Use of Crop Model in the Integrated Assessment Framework to Estimate the Biophysical Potential of Wheat Production in Sub-Saharan Africa
0 likes405 views
This document discusses the use of crop models to assess the potential for wheat production in Sub-Saharan Africa, particularly through collaboration between various agricultural institutions. It details the simulation of wheat yields based on different soil types, climate conditions, and fertilizer applications across selected countries in the region. The findings suggest that, even with changing climate conditions, the potential for rainfed wheat production remains viable, especially with the influence of CO2 fertilization considered.
![Use of Crop Model in the Integrated Assessment Framework to Estimate
the Biophysical Potential of Wheat Production in Sub-Saharan Africa
IFPRI: Jawoo Koo j.koo@cgiar.org and Zhe Guo (2033 K St., NW., Washington, DC 20006, USA), CIMMYT: Sika Gbegbelegbe, Kai Sonder, Bekele Abeyo, and Uran Chung
(56130 Texcoco, State of Mexico, Mexico), and U. of Florida: Senthold Asseng and Davide Cammarono (PO Box 110570, Gainesville, FL 32611, USA)
ME Agro-ecological characteristics Representative site Benchmark cultivar
SUMMARY ME1 Low rainfall irrigated, coolest quarter (3 consecutive Yaqui Valley, Mexico Seri M 82
months) mean min temp 3-11C (Ciano or Obregon)
CIMMYT uses process-based crop models in
ME2A High rainfall in summer; wettest quarter mean min Kulumsa, Ethiopia Kubsa
collaboration with other CGIAR centers and temp 3-16C, wettest quarter (3 consecutive wettest
months) precipitation > 250 mm ; elevation 1,400m
leading academic institutions, as an ex-ante ME2B High rainfall winter rain; coolest quarter mean min Gorgan, Iran Tajan
impact research tool at national, regional or temp 3-16C; elevation 1,400m
ME 3 High rainfall acid soil; climate as in ME2 and pH < 5.2 Passo Fundo, Brazil Alondra
global scales, to estimate the potential ME 4A Low rainfall, winter rainfall dominant; coolest quarter Aleppo, Syria Bacanora
mean min temp 3-11C; wettest quarter precipitation
performance of new varieties in various 100-400 mm
environment conditions and to target scale out ME 4C
areas for these, to assess the impact of climate
Mostly residual moisture ; coolest quarter mean min
temp 3-16C; wettest quarter precipitation > 100-400
mm
Indore, India HI 617
2 WHEAT ME’S Climatic-based
generalization of wheat
production systems, developed for
ME 5A High rainfall/ irrigated, humid; coolest quarter mean Jessore, Eastern Kanchan breeding and priority settings
change on food security, to inform the min temp 11-16C Gangetic plains in
Bangladesh
breeding programs about future threats and
opportunities and to develop and test the best 1 DESCRIPTION OF ME’S AND THEIR REPRESENTATIVE SITE/CULTIVAR Genetic coefficients of each cultivar are being
developed as of writing; this study used the coefficients available as of Mar 2012.
3 SIMULATED WHEAT YIELDS
Rainfed system with current climate
conditions with 100% of recommended
management practices in maize and wheat production fertilizer application rates
Planting month: Climate scenario-specific most-likely rainfed planting
systems. month, generated by applying the spring wheat growth requirements
In this study, the CERES-Wheat model of DSSAT v4.5 was Soil: Gridded soil profile database generated using FAO Harmonized
used in the biophysical-socioeconomic integrated World Soil Database v1.1 and ISRIC WISE Soil Profile Database v1.1
assessment framework to help identifying which areas in Nitrogen fertilizer: Three levels: (1) not applied, (2) medium
the selected Sub-Sahara African countries, beyond its intensification with 50% of recommended with 50 kg of DAP and 25 kg
of urea ha-1 (21 kg[N] ha-1), and (3) high intensification with 100% of
current spatial distribution in the region, have the
recommended: 100 kg of DAP and 50 kg of urea ha-1 (42 kg[N] ha-1);
potential of rainfed wheat production of smallholder
split-applied on 1 and 30 days after planting.
farmers enough to compete with imports. Using spatial
analysis and agro-climatic databases, potential areas were RESULTS & DISCUSSION
identified and their site-specific wheat yield responses to
Yield Responses
fertilizer applications were simulated on 5’ grids.
Crop simulation model showed a positive and significant yield response
CIMMYT’s Wheat Mega-Environment was superimposed to fertilizer application overall (see 3 and 4). Simulated yields varied
on the area to identify which representative variety can within and across countries depending on agro-ecological conditions, but
be used in the simulation. Each variety’s genetic were generally highest in the Eastern and Central African highlands and
coefficients were calibrated using CIMMYT’s variety trial mid-altitude growing regions. At medium levels of intensification, wheat
data. Soil properties were compiled from existing yields averaged between 1.2 and 3.5 t ha-1, but also reached about 4 t ha-1
in the highland agro-ecology zones.
databases on the regional soil characteristics and profiles.
