Determination of spatiotemporal distribution of Agricultural drought in Central Serbia (umadija)

Abstract

Over the past several decades, extreme drought events have become increasingly frequent in Central Serbia and have affected vast areas of land. In agriculture, the impact of drought is dramatic, especially if the soils in the affected area are shallow, like in most of the Sumadija Region where field crops, fruits and grapevines are traditionally rainfed. Exceptionally dry years have a multiple-year effect. Investment in irrigation is imminent but not feasible without state subsidies. To arrive at an impartial decision as to which part of a region should be given priority, it is extremely important to study the spatial distribution and severity of drought. Assessments made to date have resulted in the identification of areas with similar precipitation patterns. However, these assessments are rather crude and have classified parts of Serbia by precipitation depth or climatological precipitation deficit; as such, they are not accurate enough for agricultural purposes. Present-day needs call for more detailed analyses on a smaller scale for specific purposes such as plant production. It is a well-known fact that drought does not simultaneously affect an entire region. It is generally more severe in a certain area, less severe in another area, and may leave parts of the region unaffected. The objective of this research was to conduct a statistical analysis to determine the homogeneity of drought within a region, in spatial and temporal terms. Towards that end, the Standardized Precipitation Index (SPI) of the growing season was studied in 12 areas over a monitoring period of 60 years. Applying the Principle Component Analysis (PCA) method, a single dominant component with a total variance of 69.47% was determined, to which all sites contributed equally. Varimax rotation was then applied and resulted in the identification of three areas. Based on this study of the spatial distribution of drought in the considered region, only one area to the east could be reliably identified as experiencing neither extreme drought nor excessive moisture, while no such claim could be made about two other areas which were assessed. Consequently, drought mitigation measures may be based on separate, homogeneous units. Key words: Agricultural drought, principal component analysis (PCA), standardize precipitation index (SPI).