Spatial and temporal patterns of rainfall variability and its relationship with land surface phenology in central east Argentina

Transition zones between dry and humid climates are highly sensitive areas where small changes in the spatial and temporal distribution of rainfall may have vast consequences on vegetation development. Rainfall regimes are expected to change worldwide and predictions include variations in total annual precipitation and timing of rain events. In this context, evaluating the relationship between vegetation activity and rainfall variability along transition zones is particularly relevant. In this study, 54 years (1961-2013) of gridded monthly rainfall data (Climatic Research Unit Time-Series Version 3.22 [CRU TS3.22]) were used to analyse annual and inter-annual rainfall variability. Land surface phenology (LSP) metrics (2000-2013) were derived from MODIS 16-day composites Normalized Difference Vegetation Index (NDVI) data (MOD13Q1 product) and relationship between rainfall variability and LSP metrics was assessed. Annual rainfall showed a strong seasonality in the northeast of the study area that diminishes towards the southwest. Inter-annual rainfall variability, which showed a significant 8-year cycle periodicity after 1980, was stronger in the southwest of the region. For most of the study area, summer rainfall (November-January) showed significant negative correlation with August-October Southern Oscillation Index (SOI) average, although this connection was not reflected in annual precipitation totals. Regarding vegetation response to rainfall variability, two main areas were identified where differences in LSP metrics between wet and dry years were larger than the average. The first one was a zone within the humid to semi-arid transition belt, where NDVI maximum, NDVI amplitude and NDVI integral showed higher values during wet years. The second one was the central north area, which showed higher NDVI amplitude values during dry years. In the study area, water availability is a sensitive issue for natural ecosystems, agriculture and cattle-raising. These results provide a better understanding of the relationship between rainfall variability and vegetation changes, which is useful information for the development of future management policies.