Solid Water Melt Dominates the Increase of Total Groundwater Storage in the Tibetan Plateau

Geophysical Research Letters, Год журнала: 2022, Номер 49(18)

Опубликована: Сен. 13, 2022

Abstract Understanding how groundwater storage (GWS) responds to climate change is essential for water resources management and future availability in the Tibetan Plateau (TP). However, dominant factor controlling long‐term GWS changes remains unclear its responses are not well understood. Here we combined multi‐source datasets including in‐situ measurements, satellite observations, global models, reanalysis products reveal that increased at 5.59 ± 1.44 Gt/yr during 2003–2016 while showing spatial heterogeneities with increasing trends northern TP glacial regions declining central southern TP. The accelerated transformation from solid (glaciers, snow, permafrost; −17.72 1.53 Gt/yr) into liquid provide more recharge groundwater, dominating total increase. This study contributes a better understanding of hydrological cycle under provides key information projecting different scenarios

Язык: Английский

---

Long short-term memory models to quantify long-term evolution of streamflow discharge and groundwater depth in Alabama

The Science of The Total Environment, Год журнала: 2023, Номер 901, С. 165884 - 165884

Опубликована: Июль 29, 2023

Язык: Английский

---

A high-resolution, 3D groundwater-surface water simulation of the contiguous US: Advances in the integrated ParFlow CONUS 2.0 modeling platform

Journal of Hydrology, Год журнала: 2023, Номер 626, С. 130294 - 130294

Опубликована: Окт. 13, 2023

Large-scale, high-resolution hydrologic modeling is an important tool to address questions of water quantity, availability, and potential recharge. Continental-to-Global scale models, particularly those that include groundwater, are growing in number. However, many these approaches simplify aspects the system connections between surface groundwater. The ParFlow CONUS platform a large-scale, hyper-resolution, model relies on integrated solution 3D partial differential equations describe soil, 2D flow. prior version, 1.0, was first large-scale included explicit treatment lateral groundwater flow for contiguous US (CONUS). Here, we present 2.0 model. This extends coastlines contributing basins consistent with NOAA National Water Model. Here document roughly five years technical development this platform, steady-state simulation results, rigorously compare results 1.0 simulations, evaluate performance based observations. Simulated table depth streamflow were evaluated using more than 635K observations from USGS monitoring wells, other compiled datasets, NHD gauges. Our demonstrate improvement both simulations over generation all Hydrologic Unit Code (HUC) basins. These suggest current has good excellent entire CONUS, almost half HUC subbasins exhibiting normalized root-square error (RSR). metric not usually compared directly at studies, good-to-excellent exhibited some regions. We also delineate two regions influence performance, one where microtopography around streams dominates (D2), another mix subsurface heterogeneity topographic gradients dominate (D1). Improvements topography CONUS1 CONUS2 generally result better performances. Advancements structure produce estimates.

Язык: Английский

---

Diagnosing evapotranspiration responses to water deficit across biomes using deep learning

New Phytologist, Год журнала: 2023, Номер 240(3), С. 968 - 983

Опубликована: Авг. 25, 2023

Summary Accounting for water limitation is key to determining vegetation sensitivity drought. Quantifying effects on evapotranspiration (ET) challenged by the heterogeneity of types, climate zones and vertically along rooting zone. Here, we train deep neural networks using flux measurements study ET responses progressing drought conditions. We determine a stress factor (fET) that isolates reductions from atmospheric aridity other covarying drivers. regress fET against cumulative deficit, which reveals control whole‐column moisture availability. find variety stress. Responses range rapid declines 10% its water‐unlimited rate at several savannah grassland sites, mild in most forests, despite substantial deficits. Most sensitive are found arid warm sites. A combination regulation stomatal hydraulic conductance access belowground reservoirs, whether groundwater or soil moisture, could explain different behaviors observed across This not captured standard land surface model, likely reflecting simplifications representation storage.

Язык: Английский

---

Estimating Groundwater Pumping for Irrigation: A Method Comparison

Ground Water, Год журнала: 2023, Номер 62(1), С. 15 - 33

Опубликована: Июнь 24, 2023

Abstract Effective groundwater management is critical to future environmental, ecological, and social sustainability requires accurate estimates of withdrawals. Unfortunately, these are not readily available in most areas due physical, regulatory, challenges. Here, we compare four different approaches for estimating withdrawals agricultural irrigation. We apply methods a groundwater‐irrigated region the state Kansas, USA, where high‐quality withdrawal data evaluation. The represent broad spectrum approaches: (1) hydrologically‐based Water Table Fluctuation method (WTFM); (2) demand‐based SALUS crop model; (3) based on satellite‐derived evapotranspiration (ET) from OpenET; (4) landscape hydrology model which integrates hydrologic‐ approaches. applicability each approach varies availability, spatial temporal resolution, accuracy predictions. In general, our results indicate that all reasonably estimate region, however, type amount required computational requirements vary among For example, WTFM levels, specific yield, recharge data, whereas adequate information about type, land use, weather. This variability highlights difficulty identifying what how much, necessary reasonable estimate, suggests availability should drive choice approach. Overall, findings will help practitioners evaluate strengths weaknesses select appropriate their application.

