A Connectivity Response Unit (CRU) approach to assess the impacts of LULC on surface flows in a tropical floodplain wetland

The Kaabar Tal is a floodplain wetland in the Kosi-Ganga interfluve of East Ganga Plains (EGP), India. It is the largest wetland of the region and a potential Ramsar site but is currently in a degraded state because of drying and fragmentation over the past two decades. We hypothesize that the drastic changes in the land-use/land cover (LULC) in its agriculture-dominated catchment is one of the reasons for its current water-stressed situation. We have used multi-temporal NDVI (Normalized Difference Vegetation Index) obtained from Landsat data to track the changes in the vegetation pattern from the year 1989 to 2017 for pre- and post-monsoon seasons. We used these time-series NDVI dataset in a CRU framework and applied geostatistical methods namely the Getis-Ord Gi* and Mann-Kendall trend test statistics to identify and quantify the regions of the catchment where the vegetation-induced impedances have potentially increased, decreased, or remained same. The CRUs are defined as the landscape units with similar hydrological connectivity potential and are the clusters of response elements. In this case, individual NDVI pixels are the response elements.

We argue that vegetation pattern embedded in LULC provides the most significant impedance to surface flows in a flat terrain such as an interfluve. Our results show that the impact of LULC change is higher in the post-monsoon compared to the pre-monsoon season, the former being the growing season. The catchment areas with 'intensifying,' 'diminishing,' and 'persistent' impacts of LULC on the surface flows have been demarcated. Field visits imply that the changing agricultural patterns are the reason for the 'intensifying' and 'diminishing' (or dynamic) potential impacts while the presence of topographic features such as other floodplain waterbodies are related with the 'persistent' (or static) potential impacts. The protocols developed in this study can potentially be applied to any other catchment which is being subjected to dynamic LULC patterns.