Analysis of relationship between cereal yield and NDVI for selected regions of Central Europe based on MODIS satellite data

Abstract

Earth observation satellite technologies allow the status of crops to be assessed at various geographical scales. The most popular vegetation index for analyzing this status is the Normalized Difference Vegetation Index (NDVI). This article studies the relationship between the NDVI and cereal-grain yield in selected regions (corresponding to NUTS2 levels) of four Central European countries: Poland, Germany, the Czech Republic, and Slovakia. The remote sensing data cover the vegetation seasons (from the beginning of March until the end of June) of 2012–2016, and were obtained using the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor of NASA's Terra satellite. The average NDVI of the data was calculated for both the entire area and the arable-land area of the regions (according to the Corine Land Cover program). Statistical analysis includes correlation and linear regression analyses between cereal grain yield and mean NDVI for entire area and arable land of the regions in each year of the study. They are strongly correlated from the beginning of March until mid-May (most of the correlation coefficients for arable land are between 0.50 and 0.85), the strongest relationship being around the beginning of April in the Czech Republic and Slovakia (most of the correlation coefficients are between 0.80 and 1.00). In most regions of Poland, the relationships are quite strong, but in German regions they are weak and inconsistent. Regression coefficients (slopes) for relationships between the NDVI in the beginning of April and the grain yield of cereals range from 10.8 to 26.2. This means that an increase in the NDVI in early spring by 0.1 unit increases the grain yield of cereals by about 1.1–2.6 t/ha. The obtained results are promising because they prove the possibility of forecasting cereal grain yield at the regional level, three–four months before the harvest, which is important for planning food policy.