmer adoption and adaptation, and link to the larger developmental research within SRT Farmer typology research outputs 1. A clear understanding of the productivity, profitability, and natural resource integrity impacts of specific interventions responding to critical entry points for specific farm typologies within and across Action Sites (addressing Research Question 1). 2. Tools to integrate the outputs from the M&E Framework towards the identification of innovation processes and useful interventions within and across Action Sites (addressing SRT2 Research Questions 1 and 2). 3. A suite of proven interventions targeting specific farm typologies, soil types within landscapes, within-farm gradients, levels of social capital, and other factors determining their outputs, to better direct scalability within SRT3 (addressing SRT2 Research Question 1). 4. New technologies for value addition, higher crop productivity, waste management, energy saving and gender empowerment developed in the course of responding to critical entry points identified in SRT1 within and across the Action sites (addressing SRT2 Research Question 2) Tradeoff and synergy analysis Figure 10. The logic used by SRT2 to move from critical entry points to best-fit interventions. An important dimension of SRT2 activities is analysis of trade-offs and synergies. The design of systems or multifunctional landscapes aims at fulfilling various objectives simultaneously. Often the impact of an intervention on two or more objectives may be in conflict or compete where one objective is enhanced while another is negatively impacted, in other cases interventions may enhance one objective without reducing others, or there may be synergy where two or more objectives are positively impacted at the same time. Tradeoffs may exhibit cases of strong substitution, in which fulfilling an objective renders a second objective impractical or cases in which complementarities or compromise solutions are possible (Figure 11). The agro-ecological intensification paradigm is based on the principle of simultaneous occurrence of positive contributions of specific interventions to 41

47 productivity, income, and natural resource integrity in ways that are equitable across gender, farm typologies, and for other stakeholders. This represents a very ambitious view of agro-ecological intensification, which extends beyond improved production based on an understanding of ecological processes into areas of economics and social equity. Absolute win situations are likely to be scarce so any intervention aiming at improving systems will need to pass the trade-off test to ensure that accrued benefits are superior to aggregated disadvantages. There are many examples where improvement in one farming dimension can have a negative impact on other dimensions. For instance, the increased sales of banana to urban markets in the Great Lakes region is exporting large amounts of plant nutrients from the farm while previously banana-based systems were sustained by tight internal nutrient cycles. Integration of a cash crop within a farming system can enhance farm income but negatively impact on food and nutritional security and the power relations within a household. Trade-offs can also have a temporal dimension, for example, tradeoffs can be made between today s productivity and long-term sustainability, or between shortterm food security and longer-term environmental impacts. A key aim of SRT2 is to find those combinations of interventions that improve the ratio of benefits to unwanted consequences and to recognize and avoid repercussions leading to system properties falling below a critical threshold. In practical terms this may often lead to the identification of necessary combinations of practices, for example proportioning fertilizer applications to banana sales. Figure 11. Schematic representation of possible trade-offs between two objectives. A change of a certain magnitude in the value achieved for objective A (ΔA, from A0 to A1) may imply a relatively large (ΔB - substitution) or a relatively small (ΔB complementarity) sacrifice in the value achieved for objective B, respectively. Source: Tittonell (2012). An integral part of the tradeoff analysis framework is an assessment of decision-making related to farm resources and the expected benefits from their utilization. Tradeoffs are central to rational decision making as resources become scarce and are prioritized differently among household typologies. As stakeholder s interests differ, resource management conflicts emerge and conflict resolution among stakeholders becomes an important outcome of tradeoff analysis, and needs to be fully embedded within the planned R4D platforms. 42

48 4.5.5 Methodology Tradeoff analysis begins with identifying the various decision makers, their objectives, and the value they attach to these objectives. Trade-off analysis will primarily focus on the farm level but may include tradeoffs-offs within the value chain as well, as when increased market efficiency is achieved by reducing transactions (and eliminating their actors) within the chain. Various ex-ante and ex-post methods will be used to perform tradeoff analysis. Ex-ante tradeoffs are analyzed quantitatively through evaluation of the consequences of specific interventions under different modeling scenarios, for example farmer resource endowment versus market access. Models used for this purpose vary in complexity. Multiple Goal Optimization Models, also referred to as Bio-economic Models (Kruseman, 2000; Janssen and van Ittersum, 2007) effectively analyze tradeoffs. This approach has an economic bias and the limitation of being static because models run for a single season and thus are unable to capture temporal variability and feedback. Inverse modeling techniques, and related global search algorithms, optimize a number of objectives by running dynamic models with a huge number of different combinations of parameters that represent farmer management decisions. Multi-Criteria Decision-Making techniques estimate solutions and partition conflicting objectives and various actors points of view (Romero and Rehman, 2003). These approaches will be incorporated into our tradeoff analyses. Farmers livelihoods involve conflicting objectives, which are often influenced by other community members and sectors. The utilization of natural resources usually involves negotiation between other farmers, local authorities and policymakers, each operating from their own perspectives. The methods developed by Gonzalez-Pachon and Romero (2007; 2011), allow including at the same time individuals (farmers) and collective (communities or authorities) preferences to draw trade-off curves between objectives. Ex-post tradeoff analysis will be implemented by looking at the M&E outcomes on productivity, income, and natural resource integrity related to various interventions. These results will be interpreted in relation to the overall objectives of the interested stakeholder groups, as a basis for prioritizing future interventions. Models will be validated and farmers decision making further explored using multi-agent modeling approaches and role-playing methods (Barreteau et al., 2001). An example of this approach is the trade-offs in farmer management of tree cover in coffee plantations in the Western Ghats of India available over the internet (see Capacities in the application of tradeoff analysis will be enhanced through training and mechanisms for incorporating them into partner institutes developed (see Section 6) Research Outputs from Trade-off Analysis 1. An improved system-modeling framework that allows for integrated assessment of the impacts of specific interventions on productivity, profitability, and natural resource integrity under different biophysical and community settings within and across Action Sites (addressing SRT2 Research Question 2). 2. A set of tools for trade-off analysis between various indicators related to specific interventions at various scales based upon the specific objectives of different stakeholder groups (addressing all SRT2 Research Questions). 3. An understanding of necessary conditions for the occurrence of triple win scenarios, where farm productivity, profitability, natural resource integrity all improve, across various farm typologies and landscapes and a framework for focusing research on reducing trade-offs and increasing complementarities through agro-ecological intensification (addressing SRT2 Research Questions 2 and 3). 4. Tools for strengthening adaptive capacity of the stakeholder groups involved in the R4D platforms to participate in tradeoff analysis and conflict resolution (supporting SRT3). 43

49 4.6 SRT2.1 Integrating Markets through System Intensification A central concept within the development of Humidtropics is that the pathway along which production systems intensification is directly related to the structure, conduct and performance of markets. Market development trajectories in relation to smallholder farming vary tremendously as large areas are locked into market stagnation resulting from prolonged cycles of poverty; other areas are focused upon relatively few, unfairly traded commodities; and yet others enjoy the benefits rural transformation where local markets and buyers, commodity trading associations addressing internal and export markets, and local processing opportunities provide a host of marketing options; and all cases in between. Obviously as a program we strive toward the latter condition but recognize that in the least developed settings first-steps must be taken based on relatively few new marketing opportunities. Markets as an area for research and institutional innovation are relatively recent and largely developed as a response to policy adjustment and market liberalization. The structure of markets in the humid and sub-humid tropics can often be characterized as having high costs of transport, high transactions costs of market participation, inefficient bulking through too many transactions along the supply chain, transactions on a cash basis that undermine emerging financial institutions, large differentials between the farm gate sales to middlemen compared to the buyers purchase price, weak market intelligence among producers, and irregular price volatility. This set of interacting structural characteristics result in low confidence among producers that in turn create disincentives toward the adoption of needed technologies. This dynamic interaction between market structure, conduct and performance and technical change has resulted in markets being a critical field for research and intervention within CRP1.2, regarded with the same importance as systems productivity and NRM. The area of smallholder s market condition and stimulation is a research area with limited analytical methods as these market contexts violate many basic assumptions of neoclassical economics. What has substituted for theory are descriptive analytical constructs, especially value chains, or the development of empirical models of market agent behavior. Value chain analysis is not deterministic and their use requires some level of judgment by the analyst. Moreover, markets, like ecosystems or landscapes, are relatively fuzzy concepts with no clear boundaries and yet how they function within a given area and time is a key element in characterizing development potential. As a relatively new research area, agricultural market analysis under smallholder conditions involves interplay of new research methods, piloting of institutional interventions, and the development of new tools to monitor the performance of relatively disorganized markets in underdeveloped areas. This research requires survey instruments designed around the joint collection of data on farming systems and prevailing market structures, starting with the Situation Analysis of SRT1 and carrying through systematic data collection in the Action Sites within SRT2. At the Action Site level, this research requires that engagement of all parties involved in market development including private business, local authorities and non-governmental agencies. The nature of the research problem and its current state of development suggests a challenging scope for innovative research addressing markets and their interaction with farming system intensification and NRM. SRT2.1 will focus on three different but related research areas. Firstly, it will explore the role of market structure on system intensification that may lead either toward diversification or to specialization, and farmer investment in the natural resource base. The focus here is on understanding how changes in markets, whether autonomous or piloted, influence farmer decision making in regards to shifts in production structure, changes in management, investment in soil and water recapitalization, or adoption of new technologies. Research at this level will also be critical for defining critical entry points on which production system 44

50 research will be based. Secondly, SRT2.1 will evaluate the role of markets in scaling up production system findings and how market innovations such as credit or insurance programs, further stimulate spontaneous diffusion of improved techniques. Expanded research questions must be framed and addressed at the Action Site level. For example, do interventions promoting warehouse receipt systems or dairy collection points lead to intensification of their respective commodities? Thirdly, SRT2.1 will assess the impact of improved market participation at the community level on poverty alleviation and gender equity. Market development provides the most powerful vehicle for poverty alleviation but at the same time those farms with higher levels of human capital and with more resources are better able to participate. Ensuring equitable market participation may require either institutional innovations or, as in the high population density areas, participation in other income generating opportunities. Possibly even more difficult has been ensuring gender equity under increased market participation. Women s crops tend to be selected basic staples, vegetables and small domestic animals. Women also do most of the on-farm processing. Defining leverage points that guarantee women s participation and control over farm assets and income generating activities remains both a principal research area as well as an area for piloting institutional innovations. Finally, we must weigh the comparative advantage of interventions that target different types of markets including specialty niches, local trading associations, urban areas, immediate cross-boundary trade and commodities intended for domestic processing and export Methodology Given the knowledge gaps described above, the principal methods to be used in this strategic research theme will only be sketched. The methods will be expanded in four critical areas, namely; 1) survey protocols and data structures for market performance integrated with farming system surveys, 2) defining critical production system entry points based on market potential, 3) piloting and evaluating market innovations, and 4) developing a monitoring system and indicators to evaluate progress towards specified targets. Methods for work areas 1 and 4 must be coordinated program-wide across Action Areas, while those for work areas 2 and 3 will be crafted to the specifics of each Action Area and their component Action Sites. In both cases, research protocol design is coordinated with the SRT1 M&E Unit so that results are accommodated within program data services. As with much of the research in CRP1.2, a balance will need to be sought between needs for centralized coordination and the delegation of responsibility and creative approaches to the decentralized research teams in the individual Action Areas. Indeed, many market advances will be localized and result from iterative problem solving but these too must be reported to the larger program. The design of an efficient survey protocol for markets integrated into a larger Action Site characterization structure requires a significant amount of planning. This involves multiple commodities with complex supply chains to principal markets and where an evaluation of costs and prices at each transaction point is essential to evaluating market efficiency. Sampling of market agents, being able to repeat interview market agents, number of supply chains to be selected, price data collection with standardized weight conversion equivalents, and efficient monitoring within the entire Action Site (and later extended across Action Areas by SRT3) pose issues relating to the cost and potential value of the data collected. There are too few examples of a panel survey of market supply chains, much less one linked to smallholders, and expertise is required to this end. Budget constraints will also dictate the choices to be made in survey design. Identifying entry or leverage points for system intensification requires understanding market demand, expected farm gate price, and strategies for assembly and bulking. Such analysis is then integrated with more traditional farm management models to define activity sets that offer the best potential for initiating an intensification pathway for representative farm types. 45

51 Piloting institutional innovations to improve market access and efficiency, such as warehouse receipt systems, value adding processing, or interlinked microcredit, insurance schemes and input supply, are important options available for research within Action Sites and accompanying monitoring and reporting mechanisms must be developed and interfaced with SRT1. Another consideration is the choice of control conditions as this arbitrarily (randomly) opens opportunities to some households and not to others. Randomized controls and specified counterfactual may be experimental requirements, but they are not necessarily fair and must be balanced with community relations. Reliance upon models of market interventions is one means to reduce this difficulty. The greater potential for market development is through commissioned pilot programs that improving market conditions under SRT3, particularly those with proven impacts on poverty and gender equity, and our research within SRT2 must remain on track toward that goal Research outputs of SRT2.1 Research outputs from SRT2.1 follow directly from the SRT2 Research Questions and the methodologies described above. 1. Integrated market and production system panel surveys conducted, compiled, analyzed and interpreted (addressing SRT2 Research Questions 4, 5 and others). These outputs will take the form of published survey results, the release of an integrated database with meta data for analysts researching agricultural change, published validation of several economic models and monitoring systems for improving market performance. 2. Research support to other program areas including SRT1, SRT2.2 and SRT3. This support includes validated models for use within the Global Synthesis of SRT1 and systems productivity research by SRT2.2; and proven market innovations for scaling up within SRT3. It also provides proven methods for implementing market innovations, packaged evaluation systems of marketing pilot activities, packaged approaches to scaling up market innovations and business and value chain analysis procedures appropriate to diagnose smallholder engagement. 3. Research products facilitating policy reform and infrastructure improvement will be released that relate to improving market intelligence and performance. These products will take the form of specific published recommendations to policymakers and local authorities and potential investors. This output is accompanied by a synthesis of business-related opportunities describing market trends useful for guiding public and private investment (addressing SRT2 Research Question 5 and supporting SRT3 and SLO 1). 4. New farm input products and small agri-businesses will be launched as a result of research findings and pilot programs. New inputs, such as specialized fertilizer blends or the design of more ergonomic farm tools, will be released as IPGs. New businesses will include local market collection points, local food processors and agents offering credit facilities and farm insurance (addressing SRT2 Research Questions 5 and 6 and SLO 1). 4.7 SRT2.2 Increasing System Productivity Systems productivity is viewed in the most holistic sense to include annual and perennial crops, remnant forests and other non-crop areas, livestock and their interrelationships at the farm and community levels. A core research question directing SRT2.2 concerns How can farm households mobilize farm technologies and ecosystem services into more productive, 46

52 profitable and efficient small enterprises to meet their livelihood aspirations?. Sustainable intensification is regarded as an overarching solution because more effective farmer decision making of useful combinations of tree, crop and animal technologies, combined in synchrony with ecological processes and proper management of the available resource base results in productive and sustainable agricultural systems, assuming the presence of market incentives, required farm inputs and supporting institutional arrangements (Pretty et al., 2011). The constraint to systems productivity periodically posed by soil moisture stress within the sub-humid tropics must also be considered as the full beneficial impact of best combinations of crop and animal technologies cannot be realized without effective water management. This research agenda is embedded within the intensification pathways for specific Action Areas and Sites (see Section 3) and based on the identified critical entry points towards intensification emerging from Situation Analysis (see SRT1). The specific components to deploy under this SRT will depend on the balance of food security improvement and income generation within household production objectives. They will also depend on the resource endowments of the farm households, and the services and skills within farmer associations and their extension providers participating in different R4D platforms. Specific entry points towards improving system productivity will be a combination of improved lines of crop and livestock; external inputs including organic inputs, fertilizer, and other soil amendments; better agronomic and pest and disease management; and better feeds and health services for livestock. Improved seed and higher seedling quality is a frequent entry point that must be accompanied by a suite of appropriate crop management practices. Improved germplasm most often involves targeted annual food crops, but also includes indigenous plants that are tolerant to local stress conditions, fodders, and multipurpose trees, including legumes contributing greater biological N fixation. It also includes numerous perennial and annual cash crops emerging as market opportunities arise. Livestock components can include improved breeds of species currently integrated in treecrop-livestock systems, providing incentives to advance along the livestock progression from better managed and smaller to larger livestock (McDermott et al., 2010). It also includes re-introduced plants that have disappeared from farming systems. Other key livestock components include improved feed options through crop residue processing, use of agro-industrial byproducts, planted forages and forage trees. Enhanced livestock health inputs will also be important including vaccines and drugs. Evidence shows that bundling of breed, feed and health improvements is key to sustainable livestock intensification. Organic inputs consist of recycled crop residues, farmyard manure, or organic resources produced through cut-and-carry systems, including agroforestry systems. Fertilizer and other soil amendments (e.g., lime, rock phosphate) will be deployed depending on their availability and anticipated profitability, thereby aiming at maximizing their use efficiency. Adapted agronomic practices can adjust plant densities and spacing patterns, adapt time of planting to explore compatibility between growth patterns of associated or rotated crops, target applications of inputs to specific phases within a crop rotation, or other practices that explore synergies between crops in mono-cropped or intercropped systems. In general, green manuring is considered as a stand-alone rotation but is better combined with need for forage and fodder. Improved pest management will exploit host resistance, agronomic as well as soil management practices, and explore bio-control and amendments, which will be more accurately targeted following improved pest and disease diagnosis. Improved soil management practices related to the use of best tillage and land preparation practices, water harvesting structures, drip irrigation on high-value horticultural crops, erosion control measures, or other practices that conserve soil and water and favor crop water and nutrient uptake will also be advanced. Crop management packaging will also draw on knowledge of agro-ecological opportunities for increasing the role of integrated control mechanisms. 47

53 These inputs and practices will be mainly derived from earlier experiences in the Action Areas and research outputs from the CRPs focusing on these specific components (see Section 9). The main difference with other CRPs and earlier investments is that SRT2.2 evaluates not only the impact of individual components on the productivity of the field production unit but also their interactions with the larger farming system. A good example is tree-crop-livestock interactions where introduction of improved cattle breeds can result in higher milk, meat, and manure production but where the need for more and higher quality feed is often a limiting factor to generate these benefits for farmers with insufficient land (Klapwijk, 2011). Besides tree-crop-livestock interactions in the moist savanna and midaltitude Action Areas, other interactions that will be considered in SRT1.2 are those between food security and tree cash crops in the humid lowlands, and between farms and farmer typologies within micro-watersheds. The Humidtropics uses the agro-ecological intensification paradigm as its basis for system intervention. Agro-ecological intensification is the mobilization of ecological processes to increase agricultural production and the productivity of external inputs, labor, natural resources, and to reduce losses to abiotic and biotic stress and the negative off-field impacts of production (Cote et al., 2010 and Dore et al., 2011). Better use of biological regulation mechanisms and agricultural biodiversity (Jackson et al., 2007) at different field, farm and landscape management levels is also postulated. An important component of agro-ecological intensification involves the use efficiency of inputs, labor, and capital as it relates to a G E M M1 framework. G E M M1 signifies Germplasm Environment Management Market interactions, as a guiding principle rather than a model. This framework acknowledges that 1) the performance of specific components will depend on the biophysical environment (e.g., temperature, rainfall, day length, and soil stresses such as acidity, aluminum toxicity, limiting nutrients) and its variability in time and variation spatially within specific Action Sites; 2) crop, soil and pest management, as affected by farmer resource endowment and understanding of ecological processes, is an important complement to assure that improved germplasm or crop biodiversity delivers its potential in a certain environment; 3) several components deployed together may interact, either positively and negatively, at the field, farm and landscape scales; 4) agro-ecological rules of thumb usefully guide the design and management to suit different household typologies and trajectories; 5) market opportunities influence farmer decisions, particularly choice of G and M. A good example subscribing to the principles of agro-ecological intensification is the Integrated Soil Fertility Management (ISFM) approach for enhancing crop productivity (Sanginga and Woomer, 2009). ISFM is a flexible approach to soil health management based on combinations of useful soil management principles, including some of the principles advocated by conservation agriculture and agroforestry, provided these make sense in a particular farming context. ISFM is defined as: A set of soil fertility management practices that necessarily include the use of fertilizer, organic inputs, and improved germplasm combined with the knowledge on how to adapt these practices to local conditions, aiming at maximizing agronomic use efficiency of the applied nutrients and improving crop productivity. All inputs need to be managed following sound agronomic principles (Figure 12) (Vanlauwe et al., 2011). The ISFM framework thus aims at maximizing the use efficiency of all production factors and acknowledges the need to foster interactions between individual soil management components. Many new biotechnologies involve the culture and release of beneficial soil organisms and these will be considered as well. 48

54 Figure 12. Conceptual relationship between the agronomic efficiency (AE) of fertilizers and organic resource and the implementation of various components of ISFM, culminating in complete ISFM. Soils that are responsive to NPK-based fertilizer (A) and those that are poor and less-responsive (B) are distinguished Methodology The first stages of the Situation Analysis of SRT1 will identify and characterize the most prominent farming systems, farmer typologies, and social networks within specific Action Areas, and provide entry points to intensification. Based on these first understanding of the Action Sites, and with attention to gender relations, policy environment and partnering institutions, and other potential drivers of change; opportunities towards enhanced system performance will be identified for all Action Sites. These critical entry points will for certain Action Sites have important productivity-enhancing components as detailed above. Even in Action Sites where investments in profitable market access may be the focus of entry points towards system improvement, productivity enhancement and natural resource integrity will nonetheless be included in work plans in compliance with the full suite of Research Questions and the overriding CRP SLOs. Within the context of the agro-ecological, infrastructural, and institutional boundaries of a specific Action Site and the identified critical entry points, specific interventions, adjusted for farmer typologies and landscape positions, will be identified through the following principles (see Box 2): Increase overall productivity and yield of crops, trees and animals, Increase capture of light, water and nutrients leading to greater conversion to useful biomass, Increase symbiotic nitrogen fixation through prominence of grain, vegetable and tree legumes Tighten nutrient cycling and reduce other forms of nutrient loss, resulting in greater retention of nutrients, Increase in effective suppression of weeds, pests and diseases through habitat management and Increase above- and belowground biodiversity to enhance system resilience. These options will be tested and evaluated with farmer associations or other social networks within the Action Sites through participatory action research using multi-locational designs, and following a demonstration-adaptation-adoption sequence, depending on previous experiences with specific interventions (Figure 13). 49