Degraded soil fertility in the region was taken into Climate Change Impact
account in the simulation of soil processes simulated When CO2 fertilization effect was simulated otherwise negative impact
on yield under future climate scenarios were largely muted (see 5). In
using CENTURY Soil Organic Matters model in DSSAT v4.5.
reality, the changes in the climatic variables under future climate
Presumed farmers’ management practices in the scenarios are closely linked with the elevated atmospheric CO2
potential areas were defined through the consultations concentration, thus the circumstance used in this sensitivity
with local experts. analysis and its result remains hypothetical. However, from the
The overall simulation results suggested large biophysical overall result we infer that the simulated climate scenarios in
the study area may not adversely impact the potential wheat
potential areas may exist beyond the current wheat
production when the CO2 fertilization effect is considered.
growing areas in the region.
Assumptions and Limitations
Even under the warming future climate in 2050s, the
Scale: We simulated the representative wheat growth and
simulated wheat yield levels were not significantly yield on the supposedly representative soil, climate, and
reduced across the region when the compensating CO2 management practices for each grid cell. We used the model
fertilization effect was considered. and the model input data on 10 km grids, assuming the grid-
Combined with a spatial analysis on the estimation of level data and modeling appropriately represents wheat
production at the scale.
farm-gate price of fertilizer and the cost of transporting
Data: Grid-based soil data’s representativeness at the 10 km
wheat to the main market, the crop model-estimated
spatial resolution needs further research. Climate projection
wheat productivity and responses to fertilizer data were spatially downscaled from much coarser resolution
applications under the current and future climates were of data; there are uncertainties associated with the climate
used in the subsequent economic profitability analysis to modeling itself and the spatial-downscaling. Stochastically
analyze the economic potential of wheat production and generated daily weather data did not introduce weather
its competitiveness to the imported wheat. extremes; simulated yields in this study are only applicable for
the mean climate.
MATERIALS & METHODS Un-modeled constraints: Crop models do not take into
Study area: 12 countries in SSA (Ethiopia, Angola, D.R. Congo, Zambia, account all biotic and abiotic constraints that farmers may face
Zimbabwe, Mozambique, Kenya, Uganda, Tanzania, Burundi, Rwanda, in the field. Damages from pests, diseases, and weeds, soil
Madagascar); 5 arc-minute grids; excluding uncultivable areas nutrient constraints other than nitrogen and organic carbon,
Model: CERES-Wheat in DSSAT v4.5 with CENTURY Soil Organic and sub-optimum management practices, for example, were
Matters model not implemented. Thus, model-estimated yields and yield
responses to the simulated management practices should be
Field history assumption: Cultivated with poor management
cautiously interpreted, especially where farmers’ good
practices, initially grassland/forest; cultivation started 30 years ago
agronomic understanding and their resources for effectively
Variety: Mega-environment (ME)-specific varieties (see 1 and 2)
managing constraints are not readily available.
Climate: Spatially-downscaled grid-based monthly climatology for
current and future, developed
BASE CNRM-CM3, A2 CSIRO-Mk3.0, A2 ECHam5, A2 MIROC3.2 MR, A2
by CCAFS (ccafs-climate.org); Fertilizer
380 ppm 380 ppm 523 ppm 380 ppm 523 ppm 380 ppm 523 ppm 380 ppm 523 ppm
Daily weather generated from
785 913 862 1008 788 922 768 891
SIMMETEO weather generator NA 880
(-10.8%) (3.7%) (-2.0%) (14.6%) (-10.4%) (4.8%) (-12.7%) (1.2%)
to re-generate monthly means; 1425 1701 1538 1837 1385 1660 1475 1751
50% 1709
(-16.6%) (-0.5%) (-10.0%) (7.5%) (-19.0%) (-2.9%) (-13.7%) (2.4%)
10-year sequence x 10 1821 2197 1948 2343 1749 2118 1915 2292
100% 2207
realizations simulated (100 runs) (-17.5%) (-0.5%) (-11.7%) (6.2%) (-20.7%) (-4.0%) (-13.2%) (3.9%)
5 CO2 FERTILIZATION IN 2050 MAY OFFSET NEGATIVE YIELD IMPACTS Average yield (kg ha-1) and its % change
(in parenthesis) in 2050s from the baseline climate condition (in bold), with and without CO2 fertilization
4 AGGREGATED SIMULATED WHEAT YIELDS BY COUNTRY AND AEZ Rainfed system with
current climate conditions and three fertilizer application rates
