Modeling and Monitoring Wheat Crop Yield Using Geospatial Techniques: A Case Study of Potohar Region, Pakistan

In countries like Pakistan, whose economy greatly depends on agriculture and predominantly crop production, the estimation of crop yield before harvesting is very important. Remote sensing allows early estimation of crop yield before harvesting. The objective of the study is to evaluate the possibility of MODIS-derived vegetation indices using GIS and RS to estimate pre-harvest wheat yield in the Potohar region, Pakistan. Two MODIS products MOD15A2H and MOD13A1 for the period 2009–2018 were used for the derivation of LAI and indices. Wheat yield data of each district for the study period were obtained from the agriculture statistics of Pakistan. Model was run using 16-days composite MODIS vegetation indices as independent variable and crop yield data as the dependent variable. To check the ability and accuracy of the model RMSE, MAE and MBE were calculated. Overall, the percentage average difference between the actual and predicted yield was within −1.986%. Average RMSE and MAE values ranged from 34.28 to 76.50 kg/ha and 108.09 to 129.99 kg/ha, respectively. The MBE value ranged from 7.20 to 62.80 kg/ha. The results concluded that accurate wheat yield predication can be made almost 2 months before harvesting using geospatial techniques along with the statistical modeling approach.