55 Figure 13. Description of the action research cycle, indicating the various steps and their geographical focus, including a potential step back to on-station activities. The rectangles on the right depict treatments tested at each step with Tx and Ty indicating promising treatments through formal and participatory evaluation and LP means local practice. The arrows to the right indicate the feedback cycles embedded in the approach where movement between the various steps is facilitated. This approach will be compared to, and possibly replaced in time with a more streamlined development of simple prioritization tools advanced through expert consultation through R4D platforms. Demonstration trials consist of a number of best-bet component technologies that have been derived from earlier research or from experiences in other areas under similar agro-ecological conditions. Observations recorded from demonstration trials include agronomic and economic performance of the tested technologies (against the local practice) and gender-segregated participatory evaluation of the technologies. Based on the performance of the tested technologies at the demonstration sites, a few of the most promising technologies are taken forward into adaptation trials, consisting of two or three technologies and a local practice. These adaptation trials are numerous (e.g per Action Site per season). These trials are can eventually be accompanied by soil and plant sampling on a sub-sample of sites to have a better understanding of plant-soil relationships underlying the performance of a specific technology. After participating households have tested and evaluated these technologies against local practice, an indicative number of them will adjust their practices accordingly. This number will vary and depend on the nature of the technology and the creation of enabling conditions for technology uptake but is generally proportionate to a technology s adoptability. This sort of research cycle is accompanied by M&E focusing on key indicators of yield, profitability, and soil conservation, quantifying tradeoffs and assisting in designing future interventions. Specific indicators related to system performance for different farmer typologies will also be included. The large number of simple trials permits interpretation in terms of position within farms and landscapes, which is in turn related to risk assessment. These activities are also used to initialize and later validate system models operating at farm and landscape level. These tools are used for the identification of subsequent generations of entry points towards 50

56 system productivity and natural resource integrity enhancement through ex-ante methods and approaches, taking into account tradeoffs between outcomes, between households within an Action Site, and between the various stakeholders interests. This approach is admittedly geared to annual crops and soil management but similar pragmatic, robust approaches toward livestock and perennial crops will be developed. Box 2. Translation of a critical entry point in a set of simultaneously applied components towards system improvement in the CIALCA region. The Consortium for Improving Agriculture-based Livelihoods in Central Africa (CIALCA), coordinated by IITA, Bioversity, and CIAT-TSBF, is active in the Great Lakes Region since 2006 with activities in DR Congo, Rwanda, and Burundi and related activities in Uganda and Kenya ( CIALCA focuses its R4D activities around banana- and legumebased systems that cover a major proportion of agricultural land in that region. Through baseline surveys and rural appraisal studies, it was evident that food insecurity and child malnutrition were major issues for smallholder farming households in Sud-Kivu Province of DR Congo, one of the CIALCA Mandate Areas. The major staple crop is cassava but yields were severely affected by the cassava mosaic virus. The productivity of beans, one of the major sources of protein, was low as a result widespread Photograph 1: Improved legume-cassava intercropping, with cassava at 2m x 0.5m, intercropped with 4 legume rows in the 1st season (left) and 2 legume rows during the 2nd season (right). This system can produce an extra 1 ton/ha of legumes without affecting cassava yields. pests and disease, poor nutrient management, and sub-optimal agronomic management. Taking these factors into consideration, together with farmer associations, CIALCA installed on-farm field trials and demonstrations that improved productivity of cassava-based systems through 1) combining improved cassava and bean varieties previously identified through participatory selection, 2i) adapted agronomic practices through row planting that adjusting the spacing between cassava plants and legume intercrops, and 3) targeted small doses of fertilizer in combination with organic inputs (Photograph 1). This combined management strategy resulted in 30-50% higher cassava yields and a doubling of legume yields. Such improved performance is sufficient to overcome food insecurity and malnutrition except that poorest households have less ability to access mineral fertilizer and mobilize organic resources, stressing the importance of farm typologies in targeting technologies Research outputs of SRT2.2 There are five research outputs resulting from activities in SRT2.2 Increasing System Productivity, with each described in detail and related to SRT2 Research Questions as follows: 1. Parameterized system models at the farm and landscape level to allow identification of best entry points towards system intensification and productivity enhancement for applicable farm typologies (addressing SRT2 Research Question 1). Identification and modification of existing models: Existing models that consider systems at the farm and landscape level will be used to evaluate ex-ante trade-offs 51

57 involving prioritized technology components on farm productivity such as biomass production and nutrient redistribution. Such models include the Africa NUANCES model that operates at the farm scale, includes crop-livestock interactions, assesses the impact of interventions on soil C stocks, and integrates resource endowment as a set of independent variables. Other models include the IMPACT (Integrated Modeling Platform for Animal-Crop systems), developed by Herrero (2007), landscape models and learning organizations (O Farrell and Anderson, 2010), or ecological network analysis (Rufino et al., 2009). Emerging models of the role of agro-biodiversity will also be identified due to its increasing importance in resilience for climate change. Parameterization and validation of models within and between Action Sites: To allow cross-site comparisons and cover the impact of variation within Action Sites, models will need to be sufficiently robust and inclusive of various farm types. This requires substantial parameterization activities using information gathered under SRT1. Information collected during implementation of SRT2.2 activities will be used to continuously fine-tune the models. The first aim of such models will be to integrate and explore potential options towards intensification rather than to predict outcome. 2. A clear understanding of the potential to enhance the productivity of system components and their interactions within a variable and heterogeneous production environment along the G E M M1 framework in the short and long term (Addressing SRT2 Research Question 2). Performance of system components: promising interventions are identified and tested among participating farmer organizations and other stakeholders through action research, coupled with a standardized M&E framework, covering aspects of productivity, resource use efficiency, profitability, and natural resource integrity. This approach allows the estimation of likely system responses in a manner that assesses both gains and risk. Quantification of interactions among system components: Agro-ecological intensification aims at optimizing resource use efficiency, partly by integrating various system components and fostering their positive interactions. Incomplete factorial designs following the G x E x M logic will be implemented within targeted landscapes in sufficient numbers to allow a analysis of the interactions between system components. Evaluation of the medium to long term benefits of system components: Most agronomic research is interested in immediate benefits of improved system components but some interventions require years of investments with potentially negative returns before their benefits are accrued. Systems-level livestock studies are similar. Within the set of activities implemented at farm and landscape scale, some will be turned into long-term field studies aiming at monitoring the accrued benefits, their resilience to stresses and their impact on the natural resource status. 3. A set of useful-fit component technologies for maximizing the use efficiency of production factors, as affected by with within-farm soil fertility gradients, landscape position, and farmer resource endowment with a specific focus on household types including gender role differences (also addressing SRT2 Research Question 2). Assessment of the usefulness of specific technologies: Interventions will be assessed based on an integral understanding of the conditions under which they deliver on their productivity and natural resource integrity outcomes. These will have necessarily passed the test of farmer and stakeholder evaluation over several seasons. Interventions not only include their component technologies, but also their adaptability to suit individual circumstances at the farm level. Development of gender and typology-specific component technologies: Female-headed households are usually more risk averse and many agricultural opportunities may bypass them. All component technologies and interventions will be assessed from a gender perspective. 4. Decision support tools become available to assess system productivity enhancement based on local diagnosis of system constraints, farmer s 52

58 production objectives and household interests. Develop and deliver diagnostic tools for assessing system constraints: Various methods will be developed to identify constraints to enhanced productivity, from local diagnostic tools to complex and formal laboratory-based analyses. Although the latter methods are important to facilitate an initial understanding of constraints, the final tools needs to be based more on efficient, low cost indicators that can be easily mastered by all stakeholder groups within the R4D innovation networks (see Section 6). Precautions will be made to assure that some tools are extensive in nature and able to accommodate scalability within SRT3. Participatory approaches will be used through the R4D platforms to deliver specific tools for decisionmaking and conflict-resolution. 5. An approach for monitoring the performance of promising component technologies and their combinations as these are adopted within R4D platforms and lead into SRT3. Assessment of system productivity at scale: Measuring system productivity is not necessarily but difficult process and standard approaches, commonly used in agronomic trials, cannot be readily scaled up to tens of thousands of households. Assessment approaches based on single real-time observation, nondestructive measures, and rapid data entry are needed to assess whether benefits to interventions are realized at scale. SRT2.2 must calibrate these rapid methods with more detailed studies for use by SRT3. Assessment of the adaptation of component technologies: While specific component technologies are often demonstrated using standard implementation protocols at the level of demonstration trials, farmers usually adapt these to better fit within their specific conditions. It is important to observe such adaptations since they may constitute the difference between large-scale adoption and rejection by rural communities. Specific M&E tools, focusing on adaptation and feasibility of best-fit technologies, will be developed in collaboration with the SRT1 M&E Unit and implemented through SRT3 activities. 4.8 SRT2.3 Natural Resources Improvement The integrity of the natural resource base is a crucial aspect of sustainable intensification, relating especially to the efficiency of resource use for agricultural production and the maintenance of ecosystem services in supporting production. In this respect, natural resource management is important for all the intensification trajectories described earlier. NRM is considered within the context of renewable and conserved farm resources that contribute to systems productivity, with particular reference to nutrient recycling, soil and water conservation, surface and ground water quality, carbon stocks, greenhouse gas emissions, and fostering beneficial organisms. Natural resource integrity is in itself very much affected by agro-ecological and biophysical conditions and the supply of its services has limits within a specific context (Box 3). Box 3. An example of NR integrity. Natural grassland on poor sandy soil may not have much agricultural value as it is low in nutrients, does not retain water, and supports vegetation of limited potential. However, the NR integrity may well be close to 100% as it is intact and delivers the ecosystem services possible under these conditions. On the contrary a primary forest thinned to establish a cocoa plantation may have been stripped of half of its biodiversity, a large portion of the standing biomass, and the soil biological, chemical, and physical quality may have dropped substantially to <50% of its initial status. Yet, this system may still deliver most of its ecosystem services relevant to man. The conversion of natural ecosystems to agriculture inevitably affects the natural resource functions related to soil, water, biodiversity, and carbon stocks. These changes will impact the ecosystem s ability to sustain its service functions for agricultural and non-agricultural land use. Changes in the natural resource status depend on the scale and type of 53

59 agricultural activities and the resilience of the ecosystem. In the humid tropics, slash-andburn is often considered the first and most dramatic step towards either repeated shifting cultivation or permanent agricultural use of land whereby the natural resource such as soil fertility, biodiversity, and carbon stocks first steeply and later on progressively decline (Figure 14). Despite this loss, however, natural lands converted to agriculture undergo an accompanying increase in agro-biodiversity, which declines with intensification and greater use of off-farm inputs, and increases as systems later diversify. The first task is defining the current NR status that serves as a baseline against which improvements are achieved. In many existing production systems, the NR status is extremely low compared with its natural condition. On the other hand, there are examples of human activity improving NR indicators, for example through introduction of perennial production systems for permanent or rotational use. Agro-ecological intensification as an approach is focused upon improving the condition and renewable performance of natural resources both in terms of contribution toward agricultural system productivity and overall system sustainability. Primary forest NR Integrity / status Slash and burn Long fallow Short fallow Permanent cultivation Intensification degradation Time Figure 14. Hypothetical progression of natural resource degradation over time. Smallholders reliance on agriculture often varies inversely with access to less disturbed natural habitat as household needs are met through gathering, hunting and fishing in areas away from the farm. In some cases, these activities, particularly fishing and extraction of forest products, are the primary source of income and farming is regarded as a means of subsistence food production. Thus, the natural resources of non-agricultural lands are used directly, compromising the NR base in a different way than conversion to agriculture. While it is generally accepted that agriculture strongly affects ecosystem functions and services, there is less known on how direct extraction of natural vegetation and animal products affects adjacent and future agricultural production. Toward this end, agricultural researchers require additional expertise from wildlife conservationists, foresters and other ecosystem experts, on how agricultural and non-agricultural ecosystem services are affected by disturbance, extraction or complete disappearance of natural resource components. Reliable indicators should be selected, allowing for affordable, quick, and robust measurements to monitor changes over time under different NR use conditions. These indicators are then compiled to quantify overall NR integrity (see Section 5). When nearby natural and cultivated land uses are compared, the latter signals the ability of the ecosystem to continue providing the ecological services that existed in its original state. 54

60 Research has yet to identify the extent to which the natural resource base can be consumed and what threshold should not be passed before its ability to deliver essential ecosystem services is significantly or irreversibly compromised. Different components are exhausted at different rates, but which are the most vulnerable and how may they be differentially conserved? Research is also needed on the extent that agro-ecological intensification can be achieved by the use of agro-biodiversity and agro-ecosystem based management. These approaches may impact on adjacent natural resources in a manner leading to more positive interaction between agricultural land use and adjacent natural areas. Humidtropics intends to identify hotspots where essential ecosystem services are at risk or where the natural resource base is no longer able to support critical services to agriculture and rural communities. Non-responsive soils, as detailed in Figure 14 are examples where critical soil-based ecosystem services may have been lost. Tools like the land degradation surveillance framework (LDSF) developed through the African Soils Information Service (AfSIS) will be used to identify these areas and candidate interventions for land restoration and management. Another important task of SRT2.3 is to understand how the increased flow of agricultural products from farms to markets contributes to soil depletion and what mitigation measures may be framed into interventions. This consideration is particularly relevant in terms of intensification pathways, comparing farm specialization and diversification, and further reflects on the complementarity of agricultural activities. For instance, the increased sales of banana to urban markets in the Great Lakes region is interrupting nutrient cycles as a growing proportion of produce is exported from the farm, while traditionally, banana-based systems were sustained by tight internal nutrient cycles. In contrast, on-farm processing of banana retains more nutrients because peels remain available for recycling. In addition, establishing leguminous trees and shrubs within banana lands replaces nitrogen loss through biological fixed nitrogen. The Resource-to-Consumption framework, developed by CIAT, extends the conventional production to consumption or commodity chain approach to include investment in natural resource management and other productive resources (social, human, financial, natural and physical) and strengthens forward and backward linkages between natural resource management, agricultural production and markets (Kaaria and Ashby, 2001). The Resource-to-Consumption approach relies upon participatory market research to assess into soil fertility management in a manner that supports enterprise development based upon realistic levels of increased nutrient removal as harvest products from the farm. Along similar lines, as systems intensify, livestock move from acting as supporting inputs to arable production, providing manure and traction, to being kept as commodities in their own right. Intensification of livestock production includes interactions with crops and trees as food sources but in its extreme has implications for environmental integrity. Key environmental implications of intensifying livestock production include increased greenhouse gas emissions, and degradation of vegetation resources. These aspects can be studied through lifecycle analysis to determine the effects of different intensification scenarios on key environmental variables. In many cases, livestock production intensification can reduce environmental burden by greater production per unit of feed and hence per unit of greenhouse gas emitted. Some interventions and technologies are able to both enhance agricultural productivity and sustain natural resources, and are currently being adopted, adapted, and disseminated. For example, in some coffee growing areas, shading through agroforestry has greater total system productivity with lower household risks than mono-cropped coffee (van Asten et al., 2011; Perfecto et al., 1996) and supports greater biodiversity (Perfecto et al., 2005) including some of the original forest cover (Philpott et al., 2008). In West Africa, it was shown that following forest clearance, the diversity and prevalence of mycorrhizal fungi declined with intensification of the system (Tchabi et al. 2008), however certain species 55

61 persisted better than others and proved beneficial to subsequent crop production (Affokpon et al., 2009; Tchabi et al., 2010). To date, most technologies are single component or modular, focusing on one or few aspects of the production system without consideration of the wider effects on natural resource condition. Many technologies related to crop management focus on soils without regard to larger resource management issues. In other cases, technologies were pioneered by farmers themselves (coffee shading) and documented from incomplete perspective. Barrett et al. (2002) demonstrated for a wide range of NR management practices, the paramount importance of the economic and risk-management requirements to implement such technologies. Hence, activities on NR-enhancing technologies have social dimension to ensure that the technological requirements can be met by smallholder endowments. A full scope of activities is necessary to address natural resource integrity at the watershed, landscape, and regional levels relating to conservation and restoration technologies, providing incentives to land managers, and establishing policies handling access and management of those resources. Specific consideration will be given to interventions that enhance diversification, system resilience, and system productivity. These win-win-win options are more likely to improve overall agro-ecosystem services across scales Methodology SRT2.3 has two major objectives: 1) maintain the natural resource condition of relatively productive agro-ecosystems, and 2) improve the natural resource status of degraded agroecosystems and their abilities to provide essential ecosystem services. In both cases, diagnosis is important. Ecosystem processes are complex and their interactions the target of strategic research in CRP5. However, in CRP1.2, we will use well-defined processes and easily determinable parameters as proxies for essential ecosystem services (see Section 5). Emphasis will also be placed upon tradeoff analysis relating increased systems output to their disturbance and consumption of natural resources and resultant impacts. Essential ecosystem services and the interactions between them will be defined at plot, farm, and a range of landscape scales, encompassing the dimensions where different ecosystem services become manifest (Box 4). Initially, the entry points and outcomes will be considered at the farm household level, however, disturbance and management of natural resources by individual smallholders can have profound cumulative effects at the community, landscape and higher levels (extending to global in terms atmospheric change and biodiversity loss). In the same way, maintenance of some ecosystem services critical for production and livelihoods such as maintenance of pollinators and protection of water quality operate at scales above the farm. Box 4. Examples of mapping ecosystem interactions across spatial scales. Erosion and erosion control is readily observed, causing major problems downstream through sedimentation or eutrophication. Erosion has consequences from the plot to watershed level. At some levels it is easy to quantify, yet at others it may be rather cumbersome and cost intensive. In addition, the consequences of erosion are site specific and of variable importance along the plot to watershed continuum. The contribution of agrobiodiversity, on the other hand, may be more difficult to see, but also has scale interactions, particularly for such functions as pest control, and pollination. A stepwise approach will first identify the ecosystem services required at the household and production system level. Within a farm, the hierarchical order of such services, and their interactions, will be determined by identifying how any component or practice affects production system components within the household or area of operations. Once essential ecosystem services within the household decision making domains are identified, a similar exercise will follow at community level with specific focus on interactions between farm 56

62 management and community-based resources. This will be of particular importance where communities have attained a certain level of division of labor with accordingly variable natural resource use requirements by different members of communities or there are existent incentives for resource protection. Work on NR integrity needs to identify processes and factors with known precision to allow reliable modeling in time and space. At scales above the farm level, particularly the community and watershed levels, GIS modeling may express proxies for key environmental services that can be calculated from readily available data. For example, parameters such as water flow calculated from digital elevation and land cover and habitat network connectivity for biodiversity protection have been effectively used to work across sites where intensive measurements to parameterize more sophisticated models are impractical (Eigenbrod et al., 2010; Pagella, 2012). Modeling is also important to extrapolate effects of system interventions in time, and to better understand tradeoffs between short-term production benefits on the one hand, and reduction of natural resource integrity on the other. Sustainability of natural resource use in crop production systems may be relatively easily determined through analysis of soil chemical, physical, and biological parameters once the rates of change and threshold values have been determined. However, it is unlikely that any individual parameter will serve as a reliable indicator of sustainable production or natural resource condition. To identify indicators for sustainable delivery of ecosystem services, higher-level processoriented measurements of the effects of soil quality, health or functions (such as biomass decomposition) can be used (Hauser et al., 2005). Monitoring specific ecosystem processes essential for crop, animal or tree production will then be developed and linked to agricultural yields, farm household well-being and parameters of NR status and integrity (see Section 5). In sub-humid parts of the Humidtropics domain, the integration of crop and livestock production are important, and with increasing market opportunities the spatial separation of these enterprises could present a threat to natural resource integrity if appropriate solutions are not sought. At the same time that natural resource status is being characterized and threshold values identified, it is necessary to conduct participatory research to determine farmers access to those resources, their patterns of resource use and perceptions of ecosystem services. This includes capturing the social and economic dimensions of natural resource use and integrity (van Asten et al., 2009). The latter allows the determination of resource use efficiency in relation to resource decline and current and potential investment in practices that maintain and restore those resources, often with corresponding increases in farm productivity. For example, the level of NR use efficiency is considered low in traditional slash-and-burn production systems as they consume large quantities of natural resources for relatively little crop yield. However, slash-and-burn systems are labor efficient and smallholders reasons to use them are largely related the time required for activities outside of farming (Box 5). As lands become more populated, a better balance between resource consumption and productivity must be achieved to prevent rapid destruction of remaining forests. Means must also be found to manage and conserve forest remnants. This sort of research will rely upon close collaboration with CRP5 in terms of methodology and tools (see Section 9). Household and community participatory evaluation approaches (Sayer and Campbell, 2004) will be used both to capture local knowledge (Sinclair and Walker, 1999) and perceptions of the current production systems and natural resource requirements (Vanclay et al., 2006). This activity serves to identify potential entry points for sustainable intensification technologies from all three SRT2 components. These technologies will be evaluated in participatory ways to ensure that they fit the labor, land and capital endowments of smallholders. Such evaluations will produce valuable data on the infrastructural requirements to increase improved technology use by smallholders. Data will be furnished 57

