info:eu-repo/semantics/article
Early assessment of crop yield from remotely sensed water stress and solar radiation data
Date
2018-11Registration in:
Holzman, Mauro Ezequiel; Carmona, Facundo; Rivas, Raúl Eduardo; Niclòs, Raquel; Early assessment of crop yield from remotely sensed water stress and solar radiation data; Elsevier Science; Isprs Journal Of Photogrammetry And Remote Sensing; 145; 11-2018; 297-308
0924-2716
CONICET Digital
CONICET
Author
Holzman, Mauro Ezequiel
Carmona, Facundo
Rivas, Raúl Eduardo
Niclòs, Raquel
Abstract
Soil moisture (SM) available for evapotranspiration is crucial for food security, given the significant interannual yield variability of rainfed crops in large agricultural regions. Also, incoming solar radiation (Rs) influences the photosynthetic rate of vegetated surfaces and can affect productivity. The aim of this work is to evaluate the ability of crop water stress and Rs remotely sensed data to forecast yield at regional scale. Temperature Vegetation Dryness Index (TVDI) was computed as an indicator of crop water stress and soil moisture availability. TVDI during critical growth stage of crops was calculated from MODIS products: MODIS/AQUA 8-day composite LST at 1 km and 16-day composite vegetation index at 1 km. Rs data were obtained from Clouds and the Earth's Radiant Energy System (CERES). The relationship between TVDI, Rs and yield of wheat, corn and soybean was analyzed. High R2 values (0.55–0.82, depending on crop and region) were found in different agro-climatic regions of Argentine Pampas. Validation results showed the suitability of the model RMSE = 330–1300 kg ha−1, Relative Error = 13–34%. However, results were significantly improved considering the most important factor affecting yield. Rs proved to be important for winter crops in humid areas, where incoming radiation can be a limiting factor. In semi-arid regions, soils with low water retention capacity and summer crops, crop water stress showed the best results. Overall, results reflected that the proposed approach is suitable for crop yield forecasting at regional scale several weeks previous to harvest.