Spatio-Temporal Analysis of Vegetation Response to Climate Change, Case Study: Republic of Serbia

Abstract

Climate change has a potentially negative impact on the overall vitality of vegetation in both forested and agricultural areas. A comprehensive understanding of the interaction between climate and vegetation across various land cover types holds significant importance from multiple perspectives. This research examined the current state of vegetation trends and their interplay with climate parameters, specifically temperature and precipitation. Additionally, it aimed to provide insights into the anticipated changes in these climate parameters in the future, across the entire area of the Republic of Serbia. The vegetation was observed using the Normalized Difference Vegetation Index (NDVI) obtained from AVHRR/NOAA 11 satellite for the vegetation season (May–October) from 1981 to 2021, while the climate data records used the examination of the relationship between climate indicators and vegetation were monthly mean 2m temperature and precipitation obtained from the ERA5-Land (from April to October). The nonparametric Mann–Kendall test implemented with the Sen's slope estimator and the Pearson correlation coefficient (r) was utilized to identify trends (for the NDVI and climate variables) and the strength of the correlation, respectively. To obtain the information of temperature and precipitation change in future (from 2071 to 2100), the ensemble mean of the eight climate models, for vegetation period and summer season (June–July–August) from the EURO-CORDEX database was used. Results show relatively high NDVI values (> 0.5) over the entire area and the statistically significant (p < 0.005) positive NDVI trend increasing (up to 0.0006 \({\text{year}}^{-1}\))from the north (mainly agriculture cover) to the south (forest cover). In agricultural areas, a positive statistically significant correlation (r = 0.4–0.6, p < 0.005) indicates that the quality of vegetation cover in rainfed agriculture is directly dependent on the amount of precipitation, which serves as the sole source of moisture input. In contrast, the situation differs in forested areas where the correlation between NDVI and precipitation is often statistically not significant (p > 0.005) indicating that forests, because of their characteristics, are less dependent on the amount of precipitation. Regarding temperature, in agricultural areas, there is a positive correlation with NDVI, although it does not reach statistical significance. Conversely, in forested areas, a significant positive correlation is observed between NDVI and temperature which even positively contributes to the development of forest vegetation. In future, the recorded decline in precipitation (a substantial 22.72% drop) and the concurrent rise in temperature (up to 4.39 °C) in vegetation period, until 2100 might impact the reduction of NDVI.