63 to policymakers to inform about the changes required to improve supply and marketing infrastructure that support implementation of sustainable intensification technologies without compromising ecosystem services. In many cases, opportunity exists to develop compensation and reward mechanisms for investment in natural resource conservation and ecosystem services (Swallow et al., 2009). Participatory rapid appraisal tools such as the TULSEA negotiation support toolbox and spatially explicit GIS tools such as Polyscape at landscape or InVest at regional scales, that explore synergies and trade-offs among ecosystem services (Nelson et al., 2009; Pagella, 2012) will be utilized and refined for integrated systems application to assess watershed functions, carbon stocks, land tenure and agro-biodiversity. These understandings will in turn set priorities and strategies for designing, testing and refining interventions through action research. Approaches described in SRT3 and Section 6 to strengthen institutional mechanisms that support improved natural resource management are particularly important here. Within SRT2.3 we intend to build on methods and tools developed in other initiatives, particularly CRP3, CRP5, CRP6, and CRP7 (see Section 9) and apply them in a systems context relating to the environmental stewardship of smallholders (van Noordwijk and Leimona, 2010). Box 5. Labor requirements versus crop yields in slash-and-burn agriculture. In sparsely populated areas, long fallow systems have large land endowments and only systems with low labor requirements can be used. The energy or labor efficiency of long fallow slash-andburn systems is the highest among all agricultural systems in terms of energy invested versus the amount of energy gained from crop yield. In the Amazon basin, only 8 % of the human energy input to cultivate a cassava field, including post-harvest processing, is used for slashing and burning. This is largely due to the absence of fossil fuel use and chemical inputs. However, the destruction of the biomass and the release of carbon in the burn are not considered in such calculations. In areas with higher population densities and consequently shortened fallow periods, economic conditions become more suitable for market-oriented, commercial farming. While labor is available, capital for agricultural inputs may be scarce. Thus, high-input, intensive agriculture is rarely practiced. With reduced fallow length, the labor requirement for clearing decreases yet overall labor increases as additional work, weeding and tillage become necessary. In short fallow systems of southern Cameroon, slashing, burning and cleaning the soil surface of weed stumps, was 10 to 13% of the total labor required to establish and harvest a groundnut / maize / cassava intercrop. Historically, the direct connection of slash & burn with labor availability was demonstrated when population density in the Mayan lowlands decreased and this was followed by a transition from intensive agriculture to long-fallow systems. Source: Hauser, S. and Norgrove, L Slash and Burn Agriculture, effects of. In: S. Levin (ed.). Encyclopedia of Biodiversity. Vol. 5, , Academic Press, San Diego, USA Research Outputs of SRT Ecosystem services identified, quantified and prioritized, and their respective threshold indicator values for sustained agricultural production established (addressing SRT2 Research Questions 2 and 3). Identification of essential ecosystem services: The identification of essential services will be linked to sustained agricultural production at the farm level and guidelines to better manage natural resources developed. The consideration of social and economic components is paramount as investments in resource conservation rely on smallholders understanding the importance of maintaining its benefits. Socio-economic farm surveys, participatory appraisals, and knowledge sharing platforms will provide data on smallholders endowments and bottlenecks related to NR management and investments. Scalability of ecosystem services: The identification of essential 58

64 ecosystem services will be based on compiled results of biophysical and socioeconomic analyses at lower levels (field, farm, household) up-scaled to community and landscape through models to understand interactions across temporal and spatial scales. Essential services such as extended agricultural production, carbon sequestration, protection of agro-biodiversity and beneficial organisms, suppression of pest and disease outbreaks, preserving surface and ground water quality, and protection against flooding, landslides, and siltation of ponds and lakes will be determined at community, landscape and regional levels. This will involve a wide range of stakeholders, including policymakers. Impaired ecosystem services threaten not only rural communities but also distant urban populations, an assessment that requires linkages with CRP5. Ecosytem service trade-offs: Different ecosystem services cannot always be improved simultaneously as many trade-offs exist. This will require a system research approach on the quality, quantity, spatial patterns and temporal processes of these services relating to the intensity of land use. For example, studies will be conducted to determine if certain NR functions of natural systems can be sustained in confined, less affected areas, while more production related resource functions will be the primary focus in areas of rapid land use change. At the global level, it may be advantageous to intensify production through specialization (e.g. perennial plantations) while protecting nearby diversified smallholds and natural ecosystems. Both land use strategies have potential to sequester carbon but the latter provides additional benefits through protection of biodiversity. An example is the carbon-rich, shaded smallholder cocoa systems with nearby protected forests (Gockowski et al., 2010) including the Bia National Park and Krokosua Hills Forest Reserve in the Western region of Ghana. Particular interest will also be on the options for resource restoration in degraded areas through reintroductions from conservation areas. These activities will contribute to the understanding and management of ecosystem services under different intensities and patterns of land use change. 2. Methods developed to determine agro-biodiversity and natural resource condition at different scales (addressing SRT2 Research Questions 1, 2 and 3). Assessment of natural resource status: The development of reliable methods to determine biodiversity and resource status and their changes as a consequence of use at different special scales will be based on existing methodology of quantifying biophysical parameters of soil, water, vegetation mass and biodiversity assessment. Development of indices related to the NR status: Research will further focus on simple and cost efficient methods that quantify combined parameters of soil, water and biodiversity into indices closely related to overall natural resource condition and essential ecosystem services. This output again requires linkages with CRP5. 3. Models developed to quantify and analyze the effect of NR interventions on the productivity, farm systems components and interactions between these components (addressing SRT2 Research Question 4). This is conducted with emphasis on understanding tradeoffs between short-term production gains and longterm NR sustainability issues at the farm level and the effects of agricultural practices on landscape level ecosystem services. Identification and adaptation of models: Research approaches will be designed to include the collection of data with specific use in model parameterization and validation. Existing models will be reviewed for their ability to support the identification of essential services and the prioritization of linkage to agricultural production. Others will be explored in terms of assessing changes in the resource base as a consequence of technological, social, economic, infrastructural and policy interventions. Landscape models, such as ones developed by CSIRO, will be modified as required to support the assessment of the current resource base, its services and changes as a consequence of land use at watershed, landscape and regional levels. Economic models will be used to demonstrate the impact of intensified agriculture on income generation, resulting from investment and improvement in NRM. 59

65 Scenario analysis: Combining research and participatory process, the above models will be run to calculate ex ante impact in a dynamic biophysical and socio-economic context, including potential technological, infrastructural and policy improvements on smallholders livelihoods and their ability to invest in resource conservation, as well as larger scale effects on more distant areas and populations. 4. Policymakers sensitized on the importance of natural resource integrity on rural development, reward systems for investment in resource integrity, and crafting an enabling social and economic environment for sustainable agricultural production. Sensitization to the importance of NR integrity: The sensitization of policymakers to the importance of resource integrity for future food production and income generation and the value of incentives for natural resource stewardship will be demonstrated through the development of ex-ante analysis of non investment versus interventions sustaining NRM. Sensitization on compensation and reward systems: The sensitization of policymakers on the importance of compensation and reward systems for investment in natural resource integrity will be demonstrated through the results of ex-ante analysis of the consequences of no compensation of smallholders implementing NR conservation measures with potential long-term compromising effects on income and food security. Such models will be up-scaled to demonstrate as well the importance of an enabling social and economic environment (market access, input supply, access to information and knowledge, access to credit, etc.) for sustainable intensification of crop, animal, and tree production. 5. Interventions combining technologies, incentives and policies that enhance productivity, reduce risks and improve ecosystem services are piloted and validated. This allows direct use of natural resources while protecting delivery of essential ecosystem services. Identification of NR improvement practices: The development of agricultural production technologies that combine increased food production and income generation with NR conservation and improved NR integrity will comprise the use of external inputs ranging from fertilizer and natural nutrient sources (manures, biomass, compost), to improved fallow management, rotations and biomass use that increase soil carbon and N-availability. For pest management, the focus will be on better understanding pest cycles and their response to abiotic and biotic factors which will permit the design of more pest- and-disease suppressive production systems. Technologies will focus on alternative land use management without compromising labor requirements and crop yields. Aspects of biomass trade-offs, croplivestock integration and the threat of spatial separation of these enterprises with increased market participation present new challenges to be addressed. This activity will be implemented together with SRT2.2 with a focus on agro-ecological intensification and the smart use of agro-biodiversity. Identification of NR conservation practices: In areas of relatively high NR integrity, production often suffers from high pest and disease incidence and severity. Environmentally sound pest and disease management technologies for key income and food security crops and for crops drawing heavily on the NR base will be developed or fine-tuned to reduce crop yield losses. Where labor is limiting, such technologies can lead to a reduction in cropped area while maintaining crop output and income. Here, research will focus on the consequences of reduced land use frequency at increased land use intensity and contribute valuable data toward best ways of conserving NR and biodiversity. This activity will be implemented together with SRT2.2. Assessment of tradeoffs between NR integrity and productivity: The effects of any technology will be quantified through a set of essential and sensible soil properties, water availability and the status of biodiversity and economic data obtained in the intervention area to determine tradeoffs between NR integrity and productivity. For example, potential trade-offs in biomass use (mulching vs. animal feeding) on short-term farm production and long-term soil sustainability can be assessed by combining experiments and modeling tools described in SRT2.2 and currently 60

66 implemented by the System-wide Livestock Programme (SLP). Lifecycle assessments of livestock: Selected technical interventions will be offered for inclusion within rural development efforts that facilitate livestock impacts on water, land and biomass use at varying levels of intensification. 4.9 SRT3: Scaling and Institutional Innovation Introduction SRT3 is in many ways extends the boundaries of CGIAR research in that it attempts to understand how complex research outputs generated in SRT2 and supported by SRT1 are brought to scale to achieve CGIAR development outcomes, particularly in different market, institutional and policy contexts. SRT3 combines research on three critical areas central to Humidtropics, namely (1) understanding how to take complex technologies often integrating productivity with sustainable resource management to scale under different market conditions, (2) evaluating alternative methodologies or frameworks for scaling up, especially in terms of their cost effectiveness, and (3) through the scaling up testing how impact on development outcomes can be enhanced, if not optimized in essence evaluating alternative impact pathways. Providing the technological and market conditions for sustainable production system intensification, especially where there are significant differences in farm resource endowments, farmer education, and quality of the natural resource base, requires innovative institutional approaches in providing information, building farmer capacities, organizing farmers, linking farmer groups to service providers, and building more efficient input and output value chains. Humidtropics adopts an agricultural innovation systems approach to linking a range of organizational actors in the agricultural sector that can promote evolving farmer investment strategies resulting in farming system change and intensification. Implicit in this developmental research is understanding why so few proven technologies become adopted at scales large enough to demonstrate impact and sufficient return on investment. SRT3 is a blend of agricultural and social sciences. Insights from innovation studies adopting a complex systems perspective clearly indicate a co-evolution between different technologies, socio-economic structures and institutional arrangements (Ekboir, 2003; Elzen and Wieczorek, 2005; Geels, 2004; Geels and Schot, 2007; Hall and Clark, 2010). This co-evolutionary process involves generating a variety of developmental strategies and comparing them in a manner that identifies which intensification realms are most viable under certain conditions. This implies that many innovation options still will not reach the larger developmental mainstream. Scaling can thus be seen as a measure of a particular intervention to have made it from a niche in an Action Site to the mainstream regimes of an Action Area. This SRT also supports other CRPs by including their most promising innovations into institutional agendas that would not otherwise have been achieved. Geels (2004) has described a multi-level model of socio-technical transitions (Figure 15). These levels, taken into the context of Humidtropics include: 1. Innovation niches exist within research institutes where technologies and interventions originate. Within the context of developmental research, some promising innovations emerge as candidates to replace current technologies and practices and contribute to a new mainstream, while other less-viable innovations fade away (Geels and Schot, 2007). 2. Within the mainstream, dominant technologies and institutional arrangements operate below their potentials. Here candidate innovations literally complete with the status quo and the most useful of which become incorporated into a better performing mainstream. Within a developmental research context, this competition occurs at different levels 61

67 (Action Areas, Action Sites, nations and regions) and involves changes in technologies, markets, policy, livelihoods and culture. 3. Meaningful change occurs within a broader landscape of natural, social and economic conditions that in turn influences innovation develop (level 1) and mainstream regimes (level 2). SRT3 addresses the challenge of building more effective strategies to enable a wider diversity of stakeholders to implement system improvements in ways that are equitable and sustainable. This demands institutional change that better coordinates the actions of these stakeholders (CAPRI, 2010; Rajalathi et al., 2008; World Bank, 2006), helping them to achieve things collectively that they would not achieve on their own. The effectiveness of institutions can, therefore, be measured by the extent to which they influence behavior in ways that contribute to achieving desired outcomes. In addition to effectiveness, institutions should also be efficient achieving the desired objective at the lowest cost and equitable in terms of how the benefits and costs are spread across all those affected. Figure 15. Stylized multi-level model of socio-technical transitions (adapted to CPR1.2 from Genus and Coles, 2008, who adapted Geels, 2004) Research for development projects increasingly recognize the need to move beyond a technology dissemination paradigm to address institutional issues through employing a range of approaches including innovation systems (Rajalathi et al., 2008; Hall, 2009), using multi-stakeholder alliances (Lundy et al., 2008), facilitating involvement of a wider range of development partners (Kristjanson et al., 2009), and conducting more participatory and action-oriented research (Scoones and Thompson, 2009). SRT3 will identify appropriate institutional arrangements to connect diverse stakeholders to better share knowledge and strengthen their capacities through joint action. This leads to robust outcomes that impact on livelihoods, enhance economic and environmental sustainability (beyond conventional short term projects) and enhance the ability of smallholders to respond positively to opportunities made available within their Action Areas. 62

68 4.9.2 Research Questions Humidtropics has adopted a sequenced set of research questions that have previously been devolved to other research or development organizations on how to evaluate alternative impact pathways through different approaches to going to scale. These are set out in respect to the three broad areas that introduced SRT3 above. Understanding How to Take Complex Technologies to Scale SRT3 Research Question 1. How do farmers operating under different resource endowments and access to input and output markets assess and adopt complex, systembased technologies? Rationale. There has been some experience with disseminating complex, system based technologies such as conservation agriculture, ICIPE s push-pull technology, agroforestry systems, or control of banana bacterial blight using only cultural practices. Adoption of such technologies often entails a staged process of sensitization, sourcing information and/or technology components, experimentation, and adoption and integration into the farming system. This process is most often supported by intensive adaptive research and dissemination projects without a framework to assess farmer evaluation and adoption decisions and often without understanding how alternative dissemination strategies impact on what is a cumulative assessment process by farmers. Humidtropics will provide a framework to systematically conceptualize and assess farmer decision making at a production system level, choice among technology options, and adoption under alternative dissemination strategies (see below). SRT3 Research Question 2. System intensification involves a phased process of technology adoption, increased incomes, and changing farmer investment priorities. What linked institutional innovations are necessary to nudge farmer investment strategies along pathways of sustainable system intensification? Rationale. Due to the CGIAR s past focus on component technologies and adoption of these, there were few studies that assessed how farmers reinvested income gains from initial adoption of those component technologies, or even if those income gains were reinvested in the farming system itself. Humidtropics has framed this problem in terms of initial adoption around critical entry points and phased investment in other areas of the farming system (often moving from crop to livestock investments), particularly in increased investment in the natural resource base. Institutional and policy innovations in the areas of micro-credit, insurance, reduced information asymmetries, and improved access to input and output markets can investment choices, particularly those where the returns are lagged by more than one cropping season. Understanding farmer investment choices is thus critical to both understanding alternative intensification pathways as well as modalities for scaling up. Evaluating Alternative Dissemination and Scaling Up Methods SRT3 Research Question 3. Given that scaling up production system research oriented to sustainable system intensification requires a combination of new technologies, necessary institutions, and an appropriate incentive environment, what are the most effective dissemination and scaling up approaches? Rationale. Over the last decade there has been substantial work on new approaches to dissemination, building on adapting ICT s to dissemination, new theories such as innovation systems, and the multiple approaches to traditional extension. At least five principal approaches or frameworks might be distinguished, namely (1) scaling up within a 63

69 communication/ict framework; (2) scaling up within a farmer capacitation framework, eg farmer field schools; (3) scaling up within a organized advisory service framework; (4) scaling up through improved market linkages and investments; and (5) scaling up within an innovation systems framework, particularly optimally deploying innovation platforms. At the same time there has been very little systematic assessment of these different approaches, primarily work on comparing FFS to traditional extension approaches (when in fact the projected outcomes are quite different) and the SSA-CP s evaluation of innovation platforms. It would be useful to compare these different approaches in terms of both cost and effectiveness in increasing farm factor productivity but at the same time in understanding the potential synergies between different approaches on farmer decision making, for example the use of ICT approaches in farmer sensitization and FFS in building farmer adaptive research capacity. SRT3 Research Question 4. Given that agricultural innovation system (AIS) approaches tend to combine most of the characteristics needed for taking production system intensification to scale, what are the design options in AIS approaches that allow cost effective scaling within the Action Areas? Rationale. Agricultural innovation systems is in many respects the leading edge in R4D conceptualization but practice has been mostly limited, apart from the SSA-CP, to a range of pilot projects applying the approach in quite specific themes (see the associated box on FIP). This has produced a range of methods but without an effective evaluation of alternative design options, particularly the scale at which innovation platforms can most effectively operate and how these platforms function when organized around other constructs than just value chains. Given the very different contexts across the different Action Areas, SRT3 has the potential to evaluate the relative effectiveness of alternative designs of AIS approaches. Evaluating Impact on Development Outcomes SRT3 Research Question 5. Inherent in the approach to scaling up (as represented in the research questions above) are direct impacts on productivity, food security and natural resource integrity. However, there may be impacts on poverty and gender equity from direct production system interventions but these cannot be predicted. How then can impacts on gender equity and rural poverty be both understood and assured in the design of scaling up strategies? Rationale. Humidtropics adopts the approach that impacts on rural poverty and gender equity are not a matter of ex-post evaluation but are inherent in the design of scaling up options. Rural poverty and gender equity are conditioned by social factors involving power relations, community social structures, and constraints on access to services and resources. Scaling up approaches across the five frameworks sketched above can be designed to assure access by poor households and women. Market participation is a critical area to ensure equity and institutional innovations in farmer organization, collective action, and access to services can be designed to improve equitable access to markets Methodology Systematic research in the areas of understanding farmer investment trajectories and associated stages in technology adoption, evaluating alternative dissemination and scaling up models, and designing institutional innovations that achieve equitable outcomes is virtually non-existent. While this represents an enormous opportunity for the work of Humidtropics, it also creates the challenge of developing new research methodologies to support this work. Inherently in this type of work there will be trade-offs between rigorous statistical and quantitative approaches and more process-oriented, action research 64

70 approaches where systematic monitoring is key to evaluating performance and outcomes. The following will provide only a brief sketch of methodological approaches to the set of research questions set out above. Research on farmer investment and intensification trajectories will build on the integrated production system and market agent panel surveys. A range of modeling approaches can be built around this data set from constrained optimization or programming models at farming system or regional agricultural economy level, potentially using a social accounting matrix that can be constructed from the panel data. More behavioral models of farmer decision-making can be estimated from the panel survey data and contextualized using the variation in market and policy conditions. Most importantly the panel data will provide the basis for monitoring production system change with alternative interventions. Whether this is done within an experimental framework will depend on the research questions being asked and whether appropriate counterfactuals can be established, a particular difficulty with a long-term research site. The rigor of research design and the use of counterfactuals and randomization are even more of an issue in the evaluation of alternative dissemination approaches. Much of this work will be done within an action research modality, where the focus is systematic monitoring and assessing outcomes in relation to a series of critical inputs, particularly cost. On the basis of this type of assessment alternative best-fit models or approaches may emerge and the choice of whether other development agencies may take these to scale outside the Action Areas or in other Action Areas may necessitate a more rigorous and costly evaluation design, which may have to be done outside the Action Area in order to have clear counterfactuals within a fully randomized design. The experience of the SSA-CP in undertaking such an evaluation for innovation platforms should be instructive in both the decision to undertake such an evaluation and the conditions under which a systematic design can be implemented. The key to the work on directing impact to gender and poverty outcomes primarily comes in the actual design of the institutional innovations and then monitoring outcomes in their development and implementation. This work will be linked to the research on alternative dissemination models and will entail some of the choices in the selection of evaluation methods. Understanding poverty outcomes at the level of the monitoring at Action Area level will be linked to the sectoral/sam modeling work above. In general it will be difficult to attribute changes in development outcomes at this level to just the work of the Humidtropics investments but some of the causal determinants may be elucidated by combining modeling and monitoring results Research outputs of SRT3 Research outputs for SRT3 follow directly from the methods outlined above, namely: 1. An interactive modeling, panel survey, and monitoring framework for understanding the interaction between farmer investment, system intensification pathways and interactions with market structures (addressing SRT3 Research Questions 1 and 2). 2. Fully evaluated, best bet models for dissemination and scaling up of production system strategies developed for potential uptake by development partners (addressing SRT3 Research Questions 3 and 4). 3. Institutional innovations to achieve gender equity and poverty outcomes designed and evaluated (addressing SRT3 Research Question 5). 4. Developed research agenda and research community around scaling up, particularly using agricultural innovation system approaches (addressing SRT3 Research Questions 4 and 5). 65

