An official website of the United States government
Abstract GRACE (Gravity Recovery and Climate Experiment) has been widely used to evaluate terrestrial water storage (TWS) and groundwater storage (GWS). However, the coarse‐resolution of GRACE data has limited the ability to identify local vulnerabilities in water storage changes associated with climatic and anthropogenic stressors. This study employs high‐resolution (1 km2) GRACE data generated through machine learning (ML) based statistical downscaling to illuminate TWS and GWS dynamics across twenty sub‐regions in the Indus Basin. Monthly TWS and GWS anomalies obtained from a geographically weighted random forest (RFgw) model maintained good consistency with original GRACE data at the 25 km2grid scale. The downscaled data at 1 km2resolution illustrate the spatial heterogeneity of TWS and GWS depletion within each sub‐region. Comparison with in‐situ GWS from 2,200 monitoring wells shows that downscaling of GRACE data significantly improves agreement with in‐situ data, evidenced by higher Kling‐Gupta Efficiency (0.50–0.85) and correlation coefficients (0.60–0.95). Hotspots with the highest TWS and GWS decline rate between 2002 and 2023 were Dehli Doab (−442, −585 mm/year), BIST Doab (−367, −556 mm/year), Rajasthan (−242, −381 mm/year), and BARI (−188, −333 mm/year). Based on a general additive model, 47%–83% of the TWS decline was associated with anthropogenic stressors mainly due to increasing trends of crop sown area, water consumption, and human settlements. The decline rate of TWS and GWS anomalies was lower (i.e., −25 to −75 mm/year) in upstream sub‐regions (e.g., Yogo, Gilgit, Khurmong, Kabul) where climatic factors (downward shortwave radiations, air temperature, and sea surface temperature) explained 72%–91% of TWS/GWS changes. The relative influences of climatic and anthropogenic stressors varied across sub‐regions, underscoring the complex interplay of natural‐human activities in the basin. These findings inform place‐based water resource management in the Indus Basin by advancing the understanding of local vulnerabilities.
more »
« less
GRACE Downscaler: A Framework to Develop and Evaluate Downscaling Models for GRACE
Monitoring and managing groundwater resources is critical for sustaining livelihoods and supporting various human activities, including irrigation and drinking water supply. The most common method of monitoring groundwater is well water level measurements. These records can be difficult to collect and maintain, especially in countries with limited infrastructure and resources. However, long-term data collection is required to characterize and evaluate trends. To address these challenges, we propose a framework that uses data from the Gravity Recovery and Climate Experiment (GRACE) mission and downscaling models to generate higher-resolution (1 km) groundwater predictions. The framework is designed to be flexible, allowing users to implement any machine learning model of interest. We selected four models: deep learning model, gradient tree boosting, multi-layer perceptron, and k-nearest neighbors regressor. To evaluate the effectiveness of the framework, we offer a case study of Sunflower County, Mississippi, using well data to validate the predictions. Overall, this paper provides a valuable contribution to the field of groundwater resource management by demonstrating a framework using remote sensing data and machine learning techniques to improve monitoring and management of this critical resource, especially to those who seek a faster way to begin to use these datasets and applications.
more »
« less
- Award ID(s):
- 2019561
- PAR ID:
- 10429391
- Date Published:
- Journal Name:
- Remote Sensing
- Volume:
- 15
- Issue:
- 9
- ISSN:
- 2072-4292
- Page Range / eLocation ID:
- 2247
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract Prediction of dynamic environmental variables in unmonitored sites remains a long-standing challenge for water resources science. The majority of the world’s freshwater resources have inadequate monitoring of critical environmental variables needed for management. Yet, the need to have widespread predictions of hydrological variables such as river flow and water quality has become increasingly urgent due to climate and land use change over the past decades, and their associated impacts on water resources. Modern machine learning methods increasingly outperform their process-based and empirical model counterparts for hydrologic time series prediction with their ability to extract information from large, diverse data sets. We review relevant state-of-the art applications of machine learning for streamflow, water quality, and other water resources prediction and discuss opportunities to improve the use of machine learning with emerging methods for incorporating watershed characteristics and process knowledge into classical, deep learning, and transfer learning methodologies. The analysis here suggests most prior efforts have been focused on deep learning frameworks built on many sites for predictions at daily time scales in the United States, but that comparisons between different classes of machine learning methods are few and inadequate. We identify several open questions for time series predictions in unmonitored sites that include incorporating dynamic inputs and site characteristics, mechanistic understanding and spatial context, and explainable AI techniques in modern machine learning frameworks.more » « less
-
Groundwater historically has been a critical but understudied, underfunded, and underappreciated natural resource, although recent challenges associated with both groundwater quantity and quality have raised its profile. This is particularly true in the Laurentian Great Lakes (LGL) region, where the rich abundance of surface water results in the perception of an unlimited water supply but limited attention on groundwater resources. As a consequence, groundwater management recommendations in the LGL have been severely constrained by our lack of information. To address this information gap, a virtual summit was held in June 2021 that included invited participants from local, state, and federal government entities, universities, non-governmental organizations, and private firms in the region. Both technical (e.g., hydrologists, geologists, ecologists) and policy experts were included, and participants were assigned to an agricultural, urban, or coastal wetland breakout group in advance, based on their expertise. The overall goals of this groundwater summit were fourfold: (1) inventory the key (grand) challenges facing groundwater in Michigan; (2) identify the knowledge gaps and scientific needs, as well as policy recommendations, associated with these challenges; (3) construct a set of conceptual models that elucidate these challenges; and (4) develop a list of (tractable) next steps that can be taken to address these challenges. Absent this type of information, the sustainability of this critical resource is imperiled.more » « less
-
Abstract Predictions of hydrologic variables across the entire water cycle have significant value for water resources management as well as downstream applications such as ecosystem and water quality modeling. Recently, purely data‐driven deep learning models like long short‐term memory (LSTM) showed seemingly insurmountable performance in modeling rainfall runoff and other geoscientific variables, yet they cannot predict untrained physical variables and remain challenging to interpret. Here, we show that differentiable, learnable, process‐based models (calledδmodels here) can approach the performance level of LSTM for the intensively observed variable (streamflow) with regionalized parameterization. We use a simple hydrologic model HBV as the backbone and use embedded neural networks, which can only be trained in a differentiable programming framework, to parameterize, enhance, or replace the process‐based model's modules. Without using an ensemble or post‐processor,δmodels can obtain a median Nash‐Sutcliffe efficiency of 0.732 for 671 basins across the USA for the Daymet forcing data set, compared to 0.748 from a state‐of‐the‐art LSTM model with the same setup. For another forcing data set, the difference is even smaller: 0.715 versus 0.722. Meanwhile, the resulting learnable process‐based models can output a full set of untrained variables, for example, soil and groundwater storage, snowpack, evapotranspiration, and baseflow, and can later be constrained by their observations. Both simulated evapotranspiration and fraction of discharge from baseflow agreed decently with alternative estimates. The general framework can work with models with various process complexity and opens up the path for learning physics from big data.more » « less
-
Chen, Guohua; Khan, Faisal (Ed.)Artificial intelligence (AI) and machine learning (ML) are novel techniques to detect hidden patterns in environmental data. Despite their capabilities, these novel technologies have not been seriously used for real-world problems, such as real-time environmental monitoring. This survey established a framework to advance the novel applications of AI and ML techniques such as Tiny Machine Learning (TinyML) in water environments. The survey covered deep learning models and their advantages over classical ML models. The deep learning algorithms are the heart of TinyML models and are of paramount importance for practical uses in water environments. This survey highlighted the capabilities and discussed the possible applications of the TinyML models in water environments. This study indicated that the TinyML models on microcontrollers are useful for a number of cutting-edge problems in water environments, especially for monitoring purposes. The TinyML models on microcontrollers allow for in situ real-time environmental monitoring without transferring data to the cloud. It is concluded that monitoring systems based on TinyML models offer cheap tools to autonomously track pollutants in water and can replace traditional monitoring methods.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.3390/rs15092247
