CERES model application for increasing preparedness to climate variability in agricultural planning - risk analyses

The role of soil, crop, climate and crop management for year-to-year variation of yield and groundwater pollution was quantified by simulation analyses with CERES-maize and CERES-wheat models over a 30-year period for four soil-crop combinations. It was established that Chromic Luvisol-maize-dry land combination was associated with the greatest coefficient of variability of yields (C-v = 43%) and drought frequency (in 22 years with yield losses more than 20%) over the analysed period. Average yield losses in dry vegetation seasons were 60% of maize productivity potential under sufficient soil moisture. Traditional and drainage controlling precise irrigation scheduling mitigated drought consequences by reducing year-to-year variability of yield to C-v = 5.6-11.6% on risky Chromic Luvisol. Long-term wheat yields were much more stable (C-v = 23-26%) than those of maize on Chromic Luvisol. In this case droughts covered 12 of the studied 30 years in which yield losses were 25-30% on the average. Soils of high water holding capacity (as Vertisol) stored 50-150 mm additional precipitation for crop evapotranspiration and thus reduced frequency of drought under both crops to 6-7 cases in 30 years. Agriculture should be more sustainable on this soil since variability of yield dropped to C-v = 13% for wheat and respectively C-v = 21% for maize. As a result Vertisol mitigated yield losses during dry vegetation periods by 10-15% for wheat and 22% for maize if compared with productivity under sufficient soil water. Thirty-year frequency analyses of seasonal nitrogen (N)-leaching, proved that ten of wheat and only one of maize vegetation seasons were susceptible to significant (10-45 kg N/ha/year) ground water pollution on Chromic Luvisol. Simulated precise irrigation scenario did not influence drainage in vegetation period. Another risky situations occurred under maize in the wettest fallow state after extremely dry vegetation (in one more of the studied years) when up to 18-33 kg N/ha/year might be leached. Earlier wheat sowing (on the 1st instead on the 15th of October) and nitrogen fertiliser split application timed as close as possible to the period of maximum crop N-uptake reduced nitrogen available for leaching under both crops. Drainage-controlling irrigation scheduling decreased maize drainage in fallow state by 30-40% in half of the years and caused 10-12% yield losses in average and dry seasons on Chromic Luvisol. (C) 2004 Published by Elsevier Ltd.