71 5 QUANTIFIED IMPACT PATHWAY 5.1 Humidtropics Impact Approach The Humidtropics approach to impact pathways is to treat it as a research area in its own right. CRP1.2 will evaluate alternative impact pathways under different market, institutional and policy contexts in its work on scaling up in SRT3. Given that production system intensification is highly location-specific, one of principal challenges in Humidtropics is how to conceptualize approaches to going to scale and this intrinsically will involve improved capacities at farmer and institutional level. In many respects the central rationale for the Action Area definition is to serve as a laboratory for understanding the interplay of markets, institutional innovations, and policy in fostering the sustainable intensification of production systems and, most importantly, how impact on SLOs is a design feature of alternative approaches to scaling up. Thus, Humidtropics moves away from viewing impact pathways as purely a set of institutional linkages from research to technology dissemination. Rather it is within a more innovation systems framework of interacting actors in the agricultural sector with a focus on integrated approaches, that are linking technology, farmers adaptive capacity, informed farmer investment strategies, more efficient supply chains, full participation by women and poor households, and meeting the needs of processors and marketing agents. The institutional arrangements that facilitate such integrated approaches is then the focus of scaling up and thereby tests alternative impact pathways. The Humidtropics R4D approach strongly includes the social, cultural and psychological aspects that are embedded within SRTs 2 and 3. The main premise is that the poor and vulnerable themselves make important decisions governing the best management of available agricultural resources and that the most effective interventions are those that build upon farmer knowledge and practice rather than approaches that seek to supersede current systems in the cause of modernization. This principle is illustrated in Figure 16 and reflects the program s impact pathways. Indeed, the underlying process of agricultural intensification involves not only technology development but also a working understanding of social and cultural frames of references. The Humidtropics approach thus considers efficient progress in basic research and technology development and supports effective poverty reduction strategies focused upon the needs of rural households. There are several ways in which agricultural research and technologies influence livelihoods, reduce risks, and build assets through different forms of capital and collective action (Adato- Meinzen, 2007). The Humidtropics Program Pyramid (Figure 16) not only provides entry points to address poverty, but also provides a better insight into how agricultural technologies can be adapted to specific situations. It also addresses institutional issues that enable effective use of technologies (Adjei-Nsiah et al., 2008; Dormon et al., 2007; Klerkx et al., 2010). Its approach starts with the needs Figure 16. Humidtropics Program Pyramid 66

72 Box 6. Constraints to systems productivity may be rooted within complex social settings. Gender inequity is often influenced by the social and cultural context and decisions made at the social-cultural level. Innovations introduced without a sound understanding of this context can lead to undesirable consequences. For example, innovations successful in increasing commercial maize yield led to deteriorating nutrition of women and children in northern Zambia because it was not understood that there were separate but interacting male and female headed systems, rather than integrated household systems (Rocheleau, 1987). Men were in control of commercial maize (that was sold) but women contributed their labor. Women were in control of growing local maize varieties (eaten, not sold) in separate home gardens more commercial maize meant less time for subsistence maize. In Bangladesh, women contribute about 80% of the labor involved in dairy husbandry. However, cultural norms do not allow them to participate in markets and this bars them from benefiting from their efforts. and opportunities of farming households within Action Sites but also considers the institutions and drivers of change that affect their lives (Figure 16). Poverty has a clear geographical dimension (Okwi et al., 2007) and Humidtropics includes challenging Action Sites where poverty and population density are high and natural resources integrity is low as a test of its diagnostic and problem solving capacities. Indeed, constraints to systems productivity may be rooted within complex social settings (Box 6). In this way, household members will choose among proven management options and their decisions will be analyzed within an economic, natural, institutional and social setting, providing better understanding of the scalability of program technologies. Innovations and technologies will be pioneered within Action Sites, the most effective and adoptable of which will then be piloted within their respective Action Areas. This is achieved through an iterative process between the targeted farm types and household income strata and the service providers. These innovations will also be exchanged among Action Areas for evaluation and adjustment to site-specific conditions. These innovations are not only technological but can also represent new social, institutional and organizational arrangements (World Bank, 2006). Open Innovation systems are needed, in which new actors and institutions are constantly being created, changed, and adapted to suit the dynamics of scientific and technological creation (Juma and Yee-Cheong, 2005). As interventions move from the pilot stage to wider implementation, it becomes more important that they be buttressed by an enabling institutional and policy environment (Klerkx et al., 2010). 5.2 Stepwise Impact Pathways Development impact and program success require targeted agricultural research outputs and outcomes. This R4D model was selected for Humidtropics as a holistic impact pathway that permits close feedback between research and developmental outcomes and readily responds to changing needs (Figure 17). This model ensures that development needs and impact serve both as a guide and goal for program investments by integrating the Strategic Research Themes through a six-step approach (Figure 17). 67

73 Figure 17. A six-step model translating research actions into developmental impacts. Step 1: Development needs identified (SRT1). Needs are identified with the participation of policy, research and development partners to achieve a holistic understanding of the problems at hand, and identify candidate solutions (Klerkx and Leeuwis, 2009; Wongtschowski et al., 2010). To reach key policy processes and practitioners communities, the first line of engagement for impact will be with NARS and International NGOs and development partners on the most promising combinations of tree, crop, and livestock use, as well as their specific natural resource, market and institutional challenges. Step 2: Research action undertaken (SRT2). Research design, management, and output are aimed at addressing the development needs in partnership with national and international research institutes, universities, and nonprofit and private sector research organizations. Findings are then translated into improved options within and between Action Sites. This research process demands vision into the potential research and development outcomes, and accompanying institutional change with which to achieve research outcomes (Dormon et al., 2007; Adjei-Nsiah et al., 2008). Mitigation of and resilience to risk are major considerations. Development of environment-friendly crop management alternatives, resistance to pests and diseases, abiotic stress tolerance, and improved tree, crop and livestock husbandry, and subsequent economic and nutritional benefits will be addressed. Step 3: Research outcomes translated into developmental actions (SRT2 linked with SRT3). Research impact is determined by the sharing of scalable research outcomes and accompanying advocacy activities with development partners and outcomes are adjusted to fit local circumstances (Douthwaite et al., 2001). A successful outcome entices partners to adoption, adaptation, and learning activities. Research and development partners are involved in the identification of extension strategies, policy advice, and advocacy, to enable a favorable environment for embedding the research outcomes. A gender-sensitive perspective is required, as women face particular challenges in accessing information, extension, advisory services and education. Step 4: Development actions examined (SRT3). Development partners and beneficiaries design and participate in activities across the larger Action Areas. They not only 68

74 disseminate tested technologies but also provide mechanisms for service provision and knowledge exchange. Contributions from public and private sector partners are translated into opportunities for input supply, market expansion, and credit facilities across the Action Area to promote rapid technology distribution and adoption. Step 5: Development actions mainstreamed (SRT3). Tested, effective development actions are incorporated into a wider Action Area development agenda. The responses from the poor and vulnerable to the diversified options made available to them provide insight into those interventions that warrant scaling up. All partners are involved in ex-post evaluations that compare baseline information to measure the impact on beneficiaries. Expected impacts are reduced poverty, improved ecosystems health, enhanced food and nutritional security, not only in smallholder households and communities, but also in urban areas that depend on rural farming systems. Mechanisms and lessons learned from these outcomes are incorporated into the program s Global Synthesis and translated into International Public Goods. Step 6: New development needs identified (SRT3 linked to SRT1). Development impacts pose new challenges that are referred back to development needs, providing additional insight into the continuing Situation Analysis. Box 7. Example using the Six-Step Model to illustrate the impact pathway for cocoa-based systems in West and Central Africa. In 2003, after an assessment of the need to revitalize the smallholder cocoa sector in four West and Central Africa countries (i.e., Ghana, Nigeria, Cote D Ivoire and Cameroon) which account for 70% of the world supply, the World Cocoa Foundation, the cocoa industry, and USAID created a publicprivate partnership called the Sustainable Tree Crops Program, managed by IITA. In Step 1, impact assessments were conducted to quantify the role of cocoa production and sales on agricultural growth and poverty reduction in the four counties. This priority assessment identified a number of key research actions for Step 2, mainly focused on management approaches to rejuvenate existing older plantations, IPM practices for control of cocoa black pod disease, strengthening collective actions toward better markets, and alternative crops for cocoa producing households. The research impact of Step 3 included scaling out of new pruning and replanting techniques in Farmer Field Schools, adoption of new quality control methods and drying techniques by cocoa cooperatives, and transfer of new cocoa varieties from Ghana to the other three countries. Step 4 development actions included National government and private industry providing farmer training and community development. Cocoa cooperatives also diversified their services to include provision of alternative crop and tree varieties to their smallholder cocoa farmers and others in the community. The Step 5 development impacts increased income for the adoption of new management by cocoa farmers, cooperatives, and communities. The private sector benefited through increased access to high quality cocoa. Income from plantain, fruit trees, and cassava, which usually serve as temporal shade for the cocoa during its earlier years of establishment rose from the access to improved planting material by the community. This impact led to a number of new challenges to be addressed in Step 6. These are the production challenges of cocoa certification schemes, sustained intensification with better use of inorganic fertilizers, sustained access to inorganic fertilizers and other inputs, need for new high-yielding varieties and better management approaches that avoid continued deforestation due to expansion of land area by smallholders. 5.3 Quantifying Production System Impacts Quantifying impacts of a large, integrative R4D project requires a series of well-planned measurements throughout all stages of program activities and at scales that link representative Action Sites to their larger respective Action Areas and beyond. The Situation Analysis performed early in the project not only prioritizes critical entry points for integrative research, but also assembles current baselines concerning households, crop enterprises, current management and yields, market opportunities and commodity prices, 69

75 the status of key natural resource endowments and the engagement of women, skilled youth and the most vulnerable members of the community. These parameters construct an essential baseline against which improvements in farm enterprises, household well-being and business activities are weighed. Key data must also be collected within the Action Sites during the process of production systems integration, particularly during the process of adaptive, on-farm research. Often, pilot production packages may be assembled and distributed to large numbers of farmers and data collected on the agronomic efficiencies of inputs, and resulting yields and economic returns. Farmer adjustments to these technologies are also important. An M&E framework must be constructed and field-tested that captures this baseline and interacts with research outputs from the Action Sites and as proven technologies are scaled up to achieve development outcomes (see Section 13). Scalability is evaluated within Action Areas and requires that information be gathered on the potential land area targeted for production gains, adjusted for outreach coverage, farmer adoption, site-specific management and marketing innovation by stakeholders. A list of parameters essential to establish farm baselines, stepwise improvement of production systems and larger impacts that result from subsequent extension campaigns is presented in Box 8. Box 8. Quantitative parameters allowing calculation of impacts from integrative farm enterprise based upon current practice and extension recommendations (after Sanginga and Woomer, 2009). Clients and area within Action Site: number of households participating within specific intervention; adaptive land area; total intervention area (total area may also be calculated through other means such as amount of improved seeds or inputs distributed) Input and service costs: seeding rate; unit cost or value of seed; current rate of soil and pest management inputs applied; target rate of input application; price and costs of inputs; total input costs; additional services to each household; total service costs; total intervention costs Baseline yields and returns: current crop yield; current production; commodity price and current value Current recommendations and returns: Agronomic Efficiency of currently recommended inputs; increased yield; increased production and value; benefit: cost ratio of current recommendations Response to integrative best management: Integrative input regime costs; Agronomic Efficiency of inputs; additional value from resource gains (e.g., N fixation); integrative yield gain; increased production and value; integrative benefit: cost ratio Intervention (project) investment summary: Total inputs mobilized; total increased production; gross and net returns; intervention benefit: cost ratio Benefits per household: additional seed and inputs mobilized; input costs; yield and value gains; net return; integrative household benefit: cost ratio Adjustments for scaling up to Action Areas, coverage, adoption and innovation: potential land area of intervention and production gains; adjustors for coverage, adoption, site-specific management and marketing innovation; total adjusted value and net returns 5.4 Quantifying Natural Resource Integrity The need exists to monitor the agricultural resources as production systems are improved so as to ensure that production gains are not accrued at the expense of resource integrity (see Section 4.8) In order to do so, a suite of practical indicators must be developed; soil baseline conditions established and then monitored over time based upon landscape, physical, chemical, biological and land management criteria applicable to target production systems. The following criteria provide means for rapid assessment and ranking of agricultural resource integrity based upon the observations of landscapes, surface water, soils, plants and beneficial soil organisms (Sanginga and Woomer, 2009). 70

76 5.4.1 Landscape criteria Exposed soils are more susceptible to erosion and compaction. This criterion is expressed as a percentage of total farm area using walking transects, and measurements should not be taken early in the cropping cycle before crop canopies have closed. Severity of soil erosion may be expressed as a percentage of total farm area using walking transects. Extreme erosion signals the need of land restoration rather than production systems integration. When fields are sloped, structures built along the contour are necessary to control soil erosion. These structures may be terraces, bunds, rock and trash lines, hedgerows or grass strips. The presence, length and distance between contour structures are important indicators of soil conservation awareness. The width of riparian strips may be either scored or measured. In addition, well-protected streams are clear and poorly protected streams are usually muddy Chemical and physical criteria Nutrient deficiency symptoms offer quick insight into soil fertility status. Plant deficiency symptoms may, however, be confounded with one another and by moisture stress, waterlogging and pathogens, so they are best calibrated through soil testing. Soil acidity is readily measured using litmus strips, inexpensive hand-held instruments and by portable soil test kits and is corrected by applying agricultural lime. Soil carbon is measured in the laboratory by acid digestion-calorimetric analyses or using complex instruments that are not well suited to rapid assessment of soil health. Two SOM fractions, microbial biomass C and particulate organic matter indicate the short and mid-term dynamics of organic matter additions to soil, but require more complex measurement. A portable, hand-held device that measures soil carbon based upon spectrography has recently become available. Stable soil aggregates resist erosion and permit a healthy combination of air and water within soil void space. Both of these measurements are conducted using carefully collected soil cores and straightforward analytical procedures within the soil physics laboratory but these laboratory measurements are difficult to include within rapid field-based assessment Biological criteria The presence and degree of selected beneficial and detrimental organisms can be described through careful and experienced field observation. The presence and effectiveness of rhizobia, the microsymbiont associated with nitrogen- fixing legumes, may be inferred by the abundance, size and interior coloration of root nodules. Soil macrofauna are important indicators of soil health because their activities accentuate soil physical properties and nutrient recycling. Foremost indicators of soil fauna are the presence of earthworms and large soil grubs (insect larvae). Termites are important soil engineers but also attack crops, trees and wooden structure and represent a mixed blessing. Detailed information on quantifying soil fauna may be obtained from Moreira et al. (2008). Striga is a serious parasite of cereals and legumes in Sub-Saharan Africa, and its presence is an important indicator of soil health because severely infested hosts do not respond well to management. Striga seed banks are quantified in soils using elutriation-density separationcounting procedures (Odhiambo and Woomer 2005). Alternatively, the severity of host infestation may be scored in the field as emergent stems per host plant. One need not know the causal organism to ascertain that root systems are not well developed. Root disorders vary among crops, with some being resistant and others chronically affected. A visual ranking of root disorders includes severely stunted, galled or rotten roots. 71

77 Box 9. A simple checklist approach to assessing agricultural resource status in conjunction with production system integration. CATEGORY AND INDICATOR RANKING SCORE Landscape criteria Proportion of exposed soil widespread (0), frequent (1), occasional (2), absent (3) Severity of soil erosion severe (0), moderate (1), slight (2), absent (3) Presence of contour structures absent (0), distant (1), regular (2), complete or no slope (3) Protection of riparian strips absent (0), narrow (1), compliant (2), copious (3) Surface water clarity muddy (0), cloudy (1), opaque (2), clear or no surface water (3) Landscape subtotal Nutrient deficiency symptoms Basal leaf chlorosis and drop severe (0), moderate (1), occasional or slight (2), absent (3) Purpling of lower leaves severe (0), moderate (1), occasional or slight (2), absent (3) Marginal leaf necrosis severe (0), moderate (1), occasional or slight (2), absent (3) Basal interveinal necrosis severe (0), moderate (1), occasional or slight (2), absent (3) Apical chlorosis or tip distortion severe (0), moderate (1), occasional or slight (2), absent (3) Symptom subtotal Chemical & physical criteria Soil acidity extreme <4.5 (0), severe (1), moderate (2), neutral (3) Biological criteria Legume root nodulation absent (0), sporatic (1), abundant (2), super-abundant (3) Nodule interior color white (0), pink (1), red (2), dark red (3) Soil macrofauna detrimental (0), largely absent (1), present (2), active and abundant (3) Striga infestation (stems/plant) >2 per plant (0), 1-2 per plant (1), <1 per plant (2), absent (3) Root disease (see key for details) severe (0), moderate (1), occasional (2) healthy (3) Root galls (see key for details) severe (0), moderate (1), occasional (2) healthy (3) Biology subtotal Farm management criteria Crop residues (see key for wasteful or sold (0), retained (1), collected (2), processed or fed (3) details) Composting (see key for details) none (0), small piles (1), large piles (2), covered and fortified (3) Manure management (see key) absent or wasteful (0), haphazard (1), regular (2) processed (3) Pre-plant mineral fertilizers none (0), 75 kg ha -1 applied (1), kg (2), >150 kg (3) Top-dressed nitrogen fertilizers none (0) applied once (1) applied twice (2), applied thrice 3 Farm management subtotal Grand Total Farm management criteria The handling and application of organic resources and mineral fertilizers are important factors maintaining resource integrity within cultivated lands. These materials include crop residues and their utilization, composting, manure management, pre-plant fertilizer application and nitrogen top-dressing. It is difficult to gauge a farmer s organic resource and mineral fertilizer handling procedures from a single visit without queries of season-long practices. For this reason, rapid information is best collected through a participatory group discussion rather than a formal on-farm survey. These criteria for field assessment of agricultural resource management, intended for use at the farm level and illustrate that key indicators, may be monitored in conjunction with an integrated rural development program. These criteria are not exclusive and additional observations include tree, pasture and fallow coverage and agro-biodiversity. It is also important that representative observations be made across the Action Sites so they may be related to the landscape and watershed approaches advanced by CRP5 (see Section 9). Nonetheless, these criteria illustrate how a farm-level checklist (Box 9) may be developed for standardized use across a wide range of land uses within the Humidtropics Program. 72

78 5.5 Quantifying Market Impacts Quantifying impacts in the area of marketing requires that several, and very different, parameters be monitored. These parameters are related to agro-dealer activities, establishment and compliance with industry standards, initial situation and changes in local marketing opportunities, national trends in commodity imports, production, consumption and exports, announced tenders and prices offered by top-end buyers, and the growth of new opportunities in rural value-added processing (Box 10). A full spectrum of marketing parameters must be considered in the baseline study and initial Situation Analysis, and then key indicators monitored with time. In some cases, it is advantageous to work through existing agro-dealer networks and training activities in order to promote farm inputs accompanying new production systems. Similarly, farm input manufactures have existing retail agro-dealerships contacts that may rapidly stock new, proven inputs. Box 10. Quantitative parameters relating to markets, commodity marketing and rural social development. Agro-dealer activities: geo-referenced location and business name; owner (male or female); years in business, fertilizer in stock and packaging; seed in stock and packaging; inoculants in stock and packaging; links to extension, NGOs and farmer groups; past training; industry certification, credit received and extended; seasonal business volume; expressed needs Industry standards: commodity moisture content; insect damaged grain; diseased and discolored grain; off-color grain; foreign matter; households with tarpaulins and sieves, frequency of compliance Local marketing opportunity: engagement with local markets, price at local market, distance to local market, membership in marketing group and fees; access and distance to collection point; price at collection point; sales in local markets. National and export marketing opportunities: top-end buyers and locations; commodity imports, production, consumption and exports; announced tenders and prices; tariffs and related policies Value-added processing: opportunities for cottage and rural industry; value of product and production costs, frequency of household and community participation, local employment (number and wages) Quality standards Compliance with industry standards is an important indicator of marketable farm surpluses. Early in the season, compliance with moisture content is easily quantified at collection points as farmers market early to gain higher prices. Later in the season, insect damaged, diseased and discolored produce threatens quality. Top-end buyers have little tolerance of foreign matter, particularly small stones that can injure mills. Working with households to introduce tarpaulins and sieves, and to take care during shelling is the best means to improve quality. Compliance is best conducted within ongoing marketing operations at local collection points where sub-standard goods can be tracked to individual farms and simple kits containing moisture meters, sorting tools, scales and quality report forms can be distributed to collection point managers.smallholders trained in compliance often achieve higher standards than larger commercial farms reliant upon mechanized processes, and in this way brand recognition for smallholders produce may be established Marketing opportunities Expanded local marketing opportunity is a core impact accompanying systems productivity research and development. Initial engagement with local markets and prices offered there serves as a baseline. Opportunities may arise from membership in a marketing group, training in post-harvest handling, access and distance to new commodity marketing collection points and prices offered there, and innovative credit and repayment schemes. In many cases, local institutes such as schools and hospitals offer more favorable prices than 73

