SEDAM, a DSS for demand and distribution simulation and scenarios evaluation in a surface irrigation district

Transcription

1 SEDAM, a DSS for demand and distribution simulation and scenarios evaluation in a surface irrigation district José M. Gonçalves. a, André P. Muga a, Pedro Mateus a, A.A.Campos a, Luis S. Pereira a a Center for Agricultural Engineering Research, Institute of Agronomy, Technical University of Lisbon, Tapada da Ajuda, Lisbon, Portugal, Fax: jmmg@esac.pt, apmuga@apmuga.com, pmateus@isa.utl.pt; almeida_campos@sapo.pt, lspereira@isa.utl.pt Abstract The SEDAM model (sector demand and distribution model) is a decision support tool developed for searching the best scenarios for the combined improvement of farm and distribution systems in surface irrigation districts for planning and design purposes. SEDAM simulates the water demand and delivery at the sector level and aggregates the results to the district level. It operates with a GIS database in combination with the simulation models ISAREG for irrigation scheduling simulation and SADREG for surface irrigation design (described in a companion paper), which are linked with the same GIS, thus feeding interactively the database used by SEDAM. It includes a multicriteria analysis module that helps the users, mainly irrigation managers, to formulate and evaluate alternatives in respect to user-selected production, economic and environmental criteria. These include land and water productivity, farmers incomes, water saving, soil conservation, and irrigation environmental impacts. The model generates scenarios based on system management actions relative to the delivery network system and the farm irrigation system. Each scenario expresses a set of measures corresponding to an improvement strategy planned at the sector level. The field scenarios correspond to alternative improvement practices at farm level relative to irrigation scheduling, land levelling or land smoothing practices, adoption of equipment for water distribution to the fields that better control the inflow rates and the non-beneficial water uses. Each field scenario corresponds to a project in the SADREG and ISAREG applications whose results are stored in a common database, which provides for a multiple combination of decision variables. Each project represents a step further in farm irrigation improvement, thus to be implemented sequentially during a selected period. The delivery scenarios are defined by knowing the characteristics of the distribution network and adopting alternative upstream inflow rates, delivery rules and daytime delivery durations, which are assumed constant during the irrigation season. The canal network delivery model uses a simplified volume balance approach, where required data is also included in the GIS database. Computations include the estimation of canal seepage and runoff, and the lag time before the steady state flow regime is established. The fields are clustered according the farmers and the secondary canal. A rotational delivery scheme is considered in relation to crop scheduling. A 10-day time step is adopted for the simulation. A multicriteria analysis is adopted to evaluate and rank the alternative scenarios using a set of productivity, performance economic and environmental indicators. Results comprise both farm and off-farm improvements and are made available through the GIS database, thus in spatial distributed formats. This paper describes the model and shows results of its application to an irrigation district in Fergana, Central Asia. Key words: Surface irrigation, irrigation modernization and improvement, decision support systems, demand and delivery hydrographs.. 771

2 1 Models and Data SEDAM models the water demand and delivery on an irrigation Sector, for planning purposes. It is a DSS to aid irrigation managers to formulate and evaluate development scenarios building up to achieve the objectives for agriculture development and soil and water use (Gonçalves et al., 2003; 2004). These one are water savings, soil conservation, yields and farmers income increase and irrigation environmental negative impacts reduce. The model establishes scenarios based on system management actions, related with delivery network system and on-farm irrigation. Each scenario expresses a set of measures to implement a management strategy to plan the water at Sector level, i.e. corresponding to a Water Users Association (WUA). It allows impact evaluation of implementation of delivery scenarios. The user can view the network operation, the field supply hydrographs and the field and sector performance indicators. Also, Sector demand for 10 days time step. The different Sector scenarios can be compared and ranked through a multicriteria analysis. SEDAM databases stores all input, output and internal data in a structure of tables following the relational concept. The GEODB database is built in MS Access and validated in ArcGIS software; thus, it stores spatial and non-spatial data. The SADREG (Gonçalves et al., 2005) workspace database store information about fields, in particular the design alternatives for irrigation improvement, according the field scenarios. GISAREG database store the information about the irrigation scheduling for each field and delivery group. Sedam integrates several models and database: to model on-farm demand it applies GISAREG (Fortes et al., 2005) to formulate crop irrigation scheduling plans; the SADREG, that applies SIRMOD model, to generate on-farm surface irrigation alternatives, in an interactively procedure to explore an optimization search based on multicriteria methodology (Figure 1). The canal network water delivery model is a volumetric approach, where the geometric and hydraulic network characteristics are integrated with device controls, delivery rules and hydraulic constraints. It considers the seepage losses, when the canals are unlined, and an initial lag time before the establishment of the steady state flow regime. The fields are clustered, according the farmers and the secondary canal. A rotational scheme is considered to organize the delivery, linked with the crop scheduling made on a 10-day time step. Figure 1 SEDAM s models and database integration 772

