Modelling Floodplain Vegetation Response to Groundwater Variability Using the ArcSWAT Hydrological Model, MODIS NDVI Data, and Machine Learning

This study modelled the relationships between vegetation response and available water below soil surface using Terra’s moderate resolution imaging spectroradiometer (MODIS), Normalised Difference Vegetation Index (NDVI), content (SWC). The Soil & Water Assessment Tool (SWAT) interface known as ArcSWAT was used in ArcGIS for groundwater analysis. SWAT model calibrated validated SWAT-CUP software 10 years (2001–2010) of monthly streamflow data. average Nash-Sutcliffe efficiency during calibration validation 0.54 0.51, respectively, indicating that performances were good. Nineteen (2002–2020) MODIS NDVI data three different types (forest, shrub, grass) 43 sub-basins analysed WEKA, machine learning tool with a selection two supervised algorithms, i.e., support vector (SVM) random forest (RF). modelling results show vary dry wet seasons. For example, generated high positive (r = 0.76, 0.73, 0.81) measured predicted values all sub-basin against flow (GW), (SWC), combination these variables, season. However, reduced by 36.8% 0.48) 13.6% 0.63) GW SWC, Our models also top location (upper part) is highly responsive to SWC 0.78, 0.70) Although rainfall pattern variable area, summer very effective growth grass type. top-point dependent on both seasons, any instability or long-term drought can negatively affect floodplain communities. has enriched our knowledge responses each season, which will facilitate better management.