WHEAT YIELD PREDICTION BASED ON MODIS NDVI TIME SERIES DATA IN THE WIDER REGION OF A CEREAL PROCESSING PLANT

• Main Authors: Szabó Andrea, Tamás János, David Odunayo Adeniyi, Nagy Attila
• Format: Article
• Language: English
• Published: Editura Universităţii din Oradea 2019-11-01
• Series: Natural Resources and Sustainable Development
• Subjects: modis; ndvi; wheat; yield loss
• Online Access: https://docs.google.com/viewerng/viewer?url=http://nrsdj.com/papers/2019-2/10.Szabo-Andrea.pdf

The environmental impacts of climate change have been identified as a central issue in of the agriculture because the emergence of drought risk is a growing barrier to crop production and resulting in a decline in the quality and yield of cereals in recent years. In recent years there was a great development made in cereal processing in Hungary. In general, food processing plants requires steady crop supply, therefore it is essential to monitor the possible yield or yield loss of the potential supply area of the new cereal processing plant in Gyöngyösvisonta. The aerial and satellite images can be used to monitor the response time of areas to biotic and abiotic stress effects, in this study the application of water stress and drought monitoring. We analyzed on the basis of the remote-sensed time series data. In the context of recordings, a number of stress indexes can be calculated to determine the productivity of biomass during a drought period. The study site covers four counties around Gyöngyösvisonta. The source of remote sensed time series data was the 16-day images of MODIS NDVI. The 250 m ground resolution images were downloaded and processed in ArcGIS software using various GIS methods from 2003 to 2018 that area. The sample area was selected that used for the production of wheat from the CORINE database, NDVI, and crop phenology. We applied time series NDVI images using different masking techniques and created the models for wheat yield prediction. During modeling calibration of NDVI data sets was performed by correlation and regression calculations with yield and NDVI data sets. Based on the results moderate correlations (r~0.8), areas with different drought risk levels can be delineated to estimate yield loss with MODIS data in May and June which is the most suitable period for yield loss monitoring before harvest. Calculating these data will allow farmers to determine the appropriate intervention time to avoid territorial degradation.