Regional Wheat Yield Estimation by Integration of Remotely Sensed Soil Moisture into a Crop Model

摘要

A field study was conducted to estimate the regional wheat yield by integration of remotelysensed soil moisture index into CERES-Wheat model. The calibration and evaluation ofmodel was performed using experimental data and then applied on the area of Faisalabaddistrict for yield estimation. Area of Faisalabad district was divided into 7929 cells for independentsimulations. The weather data of the wheat season were used uniformly to all cells,while site-specific soil data were used for each cell. Recommended crop management practiceswere used in the model for all cells. Median normalized difference water index (NDWI)were used to estimate the irrigation amount for each cell. The estimated yield was validatedwith observed yield of 25 random farms. Model calibration results showed a good agreementbetween observed and simulated values of grain yield (RMSE = 284.8 kg ha~(-1)). Thevalidation of model at regional scale showed a close association with simulated andobserved yield of 25 farms (R~2 = 0.71). The regional yield estimation results indicated thatgrain yield varies from 1500 to 3593 kg ha~(-1) in Faisalabad district. The estimated mean yieldwas 2979 kg ha~(-1), which was 5.2% higher than the yield reported by Crop ReportingService (CRS), Punjab.