Explainable AI reveals new hydroclimatic insights for ecosystem-centric groundwater management
Bibliographic Data
| ID | 15546100 |
|---|---|
| Authors | Debaditya Chakraborty (0000-0002-2165-9440, The University of Texas at San Antonio, corresponding author), Hakan Başağaoğlu (0000-0002-0729-0296), Lilianna Gutierrez (0000-0003-1257-8048, The University of Texas at San Antonio), Ali Mirchi (0000-0002-9649-2964, Oklahoma State University) |
| Year | 2021 |
| Volume | 16 |
| Issue | 11 |
| Pages | 114024-114024 |
| Publication date | 2021-10-15 |
| Open Access | Yes |
| Type | ARTICLE |
| Venue | Environmental Research Letters (JOURNAL) |
| Journal identifiers | ISSN: 1748-9326 • E-ISSN: 1748-9326 |
| Publisher | IOP Publishing (PUBLISHER • GB) |
| DOI | 10.1088/1748-9326/ac2fde |
| OpenAlex | W3205832097 |
| Language | EN |
| Citations received | 4 |
| References cited | 44 |
Trustworthy projections of hydrological droughts are pivotal for identifying the key hydroclimatic factors that affect future groundwater level (GWL) fluctuations in drought-prone karstic aquifers that provide water for human consumption and sustainable ecosystems. Herein, we introduce an explainable artificial intelligence (XAI) framework integrated with scenario-based downscaled climate projections from global circulation models. We use the integrated framework to investigate nonlinear hydroclimatic dependencies and interactions behind future hydrological droughts in the Edwards Aquifer Region, an ecologically fragile groundwater-dependent semi-arid region in southern United States. We project GWLs under different future climate scenarios to evaluate the likelihood of severe hydrological droughts under a warm-wet future in terms of mandated groundwater pumping reductions in droughts as part of the habitat conservation plan in effect to protect threatened and endangered endemic aquatic species. The XAI model accounts for the expected non-linear dynamics between GWLs and climatic variables in the complex human-natural system, which is not captured in simple linear models. The XAI-based analysis reveals the critical temperature inflection point beyond which groundwater depletion is triggered despite increased average precipitation. Compound effects of increased evapotranspiration, lower soil moisture, and reduced diffuse recharge due to warmer temperatures could amplify severe hydrological droughts that lower GWLs, potentially exacerbating the groundwater sustainability challenges in the drought-prone karstic aquifer despite an increasing precipitation trend
Aquifer · Biome · Climate change · Ecohydrology · Ecosystem · Evapotranspiration · Groundwater · Groundwater recharge · Hydrology (agriculture · Sustainability · Water resource management · Water resources · Wetland · Environmental Science · Hydrological Forecasting Using AI · Hydrology and Drought Analysis · Hydrology and Watershed Management Studies · Ecology · Geology
Constraints on future changes in climate and the hydrologic cycle
Evidence for intensification of the global water cycle
TerraClimate, a high-resolution global dataset of monthly climate and climatic water balance from 1958–2015
Hydrological drought explained
From local explanations to global understanding with explainable AI for trees
Explaining prediction models and individual predictions with feature contributions
A Survey of Methods for Explaining Black Box Models
The Mythos of Model Interpretability
Rcp4.5
RCP 8.5—A scenario of comparatively high greenhouse gas emissions
Long-term climatic change and sustainable ground water resources management
Domestic well vulnerability to drought duration and unsustainable groundwater management in California’s Central Valley
Vulnerability to the impact of climate change on renewable groundwater resources
How can Big Data and machine learning benefit environment and water management
Machine learning and artificial intelligence to aid climate change research and preparedness
| Citations per year | 1.33 |
|---|---|
| Citation span | 2023 - 2026 (4) |
| Highly cited | No |