Calibration and Parameterization of APSIM-Wheat using Earth Observation Data for wheat Simulation in Kenya
Keywords:Normalized Difference Vegetation Index (NDVI), Moisture Stress Index (MSI), Normalized Pigment Chlorophyll Ratio Index (NPCRI), Phenological Statistics, APSIM-Wheat Module
The ability to accurately translate the current condition of the crops into yield foresight expected at the end of the growing season helps the governments and other policymakers around the world to make informed decisions on matters relating to food security and economic planning. While the Agricultural Production Systems Simulator (APSIM-Wheat) is the widely used wheat-yield simulator in the world today, its major challenge is the lack of adequate data for calibration and parameterization of the model in many developing countries. This aspect inhibits the model's performance. This study utilized earth observation data derived from sentinel-2 to calibrate APSIM-wheat (version 7.5 R3008) to compensate for the data inadequacy and improve the model's performance in developing countries. The phenological statistics generated from sentinel-2 were integrated into the model as part of the input parameters. The phenological statistics were based on NDVI, MSI and NPCRI and were used with other crop management data collected at the field level. When the phenological statistics from sentinel-2 were used to calibrate APSIM-Wheat, the improved model outperformed the conventional APSIM-Wheat by 18.65% since the RRMSE improved from 25.99% to 7.34%; RMSE from 1784 Kgha-1 to 501 Kgha-1 and R2 from o.6 to 0.82 respectively.