79 middlemen buyers and monitoring must remain alert to these buyers. Assessments of this sort are best performed within the Action Sites and then successes scaled up to the Action Areas. Finally, engagement in value-adding cottage and rural industry, the value generated through processing and local employment offered must also be considered against the simple marketing of produce. Cottage and rural industries are best described as case studies within Action Sites, with successes then replicated elsewhere Commodity trade At the broadest scale, marketing impacts are assessed at the national level. First, estimates are made of commodity imports, production, consumption and exports. As development outcomes are realized, the contribution by the Action Areas to changes in these statistics is assessed. Another approach is to identify the replacement of imports by commodities produced within Action Areas. This is achieved by identifying top-end buyers relying upon imports and securing their commitment to purchasing domestic produce. As domestic supply improves, then markets may also be monitored through publicly announced tenders and prices. Commodities intended for exports may be monitored in a similar manner through contacting exporters and their agents. In some cases, buyers and exporters may operate within the Action Areas, and monitoring these marketers alone will provide sufficient information to estimate project impacts without confounding the inclusion of production and marketing at a national level. In other cases, large-scale buyers are few and routine monitoring at a national level becomes more feasible. 5.6 Assessing Scaling-up Options and Development Impacts Quantifying outcomes within SRT3 is in many respects the most difficult challenge of all and will require new methods as the structure of CRP1.2 is put in place and the work progresses. The following will only outline some of the principal issues, and this starts with some framing of alternatives for scaling up. Five approaches or institutional frameworks for scaling up might be distinguished and this list is not necessarily comprehensive, namely 1) scaling up within a communication/ict framework; 2) scaling up within a farmer capacitation framework, e.g., farmer field schools; 3) scaling up within a organized advisory service framework; 4) scaling up through improved market linkages and investments; and 5) scaling up within an innovation systems framework, particularly optimally deploying innovation platforms. These constitute very different options with significant differences in both cost and potential effectiveness, especially in relation to the particular needs inherent in sustainable production system intensification. Within each it is also possible to build processes that particularly target women or poor households, which becomes an additional dimension in the evaluation of scaling up approaches Evaluation Approach CRP1.2 will be faced with some difficult methodological choices in how to evaluate the outcomes of this type of work, with the choices resting across a continuum from monitoring and adaptive management within an action research modality to fully randomized control trials (RCTs) with well-defined counterfactuals. Suffice it to say at this point that each has particular strengths and weaknesses and that there is not much experience with RCTs in agricultural research or rural development (as compared to their use in the education and health sectors). Experience in both the SSA-CP and HarvestPlus suggests that this approach is costly, is not conducive to design changes or adaptive management, and that to be effective the research question is usually quite narrowly defined. It is anticipated that more of the work will be done within a well-structured action research framework and that use of an RCT approach will be limited to the comparison of well-defined approaches that trade-off cost with effectiveness. 74

80 5.6.2 Monitoring Development Outcomes Impacts on system level objectives will be monitored at both the Action Site level through panel surveys and at the Action Area level through both the monitoring work related to the scaling up research and less intensive monitoring of baseline survey households. Performance of the Humidtropics program and impacts on development outcomes will be evaluated at the level of the Action Area. At this time there are no plans to evaluate spillovers outside the Action Areas. Assessment of the potential for impact is continuous in CRP1.2 and built into the overall research framework. What will be most interesting is to assess what types of modifications might be made in the program design that better allow Humidtropics to achieve its targets. It is probably more important to understand why the program is not achieving its targets than to have a research program that shifts much of its effort into more downstream development activities in order to realize those targets. This framework of continual assessment is a better alternative to the latter focus on meeting targets. 75

81 6 PARTNERSHIP MANAGEMENT AND CAPACITY DEVELOPMENT 6.1 Partnerships Horton et al. (2009) defined partnership as a sustained multi-organizational relationship with mutually agreed objectives and an exchange or sharing of resources or knowledge for the purpose of generating research outputs (new knowledge or technology) or fostering innovation (use of new ideas or technology) for practical ends. CRP1.2 adopts and builds upon this definition. Partnership within the Humidtropics intends to combine efforts to take fuller advantage of opportunities; consolidate operational and management activities: operate more efficiently by sharing costs and skills; integrate ideas, activities and goals with others; better attract needed expertise and experience; make better use of shared knowledge and ideas; more rapidly discover solutions to complex problems; add additional value to products, services, programs and activities; increase lobbying ability, influence, credibility and standing; and eliminate duplication and overlap between similar groups. The Humidtropics Program relies entirely upon partnership and much attention will be paid to how these partners are managed. Viable partnership requires understanding by all parties, and recognition of their respective roles and complementarity. Partnership must be mutually agreed, properly developed and well maintained, and when necessary ended. This is particularly true when dealing with different stepwise components within production, marketing and innovation systems. Productive partnership offers more than just the sum of the individuals or organizations working together but rather a more efficient approach to planning multiple activities, identifying and solving real problems, and expanding the scope of program successes. The Humidtropics is built upon a purposefully designed partnership structure intended to empower individuals, teams and organizations to change for the better. Promoting innovations of economic importance requires a coalition of interests among stakeholders such as farmers, entrepreneurs, development workers, researchers and supporting institutions that can work together to generate and use different types of knowledge, such as scientific, technical, organizational, institutional, market-related and managerial, that take advantage of emerging economic opportunities (Hall et al., 2005). There are no blueprints for ensuring the effectiveness of partnerships, and each case is a learning process for the partners involved. Hence learning to innovate (Douthwaite et al., 2009) for effective system interventions will be an integral part of monitoring and assessing how partnerships work in Humidtropics. The following is a general overview of different types of partnerships Humidtropics considers at the Action Site, Action Area, Regional and Global levels and some of their characteristics: Humidtropics will engage in and with community-based partnerships in Action Sites through a wide variety of initiatives. We refer to partnerships as being "communitybased" when they take place in a community, have community members involved and have a direct impact on the community, rather than within an organization or institution. Partnerships in either a rural or urban area have different advantages and drawbacks, depending on the focus and requirements of the partnership. For example, urban areas have the benefit of a larger population base from which to draw, and often have better access to resources and influence. People in rural areas tend to work more closely together due to their history and familiarity, and the fact that they usually know what local resources are available and how to get them. Humidtropics will thus consider these differences, especially when looking at innovations and value chains. Partnerships can span several people, communities, regions and/or nations at a time. While the poor and vulnerable people in Action Sites themselves face their own unique challenges, there are many common issues across the Action Areas and the program 76

82 as a whole that are best addressed by a more encompassing partnership. Humidtropics will explore the economies of scale associated with those partnerships to address longer-term, more complex issues and creates a bigger picture of the opportunities and innovations that are possible, while providing different locations to test or pilot solutions and models. Engaging a wide range of stakeholders and actors from the outset of the program and at key points throughout program implementation will strengthen the program s influence as well as the likelihood and rate at which system recommendations are adopted. Complementarities in knowledge and skills of the different partners, and comparative advantages will generate synergies and enhance cost effectiveness of interventions at Action Sites and Action Areas and contribute to the Global Synthesis and IPGs. In the same light, opportunities for sharing research facilities and administrative services will be implemented within all Action Areas. Our management of partners will be based on previous experience. For example, the Ford Foundation supported a program of organizational change, which analyzed experiences with partnerships and produced guidelines (Spink and Merril-Sands, 1999).In addition, private and civil society partnership will be guided through lessons from key sources (e.g., Spielman and von Grebmer, 2006; Smith and Chataway, 2009). 6.2 Organizing and Mobilizing Partnerships To leverage collective resources and expertise in each of the Action Areas, the program will seek to develop partnerships with diverse international, regional, national and local organizations, according to capabilities and expected roles (Table 11). The rationale for partner selection will consider: 1) overlap of agendas between the organization and Humidtropics, 2) complementarity of research or development strengths, 3) filling a significant gap in research or development expertise needed by Humidtropics and other partners, 4) clear geographic focus in one or more of the Action Sites, 5) comparative advantages for delivering promised SRT outcomes within the Action Sites and Action Areas, and 6) ability of contribute additional resources to co-finance or extrapolate program activities. Initial identification of potential partners resulted from the participatory consultation process carried out as part of the proposal preparation. It is clear that there are different types of partnership arrangements already being implemented in different Action Sites (i.e., CIALCA and SSA-CP), and therefore, Humidtropics should strengthen such partnership relationships in some cases, while in others, new platforms must be formed. Once Humidtropics starts functioning, a process for formalizing partnerships will be carried out with key stakeholders. The first-level platform for partnerships will emerge from the Action Sites where most of the initial characterization, research and development activities will be conducted. CGIAR centers will be in charge of coordinating and facilitating partnership formation, playing the role of innovation brokers (Klerkx et al., 2009; Winch and Courtney, 2007) around specific tasks related to the SRTs. Later in the program, partnership platforms more oriented to deliver development outputs throughout the Action Areas will be formed, as well as new partnerships forged to share IPGs. 77

83 Table 11. Partnerships of different types and scales, and their expected outcomes SRDP Note that local development partners include the target groups, community-based partnerships and public and private sector organizations Partnership in delivering research outputs. Table 11 illustrates partnerships of different types, and their expected contribution to program outcomes at different scales of operation, the Action Sites, Action Areas and global levels. Additional information for each of these levels follows: Partnership within Action Sites. Local research and development partners, such as a NARS, community-based partnerships, local universities, development organizations, local governments and policymakers, and private sector will participate more actively to codevelop and deliver research outputs related to the SRTs 1 and 2. This will include situation and priority analyses and the improvement of integrated systems, particularly focusing on the development and delivery of research outputs of local application, validation of action research methods, better understanding of adoption processes, and interdependencies of biophysical and socio-economic systems at Action Site level. The conceptual framework and methodological support will be provided by CGIAR and other international research organizations, while the research implementation process towards developing integrated strategies using participatory methods will be the responsibility of local partners. In addition, capacity-strengthening activities will be planned and implemented as needed with partners. A communications plan will be developed and implemented in consultation with all the partners in each of the Action Sites. Partnership within Action Areas. Partnerships at this level have regional coverage and develop regional research outputs or public goods. They will be key for achieving the development outputs included in SRT3. This level will include NARS, universities with regional influence, regional R4D organizations, development organizations, civil society 78

84 organizations and the private sector. The CGIAR Centers and other international research organizations will provide methodological and conceptual backstopping for the integration effort at Action Areas. The types of research outputs to be delivered include decisionsupport tools, analyzing system trajectories, market-nrm relations, policy analysis and partnership approaches in a comparative basis. In addition, capacity-strengthening activities will be planned and implemented as needed with partners. A communications plan will be developed and implemented in consultation with all the partners in each of the Action Sites. Global partnership. Strategic partnerships will be operationalized with a long-term perspective, particularly to deliver the SRT1 outcomes dealing with the Global Synthesis and the release of IPGs. Initially, the main partners will be CGIAR centers and other international research partners such as WUR, AVRDC, and icipe, interested in a global learning process for the humid and sub-humid tropics. As rightly pointed out in the SRF document, The CGIAR is one of the few institutions that can provide a bridge between the local and the global levels. This is of paramount importance when dealing with topics such as climate change, which requires widespread interventions for mitigation and adaptation. The research needs for most of the Strategic Research Themes will require partnership with Advanced Research Institutions such as universities and other national government entities, who have research skills and focus that complement that of the main international partners and can support PhD training. Partnership with other CRPs. Another important type of partnership for output delivery will be among CRPs. Section 9 describes the types of interactions across CRPs and envisions that Humidtropics will use research outputs from other CRPs and create demand for new ones and provide feedback about relevance and applicability of their research products. Therefore, Memoranda of Understanding will become formalized among CRPs in order to coordinate action, particularly within overlapping Action Sites and Action Areas Mechanisms for facilitating partnerships At the Action Site level, partners will participate through R4D Platforms, first building on existing relationships but later broadened to include new and more diverse sets of partners, depending on the Situation Analysis and priority interventions identified under SRT1. These platforms must include partners along the research-to-development continuum that focus on: 1) identifying best practices and interventions for known solutions to priority challenges (e.g., resistant varieties to a known disease); 2) implementing strategic, adaptive, and applied research to address constraints to system intensification where solutions are not immediately at hand; 3) evaluating the performance of new practices and approaches and monitoring their sustainability; and 4) addressing new constraints and challenges as they arise, including those generated by better understood drivers of change. The local partners will participate in result-oriented planning sessions and common M&E activities, including standard data collection schemes. R4D Platforms (R4DPs) differ from Innovation Platforms, as employed in the Sub-Saharan African Challenge Program, in the sense that the latter rely on existing solutions to identified problems and build on what is already known. Humidtropics will also build on existing knowledge but also actively engage in research activities leading to improved productivity, profitability and resource use efficiency, with particular focus on increasing institutional effectiveness. The current SSA-CP pilot learning sites will transition from innovation platforms to R4DPs by the third year of Humidtropics. This will be the key focus for integration of the Challenge Program into the CRPs where lessons learned and evaluation of outcomes and impacts will be part of the inherited research focus. In this way, R4DPs will also serve as a primary source of information for system intensification within the Action Areas. 79

85 There will be specific efforts to facilitate exchange of lessons among platforms within Action Areas, regions and also at global level, so that a process of synthesis of experiences can be promoted, documented and shared at wider scale. An additional function of the R4DPs will be to integrate efforts from other CRPs according to the needs of the specific projects to be implemented at Action Sites Partners for delivering research and development outcomes Achieving significant development requires partnerships with a medium- and long-term perspective, for which the Humidtropics program will play a catalytic role. Table 12 presents the types of partnerships for achieving these outcomes at local, regional and global levels. Within SRT3, partners ensure that research outputs from SRT2 are scaled up to the Action Area level and contribute to poverty reduction and sustainable natural resources management. The role of local and regional partners in this endeavor will be more important than that of CGIAR centers and other international research organizations. Outcome partners in Action Sites. The participation of local development organizations from both the government, non-government and private sectors are required to achieve improvement in system productivity while improving natural resources, enhancing gender equity and empowerment of local organizations, and strengthening value chains at Action Sites. These partners will be actively engaged in transforming research outputs into development outcomes. CGIAR and other international research centers will participate to extract lessons and best practices and support monitoring and impact assessment. Outcome partners in Action Areas. These same partners will participate in the Action Areas, but partnership also incorporates regional development organizations and businesses, particularly farm input manufacturers, commodity buyers and food processors. This expansion facilitates regional interaction, sharing of lessons about development outcomes and impacts among R4DPs, so that regional policies are strengthened and investment in developmental research is increased. Outcome partners at global level. Partners with global responsibility, mainly CGIAR and other international agricultural research centers, together with global development organizations will participate in the process of extracting lessons and synthesis from the specific outcome oriented interventions. In this way, lessons will be useful for designing global and regional investment policies for improving the humid and sub-humid tropics. In addition, partners will adopt common agendas relating to rural poverty reduction, food and nutritional security, and the enhancement of natural resources, which will better position partners to cope with climate change and future food demands. 80

86 Table 12. Candidate partners active within Action Sites, Action Areas or at global level. Partners and emphasis of intervention Category Action sites Action areas Global level Farmer organizations Public sector ministries and line departments Agricultural Extension Services; Voluntary organizations and NGOs Regional & National Agricultural Research Organizations and Platforms National universities and institutes of higher learning Private sector Regional/Global organizations CGIAR and other international agricultural research organizations Advanced research Institutes (ARIS) and universities from developed countries Development Investors National Farmers Association of Malawi, Conservation Farming Unit of the Zambia National Farmers Union, Kenya Federation of Agricultural Producers, Kenya Soybean Farmer Association (and similar commodity groups) Ministries of Agriculture & Forestry, Rural Development, Health, Education and others International Federation of Agricultural Producers (IFAP) Governmental services often offered through the Ministry of Agriculture but in other cases privatized International: CRS, Heifer Int.,CRS, World Vision, Care, Oxfam, Plan Int. Regional: Inades Formation (Africa), CONDESAN (Andes) National NGOs:action area specific, such as Diobass (DRC), Action de Sud (Rwanda) and ARDAP (Kenya), FIELD Indonesia Foundation; KEUMANG, BRAC Bangladesh, BAIF, Pradhan, National: EMBRAPA (Brazil); Indian Council of Agricultural Research (ICAR); PCARRD (Philippines); AARD (Indonesia); KARI (Kenya), NARO (Uganda) IIAM (Mozambique), CARBAP, INERA, ISAR, ISABU, IRAZ, UCB, ICTA (Guatemala), DICTA (Honduras), INTA (Nicaragua), CENTA (El Salvador), Corpoica (Colombia), INIAP (Ecuador), INIA (Peru), INIAF (Bolivia), NAFRI (Lao PDR); WREA (Lao PDR); SFRI (Vietnam); VAAS (Vietnam); MOWRAM, Indonesian Vegetable Research Institute; Assessment Institute for Agricultural Technology; Agricultural Extension Service; Indonesian Tropical Fruit Research Institute (ITFRI);Indonesian Center for Horticultural Research and Development (ICHORD). Engineering Department Regional: CATIE, Amazon Initiative, ASEAN-AVRDC,IICA - Interamerican Institute for Ag Sciences, FARA, Makerere University (Uganda),University of Nairobi (Kenya), University of the Philippines Los Banos, University of Costa Rica, University of Sao Paulo, National University of Rwanda, Zamorano University (Honduras) Small and medium size enterprises; large agribusinesses; traders and transporters; in various Action Areas. For example, Semente Perfeita Seed Company (Mozambique), Dengo Seeds (Malawi), Western Seed Company, Smart Logistics (Kenya), Lapanday Fruits Corporation (Philippines); Nasuntara Tropical Fruits (Indonesia) Regional private-sector agribusinesses organizations. Syngenta (East Africa), MEA Fertilizers Ltd.(East, Central and Southern Africa) Cocoa and coffee industry FARA, AGRA, FAO, ASARECA, CORAF/WECARD, SADC-FANR APAARI IITA, ILRI, ICRAF, Bioversity, IWMI, CIAT, CIP, ICIPE, AVRDC, IRRI, AfricaRice, WorldFish, CIFOR, CIMMYT CIRAD, KUL Belgium, University of Florida, Stellenbosch University, University of Santa Barbara, University of Hohenheim, Stirling University, WUR, Michigan State University USAID, DFID, BMZ, EU, FONTAGRO, B&MGF 81

87 6.3 Partnership Management Humidtropics specifically allocates resources for partnership formation and facilitation to ensure a common vision of objectives and outputs as well as the co-creation and dissemination of solutions. It is expected that CGIAR and other international research organizations will play the role of facilitators for partnership formation and function through the following steps: 1. Partner selection. The choice of the partners will be based on expertise and experience for either research outputs or outcomes at local, regional or global levels. As part of the SRT1, the program plans to undertake mapping to determine who are the main actors and their activities in the Action Areas and their potential contributions and benefits. Some specific criteria related to the selection of partners include: 1) where possible each Action Site should have at least one partner from each of these categories; farmer organizations, NGOs, extension workers, and the private sector, 2) the partner should already be active in the Action Site, 3) the partner should be willing to disseminate technologies recommended by the program 4) the partner should be willing to participate in capacity development activities. A non-exhaustive list of candidate partners also appears in Table Partner eligibility. Where possible each candidate partner that will be receiving funds from Humidtropics partner must provide 1) a copy of their organization s registration, charter and by-laws; 2) a brief narrative description of their mission and ongoing activities; 3) a list of their governing board and key officers; 4) recently audited accounts of their organization; 5) a description of recent grants made by other organizations or units of government to their organization; and 6) details of their bank account. 3. Partner engagement. Overlapping objectives and strategies, and complementary resources (human, financial and expertise) will be criteria for initiating partnership formation at the Action Site level. Partners will have to invest institutional resources to participate in the platforms. The relationship will be formalized through Memoranda of Understanding specifying roles and responsibilities. Partnership formation and operation should be flexible for facilitating both engagement and disengagement. 4. Participatory planning. An important building block of partnerships is clarity in expectations and transparency in roles, responsibilities and resource allocation. We distinguish different types of partnership roles. There will be partnerships more oriented to research on component integration, system integration or market development, and others more oriented toward policy and capacity building. Planning meetings will be essential for the efficient performance at each Action Site although some planning for developmental activities and capacity building is better conducted within Action Areas. 5. Partner communication. All partners must regularly and reliably communicate, and partners that persist in communicating poorly are by definition poor partners. Communication will occur at two levels, among participants through program reporting, and among a wider outreach circle. Cost-effective means for internal and external communication include websites, electronic fora and networks, regular face-to-face interactions and teleconferences. CGIAR partners must play a key role in facilitating communication at all levels of program operations. 6. Joint M&E. All partners will be actively involved in monitoring and evaluation of activities in order to gauge progress, adjust methods and stimulate joint learning. This will also provide an opportunity to detect potential conflicts among partners and propose solutions. This involvement reinforces partners rights to program resources but also assures their awareness of program obligations. 7. Co-funding for increased coverage. Resources indicated in the program budget clearly cannot improve the lives of all persons throughout the humid tropics, and means of amplifying and replicating program success must be found. Therefore, the program 82

88 will seek co-funding strategies with other research organizations, large development organizations and governments interested in reducing poverty and improving natural resources management. This may be achieved through buy-ins of the program s IPGs. 6.4 Communication Influencing behaviors Humidtropics views communication as a means to change people s behaviors and as a research area in its own right. Communication becomes more effective when what is implicitly assumed becomes explicitly stated. Knowledge sharing and communication activities are intrinsic to the program and directly related to partnerships, capacity development adoption pathways. Communication tactics will be based on audits of the current Knowledge, Attitudes and Practices (KAP), communication methods and communication technology and relate those to the envisioned adoption of innovations and technologies. For communication and information sharing, we will consider the following: A major impediment to innovation in integrated systems relates to inadequate knowledge and information flow between different elements of the various systems. This includes both lateral flows between actors at a given scale and vertical flows from local to higher scales and back again. An important outcome of the program will be enhanced knowledge and information flows through social and technological networks at different scales, to facilitate timely decisionmaking, enhance innovation and improve its uptake leading to impact. Lessons on how to foster innovation through better information flows among actors in networks and between scales will be obtained from across the Humidtropics and shared with the wider R4D community including other CRPs as part of both the Global Synthesis and M&E framework Knowledge exchange The communication strategy will be embedded within the research conducted as part of the program (SRT1). We will include a knowledge exchange element whereby lessons about the processes are captured and made available beyond Humidtropics Action Areas. This will capture, document and extract lessons within Action Sites, and also across collaborating CRPs. A set of communication processes and tools will be used to maximize interaction, learning and knowledge sharing among the participants, partners and where applicable the larger society. Lessons emerging from this program will be shared with interested parties, including development agents, donors, policymakers, and the R4D community among others. A number of principles define the communication and knowledge sharing activities of the Humidtropics program: A commitment to early sharing of soft outputs (activity reports, meeting reports, presentations at local meetings, videos, etc) through publicly accessible and permanent web spaces and services, including a global electronic program repository. Establishment of mechanisms to aggregate stories and lessons emerging from Action Areas through a Humidtropics Portal, blog postings and other communication tools. A commitment to use open access journals and communication tools (e.g., blogs, Facebook, wikis), and open source licenses that encourage the wider reuse of program discoveries. For learning within Action Areas, a full range of communication opportunities will be offered through the establishment of Action Area Learning Groups. Although face-to-face interactions at least once per year would be beneficial, other means of communication using telephone and Internet may be more feasible. Reports, meetings and presentations will be regularly posted into the Humidtropics e-repository. 83

