The link between the normalized difference vegetation index in major crops and meteorological factors

Abstract

Normalized difference vegetation index (NDVI) is frequently used in monitoring of meteorological events. The goal of this study was to establish the relationship between the spatial NDVI with air temperature and precipitation amounts in major crops cultivated in the steppe zone of Ukraine, namely, winter wheat, winter rapeseed, grain corn, soybeans, and sunflower. The study included the croplands of eight regions: the Crimea, Kherson, Mykolaiv, Odesa, Zaporizhzhia, Dnipropetrovsk, Kirovohrad, and Kharkiv. The yield data were retrieved from the official bodies of the State Statistical Service of Ukraine. The meteorological data on monthly air temperature and rainfall were retrieved from regional hydrometeorological centers. The NDVI values were retrieved from the GIMMS Global Agricultural Monitoring System, which provides Terrain MODIS NDVI 8-Day smoothed time series with 250 m resolution. The study was performed for the 2021-2022 for winter rapeseed and sunflower; 2017 and 2020 for grain corn; 2022 for winter wheat; and 2017 for soybeans. We found almost no or weak correlation between monthly NDVI values of the studied crops and rainfall amounts. A stronger correlation was found for air temperature, with the greatest values of the correlation coefficient of 0.76, 0.72, –0.72 for sunflower, soybeans, and winter wheat, respectively. The linear regression models, developed to predict NDVI based on the air temperature for the mentioned crops, provided good prediction accuracy with the relative error within 10–20%. The best overall fit and accuracy was produced by the model of sunflower’s NDVI. Only winter crops were observed to have a negative correlation with air temperature, suggesting that the cultivated varieties of these crops are heat-intolerant.