Presented at the Wheat for food security in Africa conference in Addis Ababa, Ethiopia / October 2012](https://image.slidesharecdn.com/ifpri-jawookoo-simulationofwheatpotential-121101150643-phpapp02/85/Use-of-Crop-Model-in-the-Integrated-Assessment-Framework-to-Estimate-the-Biophysical-Potential-of-Wheat-Production-in-Sub-Saharan-Africa-1-320.jpg)
Use of Crop Model in the Integrated Assessment Framework to Estimate the Biophysical Potential of Wheat Production in Sub-Saharan Africa
- 1. Use of Crop Model in the Integrated Assessment Framework to Estimate the Biophysical Potential of Wheat Production in Sub-Saharan Africa IFPRI: Jawoo Koo j.koo@cgiar.org and Zhe Guo (2033 K St., NW., Washington, DC 20006, USA), CIMMYT: Sika Gbegbelegbe, Kai Sonder, Bekele Abeyo, and Uran Chung (56130 Texcoco, State of Mexico, Mexico), and U. of Florida: Senthold Asseng and Davide Cammarono (PO Box 110570, Gainesville, FL 32611, USA) ME Agro-ecological characteristics Representative site Benchmark cultivar SUMMARY ME1 Low rainfall irrigated, coolest quarter (3 consecutive Yaqui Valley, Mexico Seri M 82 months) mean min temp 3-11C (Ciano or Obregon) CIMMYT uses process-based crop models in ME2A High rainfall in summer; wettest quarter mean min Kulumsa, Ethiopia Kubsa collaboration with other CGIAR centers and temp 3-16C, wettest quarter (3 consecutive wettest months) precipitation > 250 mm ; elevation 1,400m leading academic institutions, as an ex-ante ME2B High rainfall winter rain; coolest quarter mean min Gorgan, Iran Tajan impact research tool at national, regional or temp 3-16C; elevation 1,400m ME 3 High rainfall acid soil; climate as in ME2 and pH < 5.2 Passo Fundo, Brazil Alondra global scales, to estimate the potential ME 4A Low rainfall, winter rainfall dominant; coolest quarter Aleppo, Syria Bacanora mean min temp 3-11C; wettest quarter precipitation performance of new varieties in various 100-400 mm environment conditions and to target scale out ME 4C areas for these, to assess the impact of climate Mostly residual moisture ; coolest quarter mean min temp 3-16C; wettest quarter precipitation > 100-400 mm Indore, India HI 617 2 WHEAT ME’S Climatic-based generalization of wheat production systems, developed for ME 5A High rainfall/ irrigated, humid; coolest quarter mean Jessore, Eastern Kanchan breeding and priority settings change on food security, to inform the min temp 11-16C Gangetic plains in Bangladesh breeding programs about future threats and opportunities and to develop and test the best 1 DESCRIPTION OF ME’S AND THEIR REPRESENTATIVE SITE/CULTIVAR Genetic coefficients of each cultivar are being developed as of writing; this study used the coefficients available as of Mar 2012. 3 SIMULATED WHEAT YIELDS Rainfed system with current climate conditions with 100% of recommended management practices in maize and wheat production fertilizer application rates Planting month: Climate scenario-specific most-likely rainfed planting systems. month, generated by applying the spring wheat growth requirements In this study, the CERES-Wheat model of DSSAT v4.5 was Soil: Gridded soil profile database generated using FAO Harmonized used in the biophysical-socioeconomic integrated World Soil Database v1.1 and ISRIC WISE Soil Profile Database v1.1 assessment framework to help identifying which areas in Nitrogen fertilizer: Three levels: (1) not applied, (2) medium the selected Sub-Sahara African countries, beyond its intensification with 50% of recommended with 50 kg of DAP and 25 kg of urea ha-1 (21 kg[N] ha-1), and (3) high intensification with 100% of current spatial distribution in the region, have the recommended: 100 kg of DAP and 50 kg of urea ha-1 (42 kg[N] ha-1); potential of rainfed wheat production of smallholder split-applied on 1 and 30 days after planting. farmers enough to compete with imports. Using spatial analysis and agro-climatic databases, potential areas were RESULTS & DISCUSSION identified and their site-specific wheat yield responses to Yield Responses fertilizer applications were simulated on 5’ grids. Crop simulation model showed a positive and significant yield response CIMMYT’s Wheat Mega-Environment was superimposed to fertilizer application overall (see 3 and 4). Simulated yields varied on the area to identify which representative variety can within and across countries depending on agro-ecological conditions, but be used in the simulation. Each variety’s genetic were generally highest in the Eastern and Central African highlands and coefficients were calibrated using CIMMYT’s variety trial mid-altitude growing regions. At medium levels of intensification, wheat data. Soil properties were compiled from existing yields averaged between 1.2 and 3.5 t ha-1, but also reached about 4 t ha-1 in the highland agro-ecology zones. databases on the regional