ABSTRACT
Wheat is a major food grain crop of main agricultural region i.e. northern plain of India. Haryana state holds a premium position in wheat production (Rabi Season) in the country. Pre-harvest yield estimation of wheat has key role in policy framing. In Haryana, Agriculture is a big support to its economy which continues to occupy a prominent position in State GDP. In present research, Agromet-Spectrals models have been developed for this purpose i.e. yield estimation in Haryana with the help of input data such as meteorological indices and satellite based NDVI (NASA’s-MODIS) from 2000-2017. Empirical models were developed for predicting wheat yield for Hisar and Karnal districts representation the two agro-climatic zone of state in Haryana, India. The models were developed used weather variable (Temperature (Minimum and Maximum), Relative Humidity (Morning and Evening) and Rainfall) and spectral indices Normalized Difference Vegetative Index viz. Agromet-model (weather model) and Agromet-spectral model (MODIS-NDVI). Weather or Agromet model was integrated with NDVI values for both location to enhanced the accuracy of models. Regression models were developed using significant weather variables and NDVI data for wheat yield prediction at both location. The result revealed that the models when integrated with remote sensing data (NDVI) gave better prediction as compared to agromet model that depends only on weather variables. Agromet-models (adjusted R2 = 0.38 to 0.78) whereas satellite data based NDVI i.e. MODIS-NDVI for both station gave best result (Adjusted R2 = 0.61-0.86) as compared to weather models. MODIS-NDVI pixel based values observed to be more effective for wheat yield predication in integrated with weather parameters. This study could help the provincial government of Haryana as well as in northern plains in estimation of yield prior harvest at first week of April by using weather spectral (NDVI-MODIS) models.
