Modelling: does it have a role in nutrient management? Dr Peter Thorburn Research Group Leader, Northern farming Systems CSIRO Ecosystem Sciences

Transcription

1 Modelling: does it have a role in nutrient management? Dr Peter Thorburn Research Group Leader, Northern farming Systems CSIRO Ecosystem Sciences SASTA Symposium on Sustainable Soil Use, 31 August 2011

2 Outline Some modelling philosophy What is a model Some examples of models in nutrient management The cost of complexity What is sensible to model Model applications in nutrient management Exploring N fertiliser management strategies Exploring crop-inputs from organic N sources Evaluating water quality under different management regimes Assessing management impacts on N 2 O emissions

3 Some modelling philosophy......a model is anything used in any way to represent anything else....models...are represented by concepts...which are formed after a conceptualization process in the mind... Wikipedia, the free encyclopaedia Models about the world v. models of the world Brian Keating (1996) Models are simplifications of the world. If they were complex as the world they would take real time to run and so cease to be useful. Calvin Rose (1986)...(all) models are wrong, but some are useful... George E P Box (1971)

4 Cane Yield (tonnes/ha) Some models in nutrient management... Simple models To maximise yield, always... Apply 160 kg N/ha Apply 160 kg N/ha to Soil kg N/ha to Soil 2 Apply 160 kg N/ha to Soil kg N/ha to Soil 2 Except after a legume, then reduce by... A rugby analogy: The Springboks always win The Wallabies always lose More complex models Response curves: Polynomial, Mitscherlisch, hyperbolic, etc. Issue: Many dependencies after Keating et al (1994) Mechanistic daily time step C & N: e.g. APSIM- SoilN, DayCent etc. Thorburn et al (2005) Rate of N fertiliser (kg/ha)

5 Model complexity power at a cost Simple models Generally static No consideration of time dimension Constant conditions Production potential Underlying process Environment Limited predictive value Limited explanatory value More complex models Generally dynamical systems models, capturing Climate variability Seasons are different Climate deviates from average, etc Soil processes Structure declines, improves Organic matter runs down, builds up Some process are mediates by temperature and moisture Stirzaker et al. (2010, Fig 5)

6 Outline Some modelling philosophy What is a model Some examples of models in nutrient management The cost of complexity What is sensible to model Model applications in nutrient management Exploring N fertiliser management strategies Exploring crop-inputs from organic N sources Evaluating water quality under different management regimes Assessing management impacts on N 2 O emissions

7 Cane Yield (t/ha) Cumulative cane yield (t/ha) Exploring/developing concepts in N fertiliser management recommendations Predicted performance of N Replacement system (Thorburn et al., 2003, 2004) Year Recom. N (160) N-200 Replace Actual performance of N Replacement system (Thorburn et al., 2011) Farm Repl Low Yield target 120 t cane / ha

8 Availability of N following soybean crops: Can we reduce N fertiliser applications? Plant 1R 2R 3R 4R Park et al. (2010) Crop Predicted reduction in N fert applied for optimum yield (%) Plant 100 Ratoons Up to 40 Soil

9 Evaluating effects of different management strategies on water quality and production Water quality issues driven over concerns about GBR health Federally funded incentives to move from D- and C-Class management Sugarcane management now regulated in three main catchments Great Barrier Reef and catchments N losses through runoff and leaching under four managements: Mackay loam (Biggs et al, 2011) Courtesy of Doug Yuille, Qld DERM

10 Exploring N 2 O emissions in different cane production systems (Green bars measured data) Mackay loam - trash ret (1) Mackay loam - trash ret (2) Rocky Pt clay Burdekin clay - low irrigation Burdekin clay - high irrigation Maryborough - low C soil Maryborough - high C soil Tully (wet tropics) Mackay loam - trash burnt Mackay loam - trash retained Mackay clay - trash burnt Mackay clay - trash retained N2O-N Emissions (% N Fert) Simulation v. measurement Good (as expected) Impact of irrigation Small Impact of soil carbon Small Impact of trash Big Impact of soil type Moderate Only with trash Impact of climate x management Substantial Thorburn et al. (2010 a,b)

11 Conclusions (for sugarcane production) Nutrient management revolves around models Complexity differentiates models Sensible to model nutrients at process level where... Processes/availability environmentally controlled Complex, non-linear feedbacks Competing objectives Models for N cycling in soil-plant system well advanced Aid understanding of complex problems Organic inputs Environmental loss problems Models under-utilised to explore optimum management...(all) models are wrong, but some are useful...

12

13 What is sensible to model? N Important nutrient Competing economic and environmental goals Biology dominates availability Environmentally controlled Non-linear interactions Dynamic process P S Cations & anions Micro nutrients Physical processes dominate Well sorbed Small environmental impacts

14 Relative Yield (%) Availability of N following filter mud: Can we reduce N fertiliser applications? 120 Burdekin example Plant 1st ratoon 2nd ratoon 3rd ratoon Thorburn et al (2008, in prep) No mud With mud Fertiliser (kgn/ha) Yield optimum predicted to occur at lower N with filter mud Similar results across wide range of environments At predicted optima, only approx half N in filter mud recovered by crops Soil

15 How could trash retention affect N responses? Three possibilities... Sugarcane yield (t/ha) Sugarcane yield (t/ha) Sugarcane yield (t/ha) Sugarcane yield (t/ha) Legend: Residue Burnt Retained 1 Same yield, lower opt N N applied (kg/ha) N applied (kg/ha) 2 Higher yield, same opt N 3 No effect (where does the N go?) N applied (kg/ha) N applied (kg/ha)

16 Evaluating effects of different management strategies on production and water quality Great Barrier Reef and catchments N losses under four managements: Mackay loam (Biggs et al, 2011) Courtesy of Doug Yuille, Qld DERM N losses and yields under 12 different managements (L, M H x 1, 2, 3, 4): Burdekin (Thorburn et al, 2011)

17