soil characteristics and profiles. Degraded soil fertility in the region was taken into Climate Change Impact account in the simulation of soil processes simulated When CO2 fertilization effect was simulated otherwise negative impact on yield under future climate scenarios were largely muted (see 5). In using CENTURY Soil Organic Matters model in DSSAT v4.5. reality, the changes in the climatic variables under future climate Presumed farmers’ management practices in the scenarios are closely linked with the elevated atmospheric CO2 potential areas were defined through the consultations concentration, thus the circumstance used in this sensitivity with local experts. analysis and its result remains hypothetical. However, from the The overall simulation results suggested large biophysical overall result we infer that the simulated climate scenarios in the study area may not adversely impact the potential wheat potential areas may exist beyond the current wheat production when the CO2 fertilization effect is considered. growing areas in the region. Assumptions and Limitations Even under the warming future climate in 2050s, the Scale: We simulated the representative wheat growth and simulated wheat yield levels were not significantly yield on the supposedly representative soil, climate, and reduced across the region when the compensating CO2 management practices for each grid cell. We used the model fertilization effect was considered. and the model input data on 10 km grids, assuming the grid- Combined with a spatial analysis on the estimation of level data and modeling appropriately represents wheat production at the scale. farm-gate price of fertilizer and the cost of transporting Data: Grid-based soil data’s representativeness at the 10 km wheat to the main market, the crop model-estimated spatial resolution needs further research. Climate projection wheat productivity and responses to fertilizer data were spatially downscaled from much coarser resolution applications under the current and future climates were of data; there are uncertainties associated with the climate used in the subsequent economic profitability analysis to modeling itself and the spatial-downscaling. Stochastically analyze the economic potential of wheat production and generated daily weather data did not introduce weather its competitiveness to the imported wheat. extremes; simulated yields in this study are only applicable for the mean climate. MATERIALS & METHODS Un-modeled constraints: Crop models do not take into Study area: 12 countries in SSA (Ethiopia, Angola, D.R. Congo, Zambia, account all biotic and abiotic constraints that farmers may face Zimbabwe, Mozambique, Kenya, Uganda, Tanzania, Burundi, Rwanda, in the field. Damages from pests, diseases, and weeds, soil Madagascar); 5 arc-minute grids; excluding uncultivable areas nutrient constraints other than nitrogen and organic carbon, Model: CERES-Wheat in DSSAT v4.5 with CENTURY Soil Organic and sub-optimum management practices, for example, were Matters model not implemented. Thus, model-estimated yields and yield responses to the simulated management practices should be Field history assumption: Cultivated with poor management cautiously interpreted, especially where farmers’ good practices, initially grassland/forest; cultivation started 30 years ago agronomic understanding and their resources for effectively Variety: Mega-environment (ME)-specific varieties (see 1 and 2) managing constraints are not readily available. Climate: Spatially-downscaled grid-based monthly climatology for current and future, developed BASE CNRM-CM3, A2 CSIRO-Mk3.0, A2 ECHam5, A2 MIROC3.2 MR, A2 by CCAFS (ccafs-climate.org); Fertilizer 380 ppm 380 ppm 523 ppm 380 ppm 523 ppm 380 ppm 523 ppm 380 ppm 523 ppm Daily weather generated from 785 913 862 1008 788 922 768 891 SIMMETEO weather generator NA 880 (-10.8%) (3.7%) (-2.0%) (14.6%) (-10.4%) (4.8%) (-12.7%) (1.2%) to re-generate monthly means; 1425 1701 1538 1837 1385 1660 1475 1751 50% 1709 (-16.6%) (-0.5%) (-10.0%) (7.5%) (-19.0%) (-2.9%) (-13.7%) (2.4%) 10-year sequence x 10 1821 2197 1948 2343 1749 2118 1915 2292 100% 2207 realizations simulated (100 runs) (-17.5%) (-0.5%) (-11.7%) (6.2%) (-20.7%) (-4.0%) (-13.2%) (3.9%) 5 CO2 FERTILIZATION IN 2050 MAY OFFSET NEGATIVE YIELD IMPACTS Average yield (kg ha-1) and its % change (in parenthesis) in 2050s from the baseline climate condition (in bold), with and without CO2 fertilization 4 AGGREGATED SIMULATED WHEAT YIELDS BY COUNTRY AND AEZ Rainfed system with current climate conditions and three fertilizer application rates Presented at the Wheat for food security in Africa conference in Addis Ababa, Ethiopia / October 2012