Digital Mapping of Land Suitability for Main Agricultural Crops in Romania

The scientific evaluation of land potential for different uses is essential for sustainable land development. Our study attempts to quantify this potential for agricultural purposes at a national scale, using GIS techniques, high-resolution spatial data, and recent climate data. The land evaluation methodology we applied in our study was developed in the 1980s by soil scientists from the National Institute of Research and Development for Pedology, Agrochemistry and Environmental Protection (ICPA) and it is still the official approach for the assessment of land suitability for crops in Romania. In our study, the application of the methodology is based on high-resolution spatial data including the 25 x 25 m resolution EU-DEM, the CHESLA climate database from which mean annual temperatures and precipitations were extracted for the 1990-2019 period, the digital soil map of Romania, the European LUCAS soil database. Firstly, we compared the evolution of mean annual temperatures and precipitations for 1961-1990 and 1990-2019 periods and found that there is a significant warming trend (an overall increase of 1.27 degrees C for the entire country, ranging from 0.9 degrees C to 1.6 degrees C) among the major landform units and a slight precipitation increase throughout the country (68.8 mm yr-1 for the whole country, ranging from 9.3 to 118.8 mm yr-1). Then, we applied the land evaluation methodology for the recent period (1990-2019), starting with the digital mapping of 15 land suitability factors, which were further aggregated to achieve the land suitability index and classes for the main agricultural crops of Romania (winter wheat, maize, sunflower, potato, and vine for wine). The results show that the most suitable landform units for wheat, maize, and sunflower are the plain areas (Romanian Plain, Western Plain) with LSI average values over 60. For potato, the suitable areas (LSI over 50-60) are less extended, being found especially in the intra-mountainous depressions and cooler plateau areas, while vines find very suitable conditions (LSI over 70) at the contact of the Romanian Plain and the Subcarpathians. To assess the model performance, we determined the shares of land suitability classes within the areas occupied by the specific crops. A second validation was carried out by correlating the total crop production at the county level with the cumulated LSI values. We found that, apart from potatoes, the model performs well for the analyzed crops. However, a methodological revision is necessary to accommodate temperature and precipitation values, which did not manifest in the reference climate period (1961-1990), but which are now part of the current climate of Romania.