A5028727891- Hydrology and Watershed Management Studies
- Cryospheric studies and observations
- Hydrological Forecasting Using AI
- Climate change and permafrost
- Meteorological Phenomena and Simulations
- Precipitation Measurement and Analysis
- Innovative Teaching and Learning Methods
- demographic modeling and climate adaptation
- Healthcare Policy and Management
- Seismic Imaging and Inversion Techniques
- Soil and Unsaturated Flow
- Hydraulic Fracturing and Reservoir Analysis
- Geophysical Methods and Applications
- Merger and Competition Analysis
- E-Learning and COVID-19
- Education and Critical Thinking Development
- Groundwater flow and contamination studies
University of Arizona
2022-2025
Use of complex, high-resolution integrated hydrologic models offer the most comprehensive and detailed representations groundwater, surface water, land processes, but are challenging to use for forecasting tasks due high computational costs parameter uncertainty. On flipside, machine learning approaches highly accurate can be computationally frugal targeted tasks, difficult audit must retrained adapt new or domains.In this work we present several case studies using deep surrogate modeling...
Abstract Hydrogeologic models generally require gridded subsurface properties, however these inputs are often difficult to obtain and highly uncertain. Parametrizing computationally expensive where extensive calibration is infeasible a long standing challenge in hydrogeology. Here we present machine learning framework address this challenge. We train an inversion model learn the relationship between water table depth hydraulic conductivity using small number of physical simulations. For 31M...
Abstract Integrated hydrologic models can simulate coupled surface and subsurface processes but are computationally expensive to run at high resolutions over large domains. Here we develop a novel deep learning model emulate flows simulated by the integrated ParFlow‐CLM across contiguous US. We compare convolutional neural networks like ResNet UNet autoregressively against our architecture called Forced SpatioTemporal RNN (FSTR). The FSTR incorporates separate encoding of initial conditions,...
Abstract. The Heihe River basin in northwest China depends heavily on both anthropogenic and natural storage (e.g., surface reservoirs, rivers groundwater) to support economic environmental functions. Qilian Mountain cryosphere the upper is integral recharging these supplies. It well established that climate warming driving major shifts high-elevation water through loss of glaciers permafrost. However, impacts groundwater–surface-water interactions supply corresponding lower reaches are less...
Background: With the proliferation of online courses in nursing education and professional staff development, future nurse educators must be prepared to teach online. The purpose this article is present an educational innovation created evaluated prepare develop, design, deliver learning module for