89 For program learning and beyond, other mechanisms are useful including the establishment of a Global Humid Systems Knowledge Exchange Group with representatives drawn from all Action Sites. Humidtropics communication spaces and tools serve as outreach platforms to wider communities. The aim is to maximize global harvesting and reuse of the public goods generated by the program and also to provide opportunities for other people to interact with and challenge the program s findings Program communication services As part of the above it is important to promote Humidtropics and ensure transparency and trust to further develop internal and external relations and communications. The following approaches will be developed and supported: Advance Humidtropics identity. Actions involve using website as a platform for information exchange and practicing virtual branding based on the Humidtropics informational and promotional releases, and scanning the media environment for potential reputation enhancing or reducing messages. Improve public awareness. Actions include media releases at all levels and posting a regular e-newsletter. Support to partner relationships. Actions include program and project briefs, online collaboration through posting planning results, survey and research findings and meeting reports. Support to investor relations. Through absolute and timely compliance with all technical progress and financial reporting requirements and the development of investment information packs. Enhance Humidtropics knowledge management. Through development of an online, open document repository, online bibliography and media library. Communication efficiency and effectiveness must be incorporated within the M&E framework. 6.5 Capacity Development Improving institutional effectiveness requires capacity development in terms of knowledge, skills, and resources. Capacity development is required beyond technical expertise, but must also include new approaches to knowledge sharing, scaling up and fostering small enterprise development. This effort will begin with an assessment of capacity building needs among research partners in the Action Sites and continue with development partners in the Action Areas and in both cases include skills in M&E. Capacity development must ultimately strengthening partner capabilities in innovative technology development, stakeholder participation and information exchange. Developing capacity to do research on production systems focuses more on individuals and households while developing capacity for scaling up the results of such research focuses more on institutions and their services. Each requires quite different strategies. Some partners may participate in both processes, such as universities for example, but more often, partners will be different for the two processes. The role of policies in establishing conducive, supportive and sustainable environments for capacity development must also be examined and a capacity response formulated. Humidtropics envisions the following capacity development outputs: Capacity of individuals and regional and national organizations strengthened to deal with emerging issues from system intensification, Emerging issues from system intensification incorporated within curricula of agricultural universities and other training institutions, Greater gender sensitivity embodied into governmental, education, non-governmental and grassroots organizations and a framework for gender inclusion included within their capacity building programs, 84

90 Enhanced capacity of individuals, organizations and businesses to promote and practice improved post-harvest and value addition technologies, Stronger abilities to adopt and update new approaches to knowledge sharing, scaling up and fostering small enterprise development, and Targeted provision of training and information leading to increased agricultural lending by the private sector, These outputs will be delivered through many activities discussed below that address both the need to strengthen capacity to conduct multi-disciplinary systems research and the need to develop capacity for scaling up results.it is implicit that research on agro-ecological systems, knowledge and information, markets and policy must cut across many sectors. Too often, these sectors are institutionally segregated within research organizations, ministries, universities and NGOs. Strengthening the capacity of these diverse interests to undertake research on complex agro-ecological interactions and agricultural value chains is central to the implementation of this program. A broad range of formal and informal learning methods will be used to strengthen such capacity, including conventional training on tools and methods as well as participatory action research that fosters learning by doing. Examples of these capacity building activities include: Diagnosis of current capacity for systems thinking and action across stakeholder groups and assessment of partner organizations as innovation brokers, Training in participatory processes for conducting multi-disciplinary, multi-stakeholder and multi-sectoral research and validation of learning approaches, Workshops to validate tools and methods for system analysis and improvement with different stakeholder groups, Development of online platforms for knowledge sharing of case studies, best practices and training materials, including on novel ways of enhancing capacity for co-production of knowledge with multiple actors, Training and mentoring through MSc and PhD graduate student scholarships within projects and opportunities for mid-career professional advancement, Commissioning guidelines, syntheses and briefs of best practices in multi-stakeholder research into post-harvest handling, markets, value chains and policies, Continuous assessment on learning, in particular, on system-oriented vs. traditional approaches to capacity building adopted by the program. A related challenge concerns forming strong partnerships that bring proven research findings into wider use. Partners directly involved in the scaling up and delivery of IPGs must be phased in as needed and quickly brought up to speed to ensure effective participation in the SRT3 activities of the program. Similarly, gender sensitivity must be incorporated into priority setting, design and implementation phases of the program. The poor and vulnerable people s capacity for building their capabilities to use the project s public goods will strengthen the innovations produced. For example, experience and institutional learning that has already been gained on the role of self-help groups in creating markets will be applied to strengthen the capacity of poor women producers to enhance their livelihoods. Finally, Humidtropics offers real opportunity for enhancing institutional learning and cross-fertilization of ideas. The cross-regional nature of the program implies that capacity development will benefit from both technical and non-technical lessons in different regions, as well as drawing from outputs of other CRPs; combining them in the Action Areas and beyond to empower stakeholders with knowledge and skills greater than the sum of the component parts. Mechanisms will be developed for capturing and sharing innovative learning processes and experiences across components and feeding these back into the program for continuous learning and improved performance. 85

91 7 GENDER RESEARCH STRATEGY 7.1 Rationale Any type of intervention that ignores gender issues may in fact increase vulnerability; hence, gender research and capacity building are integral within Humidtropics. Men and women have individual and shared roles, but women have critical productive, reproductive and community roles, often making them more vulnerable (Gladwin et al., 1997). The rise of women s studies and advocacy on women s rights over several decades has led to investments in women s empowerment and corresponding improvements; but in general, too many rural women remain disadvantaged due to male-bias and repressive cultures and policies that do not account for gender differentials. In contrast, where gender issues are appropriately addressed, sustainable and productive livelihood outcomes are likely. Women play important roles in tropical agriculture by cultivating and marketing crops and managing mixed enterprises while, at the same time, attending to family and social obligations (Ashby et al., 2008). Ironically, the benefits from their efforts are not always realized at the personal and household level due to entrenched inequities between gender relations tolerated by culture and society. Women face particular challenges in accessing information, extension, advisory services and education (Saito and Weidemann, 1990), as well as in owning or acquiring land, assets, and technology (Ezumah and Domenico, 1995; Kaaria and Ashby, 2001). But despite the lack of attention to their needs, women have developed organizational capacities to learn and support each other even when there is no direct economic benefit. Women comprise an average 43% of the agricultural work force in developing countries, ranging from 20% in Tropical Americas to 50% in Eastern Asia and Sub-Saharan Africa. Yet, women have less access than men to agriculture related assets, inputs and services. For example, only 5% of current agricultural extension efforts and resources are directed to women and they invariably earn far lower wages than men for equivalent work. Had they enjoyed the same access to productive resources as men, women could boost yield by 20 to 30%. This gain in production could lessen the number of hungry people in the world by 12 to 17%, besides increasing women s income (FAO, 2011). Studies show that when women control gains in income, they are more likely to be spent on food and children s needs. By investing more in women, we amplify benefits across families and generations. There is consistent and compelling evidence that when the status of women is improved, agricultural productivity increases, poverty is reduced, and nutrition improves. Reducing gender inequality and recognizing the contribution of women to agricultural productivity are therefore critical to achieving global food security. Women s participation has been sought in agricultural development interventions but there are considerations that require serious attention. Participation in meetings and training alone does guarantee positive results in terms of empowering women, and in some cases, women are more burdened as they take on new roles that are either not traditionally expected or externally imposed on them. An electronic consultation recently asked for the level of agreement on the assumption that R4D programs hardly address the specific needs of women and of the 91 respondents, 25% disagree, 47% partially agree, 20% agree and 8% had no opinion on the matter. The responses are inconclusive, indicating that more research needs to be done to verify the assumption. More conclusive were the responses to the assumption that The role of women in household decision making processes needs to be improved" with 79% agreeing and only 3% disagreeing. The key objectives of the Humidtropics gender strategy are to: 1. Ensure that men and women, young and old have equal access to assets, inputs, and technologies, including land and other productive natural assets, extension services, 86

92 financial services, agricultural inputs, and the knowledge to enable them to participate in and obtain appropriate returns from the agricultural system. 2. Expand the involvement and participation of women in decision-making at all levels of the program including management and technology development. This effort will help ensure that women have a voice so their contributions can be recognized and their needs better met. Men need to be engaged in these change processes to achieve sustainable outcomes for the entire community. 3. Ensure the interests of women and men are reflected in activities, interventions and projects of CRP1.2, including efforts to help ensure the participation of women, establish reasonable targets for the participation of women, and monitor and evaluate program impact on both men and women. On the basis of the above, this component aims to improve the quality of women s participation and increase benefits to women from research and development processes and reduce gender inequity and will be addressed through 1) gender mainstreaming and integration into the research and development processes, 2) strategic gender research and 3) institutional support and collaboration. 7.2 Mainstreaming and Integrating Gender Gender will be mainstreamed into all research and development processes, and decisionmaking structures of the program. A key approach to integrating gender into Humidtropics research and development agenda will be to use gender analysis to inform strategies to address gender inequalities. The gender analysis will take into account women s and men s lives and experiences, needs, issues and priorities and the fact that these are different. These issues and priorities will also differ based on their social position and age. Gender analysis will focus upon the following issues: 1. The key gender constraints and opportunities in the crops, livestock and other system interventions of interest, 2. The gender division of tasks and responsibilities around key crops and livestock, 3. Access to, ownership and control over productive resources as well as the constraints faced by men and women in accessing improved technologies, services, inputs and markets, 4. Intra-household decision making and how these influence agricultural production and marketing, 5. Men and women farmers preferences for technologies, inputs and services and 6. Opportunities that the research and development processes provide for reducing gender inequalities Results of the gender analysis will be used to develop strategies to better address gender based constraints in different areas of the Humidtropics research and development process. The outcome of this gender analysis will improve the understanding of gender roles in the productive, reproductive and community development spheres, better understanding key institutional, cultural and policy contexts that ingrain gender inequity at each Action Site, and identify gender-sensitive policy interventions appropriate at different scales and settings. Understanding ways in which women s roles reinforce social support systems and identifying options that enable women to improve their livelihood and living conditions allows for better targeted interventions for both men and women. Using the results from the gender analysis, gender responsive approaches and gendertargeted interventions will be implemented as part of the interventions. For each set of activities or interventions, gender has been integrated in nine important ways. 87

93 1. Problem and Context Analysis: In each of the Action Sites, a Situation Analysis is conducted that includes the identification of social, legal and cultural barriers that could prevent women from active participation in the various activities of the project. 2. Gender responsive objectives: Program targets must show the expected outcomes of the interventions for both men and women as well as marginalized groups. 3. Methodology and implementation approaches: Gender responsive approaches are required for different phases of the implementation process. These include technology development and evaluation with men and women, market analysis and development reflecting the priorities of men and women, and evaluation methods that focus on impacts of technologies and other innovations on men, women and on reducing gender inequalities. 4. Monitoring and evaluation: For all the Humidtropics interventions, the monitoring and evaluation plans will include gender-specific outcomes and targets; gender disaggregated indicators and data collection tools. Some of the specific outcomes are: reduction in gender disparities in access to inputs, services and technologies; women s empowerment for decision making and income management, increase in productivity in men and women managed farms, increased income for men and women. As a result gender disaggregated data would form an integral part of indicators developed and the M&E data collection protocols so as to acknowledge the roles of gender norms and inequalities and to develop actions that will adjust to and compensate for them. 5. Budgets: Specific budgetary allocation will be clearly allocated out for specific gender research. Where gender is included as part of other activities, the proportion of the budget to be used for gender will be specified. 6. Work plans: Gender activities will be reflected in the work plan and the necessary expertise identified from within Humidtropics and its partners for implementation of those activities. 7. Gender expertise: Humidtropics will appoint a gender focal person, who will lead the mainstreaming of gender into research, lead the strategic gender research and be the link person with the other CRPs. While the gender focal person provides the specialist gender expertise, it will be the responsibility of all the researchers and partners to mainstream gender in their activities. 8. Capacity Building: Capacity building will be targeted to both men and women to promote complementary skills and interventions for systems improvement, and to facilitate equitable innovation processes and outcomes. Women need to be empowered to leverage their roles to invest in, and gain from, agriculture production and derived activities. As a result, women will be able to attain organizational positions and power to influence institutional effectiveness. 9. Impact Assessment: Indicators that provide evidence of gender inequalities and asset disparities will be developed and measured at the household level for both male and female headed households in order to capture gender differences. Baseline and impact and evaluation studies will also focus on the effects of new innovations and new institutional arrangements to empower women, with a key focus on gender division of labor and appropriation of resources and benefits within households and communities. Household level impact study surveys shall be designed such the views of both men and women are heard, not the views of only the household heads. There will be gender representation in all the decision-making levels of the Humidtropics. The gender expert will be part of the Program Advisory Committee in order to ensure integration and commitment of gender. At the implementation level, the program will set targets (with the minimum internal target of 30%) for women s representation in the different implementation structures and activities. 88

94 7.3 Strategic Gender Research To achieve the above objective, funding will be allocated for strategic gender research across the regions to allow for regional analysis and generate robust knowledge and experience in addressing gender issues. The research will focus on developing, testing and evaluating approaches for increasing women s participation and benefits from research and development processes; analysis of what technological, institutional and market innovations are most appropriate for reducing gender inequalities and equity in women s representation at various levels of decision-making, as well as in benefit sharing. This research will take an experimental design approach and will be carried out across Action Sites in order to ensure validity and replicability of the results. Other issues for strategic gender research will be continuously identified within the program and in collaboration with the gender expert at the consortium office and in the gender network that will have membership of gender experts from the other CRPs. 7.4 Institutional Support and Collaboration It is critical that an enabling environment be provided to enable gender mainstreaming and research and linkage to policy and practice, including relations with the gender research teams in other CRPs. The Humidtropics leaders will act as gender champions and will seek commitment and support from all partners to ensure that gender researchers are given institutional support. A virtual support group of gender researchers will be created and regular communication, knowledge sharing and capacity building undertaken. Gender research capabilities within Centers involved will be enhanced and knowledge sharing expanded through collaborative undertakings with gender expert groups such as the WOCAN. Whenever appropriate and necessary, collaboration with these groups will be supported to advocate for policy change at the national or international levels, support capacity building of R4D projects, improve gender M&E indicators and contribute to ex-ante and ex-post outcome assessments. A gender guide will be developed for Humidtropics to offer direction and recommended best practices that will be used in all the program activities. It will also include advice on M&E indicators and organizational mechanisms such as collective action that allow women to participate in and benefit from program activities. Because women in most parts of the humid tropics are disproportionately below the poverty line and at the same time are central to decision making in agricultural households, understanding how women can improve control over income streams arising from systems-oriented technological innovations will be a significant component of the impact monitoring. 89

95 8 INNOVATION Research on production systems in the CGIAR represents the evolution of understanding of and the changing context for the improvement of agricultural production. The establishment of the CGIAR in 1972 divided its mandate between two commodity centers, IRRI and CIMMYT, and two centers that focused on production systems in the lowland tropics, IITA and CIAT. However, the commodity centers gained prominence with the success of the high yielding rice and wheat varieties in the irrigated lands of Asia and Mexico, based on the advances of breeding programs begun before the founding of the CGIAR. By 1978, CIAT was essentially four commodity programs, organized around germplasm improvement. Ironically, these four commodities, cassava, beans, rice, and tropical forages had little potential for integration at the level of farming systems in the tropical lowlands. While IITA maintained a focus on moving slash-and-burn agriculture to sedentary, intensified systems, much of the research was organized around breeding programs as well. It was soon apparent that the Green Revolution would not be reproduced in rainfed agriculture, particularly in terms of the wide adoption of improved varieties (Okigbo, 1990). Instead, changes in farmer management and agronomic practices were required to achieve sustainable increases in productivity from improved varieties. Farming systems research (FSR) became the framework by which to test and adapt new varieties and management practices to farmers local conditions (Chambers et al., 1989). FSR led to the development of methods for what eventually would be known as adaptive research. In rainfed systems this was seen as a crucial link between applied, on-station research and extension, and FSR was designed to provide feedback to applied research teams. FSR went through various stages as it moved away from research on farming systems per se, first simplifying to commodity research within a farming systems perspective, then to an increased emphasis on farmer participation, and finally to more of a focus on extension than on research methods, in many ways a forerunner of Farmer Field Schools (Braun et al. 2000). FSR was something of a misnomer, as the focus was essentially on cropping systems and purely in a testing modality. Moreover, in this period market conditions were taken as a given, and there was no integration of market and technological interventions. While adding adaptive research and participatory farmer research to the research toolbox, FSR never evolved into a research program on farming systems but rather evolved into methods for local adaptation and extension. The 1990s were the period of expansion of the CGIAR into natural resource management, attempts at developing eco-regional programs, and the development of cross-center, system-wide programs organized around specific issues of integration at a farming systems level. This was more a period for exploring the potential for inter-center research collaboration, and much of this work was organized around sub-systems such as foddergrain trade-offs in crop-livestock systems, integrated soil management, or integrated pest management. This was also the period of moving away from core funding to project-based funding, which created more competition than collaboration between Centers. In the end, system-wide programs operated at the margins of the Centers mandates and were never central to their core research programs. Unfortunately, these programs and initiatives were never able to exploit the complementarities across the CGIAR that could underpin a dedicated research program on production systems. This last decade leading up to the CGIAR reform was the period of experimentation with another cross-center institutional arrangement, the Challenge Programs. Two of these focused on exploiting the potential of molecular biology in breeding programs, HarvestPlus and the Generation CP, while the other three had a particular systems focus, namely the Water and Food CP, focused on river basins, the Climate Change CP with a focus on 90

96 farming system adaptation and the SSA-CP with a focus on testing innovation platforms as a vehicle for integrating productivity, NRM, and market innovations. None of these programs had a focus on research on production systems, per se, nor on exploring the potential for the synergistic integration of crop, livestock, and management options within a dynamic process of system change. This brief history is testament to the difficulty of organizing research programs on production systems and outlines the extensive base of systems learning within the CGIAR on which Humidtropics will build, including the reasons why a full-fledged systems research program never developed within the CGIAR. Three factors provide the pre-conditions for the success of the Humidtropics program over the false starts of the past. Firstly, the CGIAR reform has provided the incentives for inter-center collaboration at the same time as reducing the transaction costs in such collaboration. Success in this program builds on the sub-system research across the different Centers and provides a platform for expanding the potential application of that research. Secondly, Humidtropics exploits the dynamic interplay of markets, technical change, and farmer investment in the natural resource base. This is within something of a stage theory of system intensification and exploits different critical system interventions under different market and farm resource endowments. Finally, Humidtropics builds on a range of information technologies and systems research methods that were not available in the past, including the integration of spatial data and analysis in location-specific research, the development of data standards in systematic panel surveys, the analytical methods underlying ecosystem services, trade-off methodologies, advances in system modeling and simulation, and the expanding use of evaluation methods in agricultural programs. Humidtropics could not have been implemented before the current CGIAR reform and in that regard epitomizes the potential of implementing integrated, systems research within the new CGIAR. At the same time, there is very little ongoing research that underpins the design of this program other than the Sustainable Tree Crops Program in West Africa (Box 11) and CIALCA work in Central Africa linked to the Lake Kivu SSA-CP, both operating through a set of integrated research activities that come closest to the vision of Humidtropics. CRP1.2 research in the other Action Areas will build on ongoing research but these activities will have to be comprehensively redesigned over the first three years to meet the requirements of the overall program design. In this way, CRP1.2 is in its most basic sense a test of the CGIAR reform process. The core areas of research innovation consist of four principal areas: Development of a conceptual structure and methods for research on production systems within a framework of system intensification, Research on scaling up integrated market, productivity and NRM interventions and how this is translated into impacts on development outcomes, Developing sampling and data structures at different scales that allow for analysis within and across Action Areas leading to global synthesis and production of IPGs and Design and testing of institutional innovations that facilitate lower transaction costs in the delivery of integrated approaches and particularly target poverty and gender equity outcomes All of these tasks must be performed within an evolving and testable conceptual framework across a set of Action Sites systematically selected within that overall conceptual framework. Success also relies upon the development and refinement of research methods at each scale with the potential for integration across scales, and the design of a set of standards for data collection on farming systems and agricultural change in tropical ecosystems. While producing its own set of innovative research outputs, Humidtropics will as well provide a research platform and methods for the global change research community, 91

