Loading...
Theses
Browse
1 results
Search Results
ThesisItem Open Access Wheat yield prediction of southern region of Uttarakhand by integrating remote sensing and WOFOST model(G.B. Pant University of Agriculture and Technology, Pantnagar - 263145 (Uttarakhand), 2017-08) Latwal, Asha; Nain, A.S.Accurate and real-time information on crop yield at national, international and regional scales is becoming gradually more important for food security in the world. Crop yield prediction may play a crucial role in advanced planning, strategy formulation and management of crop production. On the basis of previous research works, it has been broadly acknowledged that regional level crop yield estimates can be improved by assimilating the remote sensing data with crop simulation models. Considering the significance of yield prediction in food security, the present investigation was carried out to predict wheat yield using WOFOST model and remote sensing. The study was mainly designed in five parts: 1) application of multi-resolution data to calculate area-weighted mean NDVI, 2) development of spectralmeteorological models for yield prediction, 3) calibration and validation of WOFOST model, 4) yield gap analysis using WOFOST model and satellite data, and 5) regional level wheat yield prediction using WOFOST model and remote sensing. The calibration and validation of the model was conducted at Pantnagar region of US Nagar district, while regional yield prediction was carried over US Nagar and Haridwar districts of Uttarakhand for the period of 10 years (2005-06 to 2014-15). The LANDSAT image of each year was classified using supervised classification technique of ENVI-4.8 software and wheat crop was discriminated in both the districts. Wheat mask of each year was generated and overlaid on the SPOT derived NDVI images to develop the temporal growth profile of wheat crop and thereby calculating area-weighted mean NDVI. The spectral-meteorological models (SMM) were developed using combinations of NDVI and decadal weather variables at different crop growth stages using SPSS 16.0 software for regional wheat yield prediction. Yield gap analysis was accomplished after the estimation of potential yield through WOFOST model. Regional yield prediction of wheat was also performed using WOFOST model and trend yield. The normal growth profile showed that as the number of days increases from November/December onwards, wheat canopy increases linearly up to the first/second week of February when peak vegetation growth stage is attained. The NDVI values starts to decrease from March onwards and decreased sharply when the crop reached physiological maturity. During calibration the closed prediction was found between observed and simulated values with RMSE 5.56% and 11.67% for phenological stages and yield attributes, respectively. The validation results revealed that observed and simulated values were quite close with RMSE value 7.86% and 19.64% for phenological stages and different yield attributes, respectively. The yield gap analysis using model explained large variability between observed and potential yield in US Nagar (RMSE = 33.10%) and Haridwar district (RMSE = 43.66%). The average yield gap for US Nagar and Haridwar district was calculated as 1.47 t/ha and 0.35 t/ha, respectively using satellite data and WOFOST model. The results for US Nagar district demonstrated that RMSE of yield prediction was 2.43%, 2.82% and 2.27% using SMM, combination of WOFOST model & trend yield and assimilation of satellite data with WOFOST model, respectively. In Haridwar district RMSE was calculated as 2.44%, 7.22% and 8.36% between observed yield and predicted yield using SMM, combination of WOFOST model & trend yield and assimilation of satellite data with WOFOST model, respectively. Therefore, it can be concluded that these approaches can be used in the study region for regional level wheat yield prediction.