3 2 Scenarios to improve irrigation water management The WUA development at water management point of view imply two kinds of decisions: delivery and on-farm irrigation. SEDAM handle this decision-making problem through building up scenarios that cover a wide range of practical solutions. From the simulation, these scenarios are explored and analyzed, providing to the user a sustainable knowledge about their performance and impacts, useful for decision aid. The global system performance at Sector level is dependent of these two types of decisions and, particularly, the management skills to induce the farmers to participate in a development process. Delivery scenarios refer to decisions-making at a collective network supply system taken at the WUA level or by the district management authority. Generally they correspond to the establishment of delivery rules, the sector main discharge and daytime supply period. The delivery rule applied is based on a rotational scheme and on a priority sequence of fields, made by the user. The delivery operative variables include the main discharge and daytime supply period, which are assumed constant during all irrigation season and play a basic role on WUA water planning. Each combination of two of these values generates a delivery scenario. The user could play SEDAM to explore the Sector impacts when changing these supply parameters. In particular, when the decrease on supply is required for water saving purposes, this simulation is useful to evaluate the feasibility of more severe supply reducing measures. On-farm irrigation scenarios, or field scenarios, corresponds to the implementation of improvement practices at farm level, including the application of an adequate irrigation schedule, land levelling or land smoothing practices, the use of water distribution equipments or devices to control a target inflow rate, and to reduce operational water losses and control runoff, or to apply a reuse technique. Each field scenario corresponds to a project on SADREG application (Gonçalves et al. 2005), resulting from a multiple combination of decision variables. Each project represents a step on field irrigation improvement, related with the modernization of farmers practices, which would be implemented sequentially during an implementation period, as explained and typified in Gonçalves et al Demand and distribution simulation The fields are clustered according the farmers and the secondary canal. From the GIS interface, the user can select the groups of fields. A field group is a set of fields supplied by the same branch canal and having homogenous irrigation scheduling from the point of view of crop and type of soil, so they are assumed to have the same delivery schedule. SEDAM application is done following several steps (Fig. 2): a) To create a new sector scenario. The user must input: the on-farm priorities (to later selecting an optimal alternative for field irrigation) the level of on-farm improvement, expressed by the proportion of field projects on global sector the main sector inlet discharge (l s -1 ) the daily main sector inlet supply (hours/day) b) Initialization of GEODB for a new sector scenario - it is made automatically and includes: load Field data o select optimal alternative o select scenario type (random, from the proportions of field projects ) o read irrigation scheduling (from GISAREG database) 773

4 GIS initializations workspace data base delivery groups Sector scenario field selection priorities field scenario proportions delivery parameters: main inlet flow rate and daily inlet time GEODB initialization determine Field data: select optimal alternative scenario type (random) read irrigation scheduling load field tables, from workspaces supply system (outlets, discharge) demand (application time vs. inflow) irrigation performances irrigation attributs water delivery for each decade, and each day select field to irrigate water applied to fields outlets, during application time of optimal alternative results saved on tables: reach_operation fields_supply decade volumes: supply, effective, percolation, runoff sector indicators Sector scenarios comparison sector supply field supply field effective field deficit canal seepage canal runoff yield productivity costs other impacts Figure 2 - SEDAM demand and delivery model flowchart load Field tables, from workspaces relative to: o supply system (outlets, discharges) o demand (application time vs. inflow) o irrigation performances o irrigation attributes c) Execute delivery simulation for a new sector scenario The canal network delivery model uses a simplified volume balance approach where required data is also in the GIS database. Computations include the estimation of canal seepage and runoff, and the lag time before the steady state flow regime is established each time that some change in the control is required. It applies the following delivery rules and procedures: 1) a rotational delivery scheme is considered in relation to crop irrigation scheduling, based on a sequential group of fields previously prepared by the user 2) for each field, to supply the maximum number of outlets, according the discharge available on network 774

