Abstract. Groundwater recharge sustains groundwater discharge, including natural discharge through springs and the base flow to surface water as well as anthropogenic discharge through pumping wells. Here, for the first time, we compute long-term (1996–2015) groundwater recharge rates using data retrieved from several groundwater-level monitoring locations across India (3.3 million km2 area), the most groundwater-stressed region globally. Spatial variations in groundwater recharge rates (basin-wide mean: 17 to 960 mm yr−1) were estimated in the 22 major river basins across India. The extensive plains of the Indus–Ganges–Brahmaputra (IGB) river basins are subjected to prevalence of comparatively higher recharge. This is mainly attributed to occurrence of coarse sediments, higher rainfall, and intensive irrigation-linked groundwater-abstraction inducing recharge by increasing available groundwater storage and return flows. Lower recharge rates (<200 mm yr−1) in most of the central and southern study areas occur in cratonic, crystalline fractured aquifers. Estimated recharge rates have been compared favorably with field-scale recharge estimates (n=52) based on tracer (tritium) injection tests. Results show that precipitation rates do not significantly influence groundwater recharge in most of the river basins across India, indicating human influence in prevailing recharge rates. The spatial variability in recharge rates could provide critical input for policymakers to develop more sustainable groundwater management in India.

No abstract

More than quarter of underprivileged global population, who lack access to basic sanitation and clean drinking water, live in India. Consequently, every year, millions suffer with enteric diseases from drinking faecal-contaminated groundwater. The UN Sustainable Development Goal lists access to safe water and basic sanitation for all by 2030, as their sixth goal. For the first time, the role of economic improvement on decrease in water-borne faecal pathogens was studied across Indo-Ganges-Brahmaputra river basin (IGB) for almost last three decades, to delineate the long-term improvement trends of groundwater quality across India, as a consequence of development. Long-term temporal (1990-2017) and high-resolution spatial (administrative block scale, n = 2217) datasets of water-borne faecal pathogen concentration in groundwater and satellite-based nightlight (NL) were used to investigate the statistical trends and causal relationships. Linear and nonlinear (Hodrick-Prescott) trend analyses, panel data analyses, Bayesian vector autoregression (VAR) and lead-lag causality (LLC) analyses were performed on aforesaid culled datasets. However, the efficiency of development in alleviating the water quality and public health, and relationship with economic development, has not been well understood. Here, for the first time, using long-term, high-spatial resolution (n = 2217), annual in-situ measurements and multivariate statistical models, we show that the spatially variable groundwater faecal pathogen concentration (FC, 2002-2017, −1.39 ± 0.01%/yr) has been significantly decreased across the basin. In most areas, increasing satellitebased NL plays a significant role (NL, 1992-2013, 3.05 ± 0.01%/yr) in reduction of FC. However, in areas with low literacy rate surpass development. Enhanced decrease of faecal coliform concentration in groundwater possibly signifies the implementation of Clean India Mission since 2014.

Despite numerous studies in recent times, there is no consensus on the primary drivers for groundwater storage (GWS) changes over India. Thus, predicting future groundwater level trends seems remote. In this context, using Gravity Recovery and Climate Experiment (GRACE)-derived GWS, WaterGap model-based groundwater recharge (GWR), and groundwater withdrawal (GWW), we show that GWW exhibits a stronger dominance than GWR on GWS change over India. Furthermore, we developed feed-forward neural network (FNN), recurrent neural network (RNN), and deep learning-based long short-term memory network (LSTM) models using multidepth in situ observations from a dense network of monitoring wells (n = 5367, 1996−2018), to simulate and forecast groundwater levels (GWL) in India. The result demonstrates the better performance of LSTM (>84% of observation wells showing r > 0.6, RMSE n < 0.7) across India, outperforming both FNN and RNN in the testing period of 5 years (2014−2018). Our estimates also reveal that besides the prevailing long-term (1996−2018) statistically significant (p < 0.1) declining GWL trends in northwest India and the Ganges river basin, higher declining trends will potentially be observed in parts of north-central and south India in the forecasting period of 5 years (2019−2023). We envisage that the forecasting approach presented in the study can contribute toward an improved urban−rural drinking water supply and sustainable crop production for 1.3 billion people in India.

No abstract