97 poverty research, and research on ecosystem services as mediated through changes in agricultural systems. All this reinforces the need for systematic design and not merely aggregating current Center activities in the Action Areas. Box 11. Intensification in action: The Sustainable Tree Crops Program. The Sustainable Tree Crops Program (STCP) is a collaborative public-private partnership between the global chocolate industry, national and international research institutes, national extension services, and various donor agencies for the development, testing, and scaling out of innovations targeting smallholder cocoa production systems and marketing services. The program is active in Côte d'ivoire, Ghana, Nigeria, and Cameroon which together account for 70% of global cocoa production. The conceptual framework and priorities of the program were benefited by a comprehensive producer baseline survey at the start of the program in 2001/2002. Major achievements of the program include the development and testing of a farmer field school (FFS) extension approach for the integrated management of cocoa pest and disease. The program also developed and tested institutional innovations with producer organizations that included collective marketing and the provision of improved planting materials. An accompanying FFS beneficiary database provides the sampling frame for the monitoring and evaluation of program outcomes of the basis of systematic random sampling. Many policy interventions have been evaluated in the course of these investigations and policy dialogues to improve or upscale policy interventions are ongoing. The program also benefited conceptually from the participation of key research personnel in the Alternatives to Slash and Burn program of the CGIAR, which has developed innovative approaches for the evaluation of environmental and economic trade-offs for a range of production systems across the global pan tropics. These approaches have been used to evaluate land-use change scenarios and productivity. Gockowski and Sonwa (2011) estimated that had intensified cocoa technology, already developed in the 1960s, been pursued in Cote d Ivoire, Ghana, Nigeria and Cameroon that over 21,000 km 2 of deforestation and forest degradation could have been avoided along with the emission of nearly 1.4 billion t of CO2. Addressing the low productivity of agriculture through research and extension should be one of the principal objectives of REDD + Climate Mitigation Programs in the humid tropics (Gockowski and Sonwa, 2011; Gockowski and van Asten, 2012). An analysis of productivity at the farm gate in Ghana has revealed significant positive impacts achieved by the Ghana Cocobod High Tech program. The major effect on productivity was the increased adoption of fertilizer and insecticides in the Western region of Ghana. Ghana has more than doubled its cocoa output in the last eight years mainly as a result of system intensification. Future sources of productivity growth are likely to include improved cocoa hybrids which account for only 8% of bearing cocoa but were four times as productive as the local planting materials most commonly used by farmers. 92

98 9 INTERACTIONS WITH OTHER CRPS Humidtropics is very clear about the boundaries that separate its work from the other CRPs, either in terms of focus, i.e., production systems, or in terms of agro-ecology, i.e., humid and sub-humid tropics. However, as an integrative CRP, Humidtropics draws on the work of all of the other CRPs, if only for shared methodologies and shared knowledge. Humidtropics serves as a R4D laboratory in the humid and sub-humid tropics and must necessarily engage in shared learning and system-level research with the other CRPs. From its very onset, it will proactively solicit collaboration with these CRPs through convening a collaboration-planning workshop, appointing specific liaison staff to work with other CRPs and allocating budget for this purpose. Figure 18 illustrates this collaboration and more specific details on entry points for collaborative action appear in Table Why Interaction with Other CRPs is Needed Humidtropics is by nature and design, a mechanism to integrate research results from the strategic objectives of other CRPs according to their contribution to food and nutritional security, poverty alleviation and natural resource integrity. Moreover, Humidtropics provides an opportunity to develop the new competency in production systems as highlighted in the SRF, addressing system constraints and opportunities in very different production systems across Sub-Saharan Africa, tropical Asia and tropical Americas. Over time, this program will also provide feedback from experimentation on system interactions, priorities, needs, and applications relating to other CRPs. While other CRPs focus on crop, livestock or theme-specific methods, models, tools and technological components (Error! Reference source not found.), Humidtropics will draw on such research outputs and integrate them into its production system and NRM research within SRT2 in those Action Areas where such outputs are relevant. Figure 18. Relationships between Humidtropics and other CRPs Knowledge sharing and capacity building dimensions are other important integrating functions within Humidtropics, facilitating the application and use of outputs generated from other CRPs for stakeholders on the ground. Table 13 summarizes the potential interactions between Humidtropics and other CRPs, but more definitive collaboration will emerge from the planning workshop. Specific interactions are also highlighted in the descriptions of the SRTs (see Section 4). Below we also highlight some of the key interactions with specific CRPs to provide further illustration of the anticipated starting points and subsequent evolution of such interactions. Box 12 describes specific elements of the interaction between Humidtropics and CRP7. Humidtropics is one of three CRPs that addresses systems complexities and in this respect, responds to the need for the adding together of 93

99 research for development in specific production systems. In this respect, considerable opportunity exists for sharing and learning from approaches to complex production system challenges that combine both institutional and biophysical dimensions. Within three years, it is important that streamlined approaches relevant to all three regimes (dryland, humid and sub-humid, and aquatic) be established leading to more rapid identification of most promising best practices. The innovative approaches to strengthening institutional effectiveness pioneered within Humidtropics will also have a direct application to CRP1.1. In addition, complex livelihood issues in both dryland and humid systems involve issues of trade-offs in biomass production and utilization best analyzed using common methodologies (see Humidtropics and CRP1.1 have a number of participating Centers, and indeed individual scientists, in common which will be an initial mechanism to foster interaction, sharing and learning. Leaders and/or scientists from these two CRPs will further strengthen this through participation in inception workshops. As the work evolves, it is anticipated that specific cross-crp learning for common approaches as described above would also be pursued. Distinguishing between CRP1.1 (drylands) and CRP1.2 (humid and sub-humid tropics) has been initially based on the simple categorization of below and above a length of growing period of 180 days. While there is no overlap of action areas, the scope for shared learning is enormous. There are many special opportunities for collaboration between Humidtropics and CRP5 because CRP5 is focused on sustaining the environment and natural resource base used across a range of sub-humid and humid zones (Error! Reference source not found., Table 13). However, CRP5 operates at landscape, watershed and basin scales with entry points predominantly through sustainable management of natural resources. Humidtropics will readily absorb these findings to better balance system productivity and natural resource integrity. There are many other examples of how differences in focus can be used to positive ends: CRP5 refines management strategies that protect water and soil resources to ensure both agriculture and ecosystem services. Humidtropics applies these findings towards intensification of agricultural production in a manner that maintains natural resource integrity, particularly through the development of guidelines for Integrated Soil Fertility Management (ISFM). CRP5 deals with component issues of soil fertility and management including the reuse of agricultural by-products that improve soil physical and chemical properties and minimize soil erosion and sedimentation, approaches that readily contribute to system integration within Humidtropics. CRP5 will identify policy and governance changes required for better management of natural resources that Humidtropics must rely upon to scale up proven production systems advancing the livelihoods of the poor and vulnerable. CRP5 examines how agricultural landscapes can be managed and regulated to better deliver critical environmental services while Humidtropics directs these findings to collaborative national institutions within the context of improved production systems so they may better fulfill their service mandates. There is mainly an overlap between CRP5 SRP Rainfed, which represent an opportunity for collaboration in Action Areas in Africa. In America and Asia, there is mainly an overlap between Action Areas of Humidtropics and SRP Ecosystems and SRP Basin. These also represent opportunities to collaborate and share research outputs, research sites, and address scalability issues. The CRP5 SRP on information systems and the outputs of SRT1 can be linked to develop more comprehensive regional and global databases. Humidtropics includes forested areas and therefore intersects with the work of CRP6. CRP6 utilizes a forest cover transition model as an organizing framework that captures the overlap with Humidtropics at various points from degraded to healthy forest ecosystems. The 94

100 greatest opportunity for collaboration or overlap will be in Components 1, 2, and 3 especially as it relates to sustainable livelihoods with tree crop integration and improved productivity and markets for forest or tree products (Error! Reference source not found.). The outputs from CRP6 will address the needs for components in SRT2 activities but will also allow for collaboration on system knowledge and integration in forest-cropland mosaics in relation to SRT1 and SRT3. Box 12. Cross-linked CRP7 example on how a CRP interaction in the area of climate change may operate. 1. Get agreement on goals that serve both Humidtropics and CRP7, with CGIAR Centers and partners. This includes conducting scenario analyses of visions for the future. 2. Standard methods and data collection in Humidtropics on agro-ecosystems characteristics, including land use (e.g., cropping, rangeland), geographical specifics (e.g., land slopes), poverty dimension, cropping patterns, crops and forages grown, livestock specifics, rotation practices, soil specifics (e.g., organic matter, fertility), water availability (e.g., precipitation, wells, access to rivers), market connectivity, value chain specifics, existing analysis on how future production systems may change under climate change. Data collection will be in collaboration with other CRPs at regional level, and will include participatory approaches so that farmer and community perspectives are understood. 3. Sharing data with modeling community. Carrying out of modeling in CRP7 using various climate change and development scenarios to identify possible mitigation and adaptation interventions. 4. Joint analysis, between Humidtropics and CRP7 and partners. Selecting sub-set of scenarios that seem congruent in their predictions. Identifying the possible sets of mitigation and adaptation interventions in terms of food security, poverty alleviation and environmental sustainability (these options may come from any points in the overall food system). 5. Developing and testing options. Humidtropics will test possible options, with co-financing from CRP7. CRP7 will provide the expertise for climate-specific components where needed (e.g., climate risk insurance methods, improved delivery of climate and weather related information for smallholders, mechanisms to enhance access to carbon markets). 6. Multilocation and multi-year trials conducted in benchmark Action Sites, both existing (with historical data already available) and new sites based on site-similarity and analogue mapping of the future production conditions for the target sites (from modeling). This will allow real-term experimentation on future predictions. 7. Joint analysis between Humidtropics and CRP7 and partners. 8. Monitoring and evaluation for learning and improvement to maximize added value. CRP6 also intends to select Sentinel Sites that could overlap with Humidtropics Action Areas or Sites. They intend to select sites as funds permit, thus collaboration with Humidtropics will inform the process with input from SRT1 on exact location for Humidtropics focus. 9.2 Realizing Effective Collaboration Between CRPs Humidtropics will ensure effective collaboration with other CRPs through close planning and liaison through the following mechanism: 1. Humidtropics has identified specific budget items intended solely for CRP collaboration including buy-ins to Action Sites and field activities of common interest. 2. Within the first three months, Humidtropics will convene a Collaborative CRP Action Planning Workshop specifically intended to develop work plans around these sites and field activities. 3. Specific liaison staff within Humidtropics will be identified that are required to implement these work plans and to regularly exchange information and lessons with counterparts in other CRPs. 95

101 4. Program scientists from CRP3 will provide recommendations and germplasm of promising varieties of maize, upland rice, grain legumes, forages, root and tuber crops and banana for assessment in humid lowlands, moist savannas and upper altitudes. Humidtropics researchers will evaluate this germplasm in the Action Areas and provide feedback on their performance. CRP4 will then evaluate the nutritive contributions of these best varieties, including biofortified varieties, and all parties will jointly publish these findings. Similarly, animal scientists will recommend breeds of animals and feeding requirements for use in the different Action Sites. 5. Humidtropics will collaboratively formulate a suite of ISFM guidelines for the humid lowlands, moist savannas and mid-altitude zones, and jointly publish them with CRP5. 6. Humidtropics will request that CRP6 provide tree-crop management guidelines and these will be tested, refined and followed in all Action Sites interfacing with forests and in tree-cropland mosaics. 7. Humidtropics will request guidelines in system carbon measurement from CRP7 and document carbon sequestration resulting from systems integration accordingly. CRP7 will be invited to measure greenhouse gas flux from all Humidtropics Action Sites. 8. The intersection of livestock and environment issues will be explored in Action Sites that overlap between CRP3.7 priority value chains including pigs in Uganda and Southeast Asia, and dairying in East Africa, Central America and India. 9. Other forms of direct and mutually-beneficial collaboration will be formalized at the planning meeting and implemented. 96

102 Table 13. Geographic scope, mutual contributions and mechanisms for collaboration between Humidtropics and other CRPs. CRP CRP1.1 Drylands CRP1.3 Aquatic CRP2 Policies CRP3 3.2 Maize 3.3 Rice 3.4 Roots and tubers 3.5 Grain legumes Geographic scope for collaboration Overlap between dry and moist savannas in West and Southern Africa Joint activities in Zambia, Uganda and Cambodia Humid Tropics Potential linkages Contribution to Contribution CRP1.2 from CRP1.2 Learning from Lessons from similar approaches innovation and to integrated, other institutional complex production strengthening systems bringing mechanisms that together are applicable in institutional and other systems biophysical dimensions. Macro-level policy, ex ante impact, and market assessment for priority setting and impact assessment Methodology for policy analysis, impact assessment, and gender analysis Component needed for testing and adoption in integrated, complex systems Meta databases form Action Sites and action areas Identification of key drivers, system opportunities and constraints, and priority research needs for sustainable intensification, NR integrity, and poverty reduction for further consideration in CRP2 research themes Feedback on components fit into systems and research needs for better adaptation or use including adjustments in breeder s objectives Household type and gender implications for component research Joint Research Strategies for understanding trade-offs in biomass use and management and developing strategies to address complex systems combining diverse biophysical, social, economic science approaches Developing competency across systems CRPs in production systems research Trajectories for intensification; models and measurement methods for system productivity, policy analysis, ecosystem services, trade-offs and other key system intervention parameters Value addition and market acceptability of varieties Action research on cropping system and implication for system integration Input market development at Action Sites for seed systems and other technology needs Mechanisms for achieving integration Joint activities at the level of action areas with potential overlap and synergies (priority given to areas as indicated in first column) Participation of leaders in CRP1.1 and 1.3 in inception meetings in appropriate regions; joint annual and other meetings especially targeted at sharing and learning lessons about implementing and improving research for development approaches Joint activities cofunded at global and action area levels. Joint planning of activities at global, regional, and national level with participation of both CRP2 and CRP1.2 scientists Center or partner level facilitation when both involved in CRP2 and CRP1.2 Center level facilitation when involved in CRP3 components and CRP1.2 Participation of CRP1.2 leader or CRP3 commodity leader in each others planning meetings at global, regional, or action site level, as appropriate 97

103 CRP CRP3.7 Livestock CRP4 Nutrition Geographic scope for collaboration Overlap with priority livestock and fish value chains in Uganda (pigs and fish), Vietnam (pigs) and dairying in Tanzania, India, Honduras and Nicaragua. Focus on two populations a) rapidly intensifying livestock in SE Asia and around cities ( CRP1.2) b) neglected populations (pastoralists, tribal, remote) ( CRP1.1) Potential linkages Contribution to Contribution CRP1.2 from CRP1.2 Information on Strategies for requirements for biomass feed inputs into management livestock value including tradeoffs; chains that may use of influence the natural management of resources in resources relation to (especially biomass livestock from crops) and production; crop choice of crop (eg and forage those with better combinations quantity and quality suitable for of residues) and livestock feed forage varieties.crp3.7 also provides a platform for the evaluation of the feed value of crop residues and cropforage rotations Nutrition-sensitive value chains Food safety in transforming animal source food value chains both to protect the health assets of value chain actors and consumers and to safeguard market access for smallholder producers Assessing and managing emerging disease in intensifying systems Improving production and productivity which leads to increased demands for food safety Joint Research Options for sustainable intensification of livestock production especially in integrated crop livestock systems (smallholder pigs and dairying) that also addresses environmental dimensions (both use of land and water for feed production as well as impacts of livestock on natural resources) in both technical and policy related dimensions. Options for biomass management that take account of trade-offs in use for animal feed or soil fertility Ability of systems to deliver food quality without trading off other outputs Food safety and emerging infectious disease as constraints to rapidly evolving value chains. Food safety is being integrated into CRP3.7 Mechanisms for achieving integration Key researchers participating in both CRPs enabling sharing of approaches and lessons in particular related to balancing demand for animal feed, natural resources and environmental issues. Opportunities for crop by products from CRP1.2 areas to be processed/transporte d and used as feed resources Joint contributions to the overall goal of better performing systems not only in terms of food production but quality, safety and environmental sustainability. Sharing of information and identification of opportunities 98

104 CRP CRP5 Water CRP6 Forests CRP7 Climate Geographic scope for collaboration Regional, benchmark basin and sentinel sites in Andean basins, West, East and Southern Africa, South Asia (Ganges) and SEAsia (Mekong) Overlap in sentinel landscapes in Central America, Peru, Cameroon, Ghana, Indonesia, Thailand, The Philippines, and Vietnam Overlap with themes 1-4 in East and West Africa and in Bangladesh and expansion areas in Tropical Americas and Southeast Asia Potential linkages Contribution to Contribution CRP1.2 from CRP1.2 Component options Local scale for interventions constraints and that balance opportunity production and assessment. ecosystem Lessons from concerns including innovation and for: rainwater other institutional management in strengthening crop livestock mechanisms that systems; improved are applicable in water productivity; other systems nutrient efficiencies, approaches carbon seeking to sequestration and balance multiple salinity livelihood management; water dimensions; access to support integrated smallholder system models. livelihoods; and Access to reuse of partnerships and wastewater. innovation Landscape and pathways at local basin scale impacts levels for uptake of land and water management interventions. Access to global conventions and conservation/environment earth system partners for uptake. Knowledge, tools, improved components for tree systems, management of conflict between NR users, and mitigation strategies for climate change in forest and tree systems Ensure intensification options are also climate proof offer potential new incentives to NRM via carbon market payments Holistic diagnosis for desirable components for complex system interventions, test results for forest and tree component research, tools and measurement for system productivity and NR integrity, trade-off and intensification implication for forest and tree systems, integration of trees across commodity production systems Strategies for biomass management including tradeoff; use of natural resources in relation to livestock production; crop combinations suitable for livestock feed Joint Research Strategies to address complex natural systems, combining diverse biophysical, social, economic science approaches to ISFM. Developing competency across systems CRPs in production systems research Joint action research testing of tree components, implication of climate change, ecosystem services thresholds, and intensification trajectories for tree systems Options for mitigating Greenhouse Gas Emissions in mixed systems Mechanisms for achieving integration Participation of key researchers in both CRPs; participation of CRP1.2 and CRP5 leaders along with key partners in annual work planning events undertaken by both CRPs; colocation of sentinel and benchmark areas/satellite sites programming of activities within the same target regions and benchmark areas Joint planning and funding of activities in both CRP1.2 and CRP6 Attendance of CRP6 scientists to CRP1.2 planning meeting and vice versa. Center or partner level facilitation of planning and implementation when both are involved in CRP6 and CRP1.2 Joint activities and meetings 99

105 10 MANAGEMENT ARRANGEMENTS FOR IMPLEMENTATION 10.1 Overview The Humidtropics executes it s research for development in Action Sites that are representative for eleven different Action Areas selected in tropical Asia, Africa and Americas within three main Strategic Research Themes. Research outputs are delivered at Action Sites, Action Areas, and globally leading to the program s development outcomes. This requires an efficient and effective governance and management framework to contribute successfully to the strategic objectives, ensuring anticipated participation, returns on investments, transparency and accountability (Figure 19). Humidtropics search for that delicate balance between inclusiveness and responsibility is reflected in the governance and management framework. It has taken guidance from the SRF and the feedback from ISPC and Fund Council in the inclusion of the key components recommended there. It is proposed that Humidtropics will have an Independent Advisory Committee with both scientific and development expertise, a Lead Center and its associated Board (namely IITA), an Executive Director and a Management Committee. There will be an Executive Office led by the Executive Director. Strategic Research Theme Leaders and Action Area Facilitators will be responsible for delivering program outputs. Humidtropics focal points for each partner will help ensure appropriate linkages including with other CRPs. Figure 19. Governance and Management Framework 100

106 10.2 Governance The Consortium Board contracts IITA as per Consortium-Lead Center contract articles 2, 2.1, 2.2, which stipulate that the Lead Center is responsible for the delivery, relevance and performance of the contract. The Consortium Board also has a conflict resolution role, should IITA fail to resolve issues with its partners. IITA as Lead Center has fiduciary and operational responsibilities for the implementation of Humidtropics. The Lead Center is thus fully responsible and accountable for the successful execution of the program and for its performance against the aforementioned contractual obligations. The IITA Board will: Account to the Consortium Board, including financial analysis and risk assessment Oversee the Lead Center s Performance Contract. Ensure efficient and effective engagement of the Advisory Committee The IITA Director General will: Be an observer of the Advisory Committee Be a champion for Humidtropics working together with the Executive Director and Program Management Committee. Work closely with the Consortium CEO on matters of Humidtropics. Advice the Executive Director as needed Resolve conflicts between involved partners Advisory Committee The Advisory Committee will have a major advisory role on Humidtropics on priority setting, partnerships, the strategic allocation of resources, and external linkages, to ensure that the needed set of Partners and Centers participate in order to achieve the goals and objectives of Humidtropics. Having an Advisory Committee ensures independence of the Humidtropics programmatic directions, shields the Lead Center from real or perceived conflict of interest, while still being accountable to the Lead Center. The Advisory Committee is composed of individuals that comprise R4D expertise and insights from diverse public and private sector partners such as farmer organizations, NGOs, Private sector, IARCs, NARs, and ARIs. The 8-person gender-balanced team is composed of 1 representative each from Tropical Americas, Asia and Africa following the general expertise and insight criteria, 3 individuals identified for their specific expertise including a Gender Specialist, Communication & Partnership Specialist and Systems Specialist, and 2 co-opted individuals on identified needs basis. The IITA-DG and a 2 nd center DG, recommended by the Consortium in collaboration with the ISPC, will be observers on this committee. The Lead Center s Board will formally appoint the Advisory Committee. The Lead Center Board and the project s Primary Partners will select the Committee members and Chair for the first year. Thereafter, the Committee will nominate its Chair and members, to be ratified by the lead Center Board and the Primary Partners. Committee membership will be for 3 years. Terms of 2-4 years in the initial year may be used to staggered turnover of members or continuity of the Chair s position. The Committee may recommend renewal of the appointment of a Committee member, once, at the end of his or her term. The Committee will meet twice per year, one of its meetings arranged back-to-back with the Lead Center Board. 101

107 The committee s role is to: Review the global program performance and the relevance of its outcomes. Make recommendations to Executive Director, Lead Center Board, and Individual Partners on opportunities for enhanced performance of the program, strategic alliances with partners, and effective engagement of R4D capacities towards greater impact. Advise the Lead Center and Executive Director on approval of the proposed workplans and budgets. Make recommendations to Executive Director, Lead Center Board, and Individual Partners on principles that assist the IITA DG in conflict resolution Periodically review the principles that guide resource allocation between projects, Primary and other partners, and the use of internal and external capacities and make recommendations to Executive Director, Lead Center Board, and Individual Partners on changes needed. The Chair of the Advisory Committee will be required to prepare an annual report to the Center Board, Primary Partners, and Executive Director. The Chair will liaise with the DG of the Lead Center and the Executive Director as needed. The Executive Director will provide quarterly progress reports to Primary Partners who are required to regularly update their constituencies on relevant progress and alert them to any emerging opportunities or threats that are of significance Primary Partners Primary partners are selected institutional research partners that through their mission, complementary skills, capacities and resources provide significant opportunities for greater innovation, accelerated development and greater impact of significant components of Humidtropics at international level. In the course of executing Humidtropics, strategic alliances with new and additional primary partners will be pursued and these partners may come from the NARS, ARIs, Centers, SROs or the private sectors. The Advisory and Management Committees will review and recommend on the inclusion or exclusion as primary partners annually in relation to the above criteria. Initial Primary Partners are those that are selected on the above criteria and co-authored the proposal, they currently are: IITA, CIAT, CIP, ILRI, ICRAF Bioversity, IWMI, AVRDC, FARA and WUR Executive Director IITA, in consultation with the Primary Partners initially and the Advisory Committee once formulated in subsequent recruitments, will appoint an Executive Director for the Humidtropics. The Executive Director Humidtropics will manage implementation through the Strategic Research Theme Leaders and Action Area Team Coordinators, lead the Humidtropics Executive Office and consult with the Lead Center Management regularly. Once per year, the Executive Director, on behalf of the Management Committee (see below), will prepare an annual report that is considered by the Advisory Committee. The Executive Office will consist of the Executive Director, Strategic Management, Communication and Partnership Manager, Finance Manager, and Assistants for Administration, M&E and Communications. It will also draw services related to Contracts & Grants, Projects, Communication and Finance from IITA s established offices on full costrecovery basis. This ensures consistency, transparency and accountability assuring that contractual obligations are met in the most cost efficient way by reducing costs of setting up additional sole-purpose offices. It also ensures that sub-contracts are managed following the Humidtropics project contracts and finance protocols. 102

108 10.6 Management Committee The Management Committee comprises of 5 Strategic Research Theme Leaders who will provide scientific leadership, oversight, and guidance for the SRTs. These will be nominated by their organization to allocate time for this global responsibility. They will be selected through a competitive process involving the Primary Partners, Advisory Committee and Executive Director. They will be supported and employed by their host organization. They will insure that the themes are appropriately planned, implemented, and monitored and advise on workplans and budgets to the Advisory Committee and Executive Director. They will review and support the quality of science within the theme and support global synthesis with the Executive Director. They will interact to insure integration amongst themes at local, regional, and global level. They will work with the Action Area Coordinators to insure that the thematic work is of priority for the region. They will facilitate links to other CRPs and to all partners. They will represent Humidtropics at international meeting related to the specific scientific focus to insure the relevance and quality of the science Action Areas and Sites Teams Humidtropics assumes that resources are best used where the action takes place and thus advocates highly decentralized operational management to ensure efficiency, flexibility, agility and relevance to improve effectiveness. Implementation of R4D project activities is carried out in the Action Areas. Assuring that projects are truly international each Action Area must span several countries within a Region (see 4.3 on Action Areas). It is recognized that most field operations are conducted at national level by Action Site Teams, but these will be internationally coordinated by the Action Area Teams to deliver their outputs. It is our vision that the Humidtropics is best managed at the tropical regional level but financial realities dictate the current approach that works at a less optimal level. When funding allows, Regional Coordination will be established to represent the Humidtropics Program regionally Tropical Africa, Tropical Americas and Tropical Asia - and to coordinate the Action Area and Action Site teams in the region for the Humidtropics Executive Director Action Area Teams Action Area Teams are executive multi-institutional, multi-disciplinary; gender balanced R4D working teams consisting of the relevant managers of project partners that individually oversee the implementation of multiple Action Sites workplans in the action areas. The composition of these teams is expected to be increasingly inclusive of non-traditional partners as they take on more projects under the program. Each Action Area Team will designate a Coordinator, who will be a scientist allocated from one of the project partners. The Coordinator leads and represents Humidtropics in the Action Area and serves on the Management Committee. The Action Area Teams are responsible for developing Action Area and Action Site R4D projects, including preparation of workplans and budgets for inclusion by the Management Committee planning for those, aligned with available funds and lead activities. They will ensure the fulfillment of their sub-contract s performance measures, including the timely submission of financial and technical reports to the Humidtropics Executive Office. They will conduct evaluations in the Action Areas as required and implement response to these evaluations. The Action Area Teams will ensure consistency in R4D approaches and methodologies across all Action Sites and organize the evaluation of activities on a seasonal basis. Monitoring and evaluation activities, as described in Section 13 will be embedded in above 103

109 activities and the lessons learnt will be fed back into the next cycle of planning and activity implementation. The Action Area Teams will interact with the Humidtropics Management Committee through the Executive Director to contribute to the production of outputs and outcomes and receive feedback on its operations. The contributions from the Action Areas will be used to make strategic decisions on approaches, partnerships, geographical focus, etc Action Site Teams An Action Site Team is a gender-balanced inclusive working team consisting of the relevant public and private sector research and development partners that execute Humidtropics R4D activities in an Action Site. Research and development partners are public and private sector institutions and individuals that through their complementary skills, capacities, and resources contribute to the development of specific research outputs of Humidtropics. Responsibilities and accountability of research and development partners that receive financial support towards the objectives of Humidtropics are defined in subcontracts executed at the action area level. The team is responsible for the R4D action of Humidtropics in their Action Site relating mainly to the NRM, Productivity and the Markets Research Themes, valuing production, nutrition and health, gender and poverty related outcomes. They will manage R4D projects in the Action Site, including developing work plans, delivering outputs, and responsibly manage allocated budgets. They will ensure the fulfillment of their sub-contract s performance measures in the Action Site, inclusive the timely submission of financial and technical reports to the relevant Action Area Team. They will conduct evaluations in the Action Sites as required and implement response to these evaluations. 104

110 11 TIMEFRAME This section describes in broad lines when the various SRTs will be operationalized in relation to the existing level and breadth of activities, cross-center cooperation, and partner engagement within the different Action Areas, as conceptualized around three tiers. Specific information is included on the status of the analytical framework, the research outputs, and the development outcomes for Action Areas belonging to different tiers (see Section 4.4) and on the consequences for budgeting Tiers 1, 2, and 3 Action Areas Within the current set of Action Areas, the level of activity and cross-center cooperation is variable from effective and operational with various Centers cooperating within a broad set of themes and through functional partner networks (e.g., the CIALCA project in the Great Lakes) (Tier 1 Action Areas) to embryonic networks with a single Centre leading a set of activities of a limited scope (Tier 3 Action Areas). Tier 2 Action Areas are intermediate between both Tiers. The long-term vision of CRP1.2, beyond the initial 3-year period is to have active R4D platforms in all Action Areas within the logic of the overall research framework underlying the Humidtropics. Tier 1 Action Areas: They typically contain on-going research for development activities focusing on various components within specific farming systems and through cooperation between various Centers and national research systems, backstopped by Advanced Research Institutes. Partnerships usually include value chain and dissemination partners and substantial investments are made in the promotion and dissemination of best technologies, supported by an effective M&E framework and partner capacity building. In Tier 1 Action Areas, priority interventions needed for transforming current partnerships in R4D Platforms include: 1) broadening of partnerships to cover all value chain components, 2) operationalization of the R4D platforms with clear goals, roles, and responsibilities, and effective interactions, 3) widening the research for development expertise to enable interventions for all system components, 4) instilling a systems approach towards intensification, 5) strengthening the dissemination tools and approaches to enable integration of complex knowledge, and 6) strengthening capacity for system interventions, evaluation, and dissemination. Tier 2 Action Areas: They typically have on-going research for development activities usually focused on one or a few system components, promoted by a single Centre. Market linkage and value chain investments are usually minimal and dissemination partners are commonly promoting and dissemination the few single components that are promoted by the Center. Systems research is virtually absent. Besides those detailed above, the following interventions will also be required to initiate R4D platforms in Tier 2 Action Areas: 1) setting up value chain partnerships, 2) engaging the major Centers dealing with the most important system components through the Humidtropics or through effective linkages with other CRPs, and 3) promotion of the R4D platform concept within the research and development community operating in that Action Area. Tier 3 Action Areas: They commonly have a very limited range of activities focused on one technology without specific market linkages and with limited promotion and dissemination channels. Tier 3 Action Areas also include parts of Action Areas with currently virtually no Center-led activities but with good future potential (e.g., Haiti in the Central American Action Area) and/or that are necessarily included in the Humidtropics for reasons related to the overall research strategy. All interventions mentioned under the Tier 1 and 2 Action Areas 105

111 will be equally required for Tier 3 Areas but without much previous experience or earlier initiative as compared to the Tiers 1 and 2 Action Areas Timeframe Within the first 3 years, the Humidtropics program will use the following logic to transform itself from an aggregation of on-going activities led by different Centers, towards an effective and integrated program that s focused around the delivery of the science and impact outcomes around the main research questions and hypotheses: Year 1 Analytical framework. First, baseline characterization activities will be implemented within the Action Areas which are areas targeted for impact within the Humidtropics. Secondly, across all Action Areas, the outcomes of SRT1 will set the scene for the identification of Action Sites within all Action Areas and ensure that the design or layout required to answer the research questions is embedded in this selection process. These activities will be implemented in the Tier 1, 2, and 3 Action Areas. It is anticipated that the critical entry points towards intensification will be identified and interactions with other CRPs formalized for the Tier 1 Action Areas in year 1. Research outputs. No SRT2 activities are expected to be initiated during year 1. Development outcomes. Only in the Tier 1 Action Areas will R4D platforms be operationalized with direct involvement of representatives from other CRPs that are crucial towards the success of the Humidtropics. This planning will include workshops to develop a more comprehensive system operational plan, identify activities and milestones, and assign responsibilities to all platform partners. We anticipate that an important part of current research being done by Centers and their partner will be integrated into Humidtropics but this will have to be assessed collectively through the planning processes. Year 2 Analytical framework. Identification of critical entry points towards intensification will happen in year 2 for the Tier 2 and 3 Action Areas. Research outputs. In Tier 1 Action Areas, the complete package of SRT2 interventions will be implemented while in the Tier 2 Areas SRT2 will start operating through some initial interventions, aiming at creating the necessary environment for SRT2 to operate in (e.g., input supply systems, value chain operationalization). Depending on the availability of new funding, this may also happen in Phase 3 countries. Interactions with other CRPs will be formalized for Tier 2 and 3 Action Areas in year 2. Development outcomes. In the year 2, Tier 2 and 3 Action Areas will be operationalized following the logic used in year 1 but with modifications based on the experiences of year 1.R4D platforms will be established and critical entry points identified. Dissemination activities under SRT3 will be initiated in the Tier 1 Action Areas, supported by an effective M&E framework. Year 3 Analytical framework. Although year 3 is early for the synthesis activities under SRT1, the concepts and methods that will have been developed to achieve this outcome will be tested and evaluated in the Tier 1 countries. 106

112 Research outputs. In year 3, SRT2 activities will be operational in all Action Areas. The level of activity will vary between the Tier 1, 2, or 3 Action Areas, depending on earlier progress with identification of interventions towards agro-ecological intensification, partner commitment, and general support for the R4D platforms. By year 3, interactions with other CRPs are anticipated to result in the added benefits that are hypothesized when focusing on system approaches. Development outcomes. SRT3 activities will be activated in all Action Areas although progress is expected to decrease from the Tier 1 to the Tier 3 Areas. For the Tier 2 and 3 countries, relatively more effort will be invested in institutional capacity building and supportive infrastructure and policies compared to Tier 1 Areas Budget Considerations In terms of budget allocations within the first 3 years and following above logic, it is clear that the scale of new investments will substantially increase over time because of two dynamics taking place simultaneously: 1) gradually more Action Sites will become operational in time, aiming at having all of them operational by the end of year 3, and 2) the intensity of interventions will increase in time with all the SRTs becoming active by the end of year 3. Moreover, a relatively smaller proportion of existing funds will be available for Humidtropics activities over time, thus requiring additional investments towards year 3. Investment needs are expected to level off by year 3 once all Action Sites are operationalized and all SRTs being implemented. The required level of funding beyond year 3 is anticipated to consist at least of a basic amount required to keep the R4D platforms operational, implement all M&E activities, and deliver a synthesis or the delivery of IPGs. Buy-in of other partners will determine the final level of investment made in the Humidtropics. 107

113 12 RISKS 12.1 Partners Limited Delivery Competences Different R4D partners involved along the impact pathway may not posses sufficient human, knowledge, capital, and infrastructural resources at the time of their essential involvement in delivery, which may seriously hamper progress and thus threaten impact. The political climate may affect policies on R4D investment, implementation and uptake. This risk will be mitigated through public-private partnerships and the development of dynamic strategic subprograms in which governance and resource issues will be anticipated and strategies and expectations adjusted accordingly Narrow Integration of Humidtropics in Wider Development Contexts Improvement of the quality of life of the poor and vulnerable goes beyond integrated agricultural systems, and should fit and collaborate with existing institutions and new initiatives in governance, health, education, social services and infrastructural development that are all influencing the poor and vulnerable. Failure to integrate risks working in isolation, and in the worst-case scenario, creates opposition and competition that may seriously affect implementation and effectiveness. This risk will be mitigated by conducting an inventory of institutions and including other programs in the development of the strategic sub-programs Limited Communication Communication is often hampered by cultural differences, technology and expectations, which may cause conflict, limited knowledge and technlogy transfer, lack of participation, and the reduced ability to monitor and evaluate progress. Through communication audits, an inventory of cultural awareness issues and potential communication challenges should be compiled and incorporated into essential solutions and development expectations as part of the R4D management system Limited Links Among Research and Development Partners There could be a risk if validated technologies are not scaled out because of inadequate linkages with extension and development partners. Thus, this issue needs to be high on the agenda when being discussed with a wider range of stakeholders Business Planning The Humidtropics program is complex and development impact will take a long time to achieve. This requires detailed business planning that brings in development partners to facilitate impact, timely progress reporting and sustained investment with long-term commitments from all R4D partners and investors. Unfortunately, long-term commitments are not common in the current research and development scenario. 108

114 13 MONITORING AND EVALUATION The Humidtropics Monitoring and Evaluation (M&E) system provides the basis for operational and performance monitoring as input into adaptive program management and accountability. M&E within Humidtropics systematically develops much of its data as a part of ongoing research activities. This is especially so for activities that integrates panel surveys at the level of the Actions Sites with baselines and periodic monitoring at the level of Action Areas. There is continual program monitoring along principal impact pathways that feeds directly into research planning and indirectly into the evolving M&E system for the whole CGIAR. To complement this part of the M&E system, monitoring instruments to capture much of the process innovation of the Humidtropics will also have to be developed, especially in the work of SRT3. This will be particularly important in monitoring what are effectively multi-institutional R4D efforts, often in an action research mode and focused on capturing institutional innovations. The following discussion of the M&E framework adds detail to Section 5 on Quantified Impact Pathways M&E Framework Figure 20. Planned M&E activities within the Action Sites and Action Areas leading to a Global Synthesis and release of the M&E Framework as an International Public Good. The M&E Unit (see further below) will work within the larger team addressing SRT1 that is responsible for the Situation Analysis. It will operate at both the program level and within the individual Action Areas. First that team will compile available information for the Action Sites and Action Areas and prepare a preliminary baseline based upon this secondary information (Figure 20). Next the unit will design a baseline survey, aligned to the Independent Evaluation unit (IEA) of the CGIAR and arrange for it to be field tested in representative Action Sites. Enumerators must be trained and operate from practical guidelines (not represented in Figure 20). As a result of field-testing, the baseline survey may be tailored to site-specific conditions but still contain sufficient standardized elements to permit cross-site 109

115 analysis. The unit will then develop the data screening, entry and analysis procedures that permit steady progress as baseline information streams from the Action Sites surveys. A synthesis of these data will also contribute to the initial Situation Analysis, a task of SRT1. Next the M&E Unit will develop monitoring tools for the field activities under SRT2. These too must be aligned with IEA guidelines and be field-tested. In this case, monitoring will be accompanied by detailed training in the use of these tools as such reporting will stem from both participatory monitoring methods and structured data collection instruments. Again, as monitoring commences, findings must be captured within similar processing and analytical protocols that consider baseline conditions so that the first increments of impact may be detected. Refinement of the M&E Framework continues as monitoring extends to the aggregation of development impacts under SRT3 to a point where the framework itself, and its accompanying survey instruments and analytical tools, are packaged as a flexible International Public Good for adoption by other CRPs and developmental research programs (Figure 20). The functions of learning and adaptive management are central to the M&E Framework. It will be designed as a continuous dynamic action-reflection-action process. This is an iterative process that allows revisiting certain reference points in order to clarify, confirm or change assumptions. Practical implications for dealing with the complexities of the integrated systems approach therefore include the need to establish continuous learning and reflection spaces and opportunities for adjustment and metrics that assess both technical and non-technical dimensions (Tubiello and Rosenzweig, 2008). The system research for R4D platforms and communication strategies described in the partnership section (Section 6) will play a key role to facilitate participatory M&E activities. Humidtropics will measure the impact of its systems-oriented research and development primarily as part of the scaling up research of SRT3 using a combination of methods for results-based monitoring, evaluating, and assessing impact in the various research themes. This M&E system will feed information back to research activities to inform research design and for development practitioners to inform the design of interventions. The framework for impact pathways (Section 5) allows for the identification and measurement of the outputs and outcomes of research, but also of its ultimate impacts on SLO targets. The impact pathway achievements and measurements or indicators (Table 14) provide a basis for the M&E Framework. Humidtropics intends to further develop methodologies for measuring and monitoring sustainable system intensification as part of tracking progress toward sustainable natural resource management. Evaluating the impacts of natural resource management via positive and negative externalities is based on an evaluation of changes in the provision of ecosystem services using tools such as choice experiments and payments for environmental services. This in turn influences the trade-offs and opportunity costs of farmers and broader aspects of community welfare (Hoyos, 2010; Swallow et al., 2009). Given the complexity in measuring changes in health and nutrition within a systems approach, household surveys will be conducted to measure changes in patterns of consumption (Smith et al. 2006). For example, a study on the patterns and determinants of dietary micronutrient deficiencies in rural areas of East Africa based upon food variety showed that staples provide somewhat more than 70% of the calorie intake of farmers in Rwanda, Uganda and Tanzania (Ecker et al., 2010). Key indicators that contribute to improving nutrition and health will involve measuring changes in household diets, and their nutritional values, with the understanding that the M&E process must link with CRP4 (see Section 9) to achieve this target. The list of initial variables to be monitored (Table 14) provides examples of indicators on which to focus from a technical standpoint, but without an assignment of priority or declared relevance to management and decision-making. A 110

116 more comprehensive list of all candidate, prioritized indicators will be compiled at the project inception workshop. Table 14. Types of partners to be involved in the PMEIA. Participatory monitoring, evaluation and impact assessment approaches (PMEIA) will be implemented in the on-farm work with farmers and in the scaling up activities in SRT3 working with a range of organizations and actors. The PMEIA will have four major and equally important objectives. The first is accountability to the beneficiaries. The second is participatory priority setting to provide project managers and Humidtropics partners with information for deciding what activities to implement, where to work, with whom, when and where to develop the livelihood frameworks. The third is accountability to investors based on feedback from project participants outside the CGIAR. The fourth is learning from the experiences to improve the design and performance of project investments over time, which will be essential in complex, system-oriented interventions. The Humidtropics M&E system will provide a comprehensive data and information system for undertaking ex-post impact studies across the four SLOs of the CGIAR. The M&E system will encompass baseline surveys at both Action Site and Action Area level, in-depth panel survey data at Action Site level, and household monitoring at Action Area level. The panel survey data will allow in-depth evaluation of more difficult to measure impacts on ecosystem services, nutrition, and poverty which then can be expanded at the level of the Action Area. To account for selection bias using the with and without project impact estimates, suitable combination of new advances in statistical techniques such as regression adjustment, instrumental variables, control function, propensity score matching, and difference in differences will be employed to reduce the variability between treatment and non-treatment groups (Baker, 2000; Godtland et al., 2003). This will be complemented with spatial analysis used on geo-referenced, disaggregated data to more clearly link outcomes and impact to local development interventions that will be complemented by process level monitoring (Thornton et al., 2003) M&E Process The M&E process is an extension of the framework as it is used to monitor field activities in the Action Sites (SRT2) and later to assess key impacts within the Action Areas (SRT3). First, Action Site teams must be commissioned to conduct the baseline survey so that it may be quickly reported for inclusion into the SRT1 Situation Analysis (Figure 20). If all goes well, these same enumerators will likely continue to monitor changes in production 111