5 3) a field only starts to irrigate if the remaining daily time is not less than the application time; if not, it will be irrigated next day 4) the volume of water not applied in the final part of the day will runoff to the drainage system 5) if a field do not receive water just by the end of decade (10-day period), it will be irrigate in the beginning of next decade; this volume of water will be recorded 6) when a field ends the irrigation, the water that stays available will be distributed to the next field in the sequential delivery table; the next field will only start the irrigation after a lag time, calculated from the reaches length; the water control required to this change is recorded on the reach operation table; the lag time is calculated based on the average flow velocity and the length of canal reach. 7) The canal reach seepage is estimated on the base of a seepage fraction proportional to the inflow rate and the reach length. d) View results of new sector scenario. The layout saved on GEODB and viewed by tables or by GIS graphics, includes: 1) Field distribution data, including start and end of irrigation 2) Network operation data, relative to the discharge at control structures during irrigation period 3) Water volumes, including effective supply, seepage and runoff 4) Performance indicators related with economic and environmental impacts: e) Compare sector scenarios The several sector scenarios developed are compared through multicriteria analysis. Criteria and attributes are described in Table 1 Table 1. Criteria and attributes utilized for multicriteria analysis Criteria Attributes Units Land Produtivity Land Economic Produtivity kg/ha EUR/ha Water Produtivity kg/ m 3 Benefits Water Economic Produtivity) EUR/ m 3 Beneficial Water Use ratio - Yield Value - Total Cost ratio - Cost Environmental impacts Total Cost per Water Use Fixed Cost per Water Use Variable Cost per Water Use Delivery operating cost Total Water Use (sector and fields) Runoff Ratio (field) Salinization Risk (m^3/ha/year) Levelling Soil Impact (cm) Soil Erosion Index canal seepage canal runoff EUR/m 3 EUR/m 3 EUR/m 3 EUR/m 3 m 3 /ha - m 3 /ha cm - m 3 /ha m 3 /ha 4 GIS Integration GIS application is a SEDAM extension designed to data input and results analysis. Since the database contains entities with geographic representation (fields, reaches or nodes), it seems logical to explore these capabilities. Some tasks that the user must perform, like create and sort delivery groups, attribute 775

6 SADREG workspaces to fields or define boundaries of each sector, become easier with GIS. In addition, the user can launch both SEDAM and SADREG models directly over the sector or a field which he wants to simulate or view data. The GIS application allows calling SADREG from the user-selected field (Fig. 3) or the SEDAM execution from a user selected sector. link SEDAM - SADREG one single field a sector objectives create the field workspace change irrigation scheduling run SADREG view the field workspace data base GISAREG objectives simulate sector demand and delivery delivery scenarios comparison GISAREG GEODB GIS GIS GEODB SADREG SEDAM SADREG Figure 3 SADREG application modes from SEDAM The GIS toolbars interface concerns several functionalities: GIS-SEDAM these functions perform interactions with the SEDAM database. o Definitions definition of SEDAM model and database path. o SEDAM run SEDAM model selecting a sector scenario. The selection is made on the map (sector) and from a list of existing scenarios for selected sector. o Delivery groups create and sort delivery groups by selecting fields from the map. o Table Results view operational results selecting a field or a reach on the map. o Graphic results view operational results selecting a field or a reach on the map. GIS-SADREG these functions perform interactions with the SEDAM database and the SADREG model and database: o Definitions definition of SEDAM model and database path. o SADREG run SADREG model selecting a field on the map. o Workspace create new workspaces and attribute workspaces to fields. o Project create new irrigation project. DataType Conversions functionalities to import and export data between SEDAM, SADREG and GISAREG databases. 776

7 5 Application and Results A SEDAM application to a Sector of Fergana irrigation District, Uzbekistan, was performed (Fig. 4). The SADREG workspace was developed and was linked (Fig. 5) to this sector (Gonçalves et al. 2005). Figure 4 Map of a sector of Fergana irrigation District and the two toolbars. Figure 5 - Workspace link from field map The field delivery groups were prepared from the GIS interface (Fig. 6). It shows to be an useful tool to plan water delivery, combining field and network knowledge to plan a better delivery schedule. 777

8 Figure 6- Delivery groups tool The output results for field supply are presented in a table through SEDAM and in a graphic through a GIS extension (fig. 7). The graphic represents the inlet volume of selected field for irrigation days grouped in decades. Figure 7 Table and graphical results of field supply for field nº 22 and sector scenario nº 1. 6 References Fortes, P.S., Platonov, A.E., Pereira, L.S., GISAREG - A GIS based irrigation scheduling simulation model to support improved water use and environmental control. Agricultural Water Management, in press. Gonçalves JM, Pereira LS, Campos AA, Fabião MS, Demand and delivery simulation and multi-criteria analysis for water saving. In: Pereira LS, Cai LG, Musy A, Minhas PS (eds.), Water Savings in the Yellow River Basin. Issues and Decision Support Tools in Irrigation. China Agriculture Press, Beijing, pp Gonçalves J.M., Pereira L.S., Campos A.A., Fabião M.S., Li Y.N., Liu Y., Cai L.G., Demand and delivery simulation and multicriteria analysis for water saving in the lower yellow river basin. In: GH Huang, LS Pereira (eds.) Land and Water Management: Decision Tools and Practices (Proc. 7th Inter-Regional Conf. Environment and Water, Beijing, Oct. 2004). China Agriculture Press, Beijing, Vol. 1, pp Gonçalves J.M., Horst M.G., Pereira L.S., Rolim J., Muga A., SADREG, a DSS for improving surface irrigation systems (this isuue). 778