Язык: Английский

---

Hyper-resolution PCR-GLOBWB: opportunities and challenges from refining model spatial resolution to 1 km over the European continent

Hydrology and earth system sciences, Год журнала: 2023, Номер 27(6), С. 1383 - 1401

Опубликована: Март 31, 2023

Abstract. The quest for hydrological hyper-resolution modelling has been on-going more than a decade. While global models (GHMs) have seen reduction in grid size, they thus far never consistently applied at (<=1 km) the large scale. Here, we present first application of GHM PCR-GLOBWB 1 km over Europe. We thoroughly evaluated simulated discharge, evaporation, soil moisture, and terrestrial water storage anomalies against long-term observations subsequently compared results with established 10 50 resolutions PCR-GLOBWB. Subsequently, could assess added value this version scale dependencies model forcing resolution. Eventually, these insights can help us understanding current challenges opportunities from formulating data requirements future improvements. found that, most variables, epistemic uncertainty is still large, issues commensurability exist respect to yet coarse used. Merely confidently state that output improves coarser due better representation river network km. However, currently available are not widely or lack sufficiently long time series, which makes it difficult performance other variables hyper-resolution. additional validation efforts needed. On side, applications require careful revisiting parameterization possibly implementation physical processes be able resemble dynamics spatial heterogeneity With km, contribute meeting grand challenge modelling. Even though was only assessed continental scale, valuable gained validity. As such, should as modest milestone on longer journey towards locally relevant output. This, however, requires community effort domain experts, developers, research software engineers, providers.

Язык: Английский

---

Physics-informed identification of PDEs with LASSO regression, examples of groundwater-related equations

Journal of Hydrology, Год журнала: 2024, Номер 638, С. 131504 - 131504

Опубликована: Июнь 15, 2024

Язык: Английский

---

Harnessing Machine Learning for Assessing Climate Change Influences on Groundwater Resources: A Comprehensive Review

Heliyon, Год журнала: 2024, Номер 10(17), С. e37073 - e37073

Опубликована: Авг. 28, 2024

Язык: Английский

---

Recent advances and opportunities in data assimilation for physics-based hydrological modeling

Frontiers in Water, Год журнала: 2022, Номер 4

Опубликована: Авг. 11, 2022

Data assimilation applications in integrated surface-subsurface hydrological models (ISSHMs) are generally limited to scales ranging from the hillslope local or meso-scale catchments. This is because ISSHMs resolve processes detail and a physics-based fashion therefore typically require intensive computational efforts rely on ground-based observations with small spatial support. At other end of spectrum, there vast body literature remote sensing data for land surface (LSMs) at continental even global scale. In LSMs, some usually represented coarse resolution empirical ways, especially groundwater lateral flows, which may be very important yet often neglected. Starting review recent progress multiple scales, we stress need find common ground between LSMs suggest possible ways forward advance use models.

Язык: Английский

---

Groundwater Level Prediction with Deep Learning Methods

Water, Год журнала: 2023, Номер 15(17), С. 3118 - 3118

Опубликована: Авг. 30, 2023

The development of civilization and the preservation environmental ecosystems are strongly dependent on water resources. Typically, an insufficient supply surface resources for domestic, industrial, agricultural needs is supplemented with groundwater However, a natural resource that must accumulate over many years cannot be recovered after short period recharge. Therefore, long-term management important issue sustainable development. accurate prediction levels first step in evaluating total their allocation. process data collection, may lost due to various factors. Filling missing main problem any research field address. It well known maintain integrity, one effective approach value imputation (MVI). In addition, it has been demonstrated machine learning better tool. purpose this study was utilize generative adversarial network (GAN) consists model discriminative imputation. Although GAN could not capture level endpoints every section, overall simulation performance still excellent some extent. Our results show can improve accuracy evaluations. current study, two interdisciplinary deep methods, univariate Seq2val (sequence-to-value), were used estimation. addition addressing significance parameter conditions, advantages disadvantages these models hydrological simulations also discussed compared. Regarding selection, analysis than those analysis. Finally, employed examine limits simulations. suggest CNNs better, while LSTM multistep prediction. beneficial providing evaluation

Язык: Английский

---