High-Resolution (1km) Data of Groundwater Level Changes from 2003-2020 for Indus Basin

This high-resolution groundwater data is a result of a comprehensive study aimed at addressing the challenges posed by the spatial and temporal limitations of groundwater monitoring. The data covers the Indus Basin from 2003 to 2020, providing biannual (July and Oct) estimates of groundwater level (GWL) changes. To overcome the data gaps, we employed a cutting-edge approach that combines machine learning, local covariates, and exiting piezometers GWL data. The geographically weighted random forest (RFgw) model, a hybrid machine learning model, was the primary tool used to generate high-resolution (1 km2) and temporally continuous GWL estimates. The accuracy and reliability of this data have been rigorously assessed. This dataset is not limited to monitored sites but extends to unmonitored locations, offering valuable insights into GWL changes in regions without in-situ measurements. The incorporation of high-resolution covariates into the RFgw model allowed for reliable estimates of GWL changes at unmonitored sites.Readers can refer to the following publication for more details on the methods.Arshad, A., Mirchi, A., Vilcaez, J., Akbar, M.U. and Madani, K., 2023. Reconstructing high-resolution groundwater level data using a hybrid random forest model to quantify distributed groundwater changes in the Indus Basin. Journal of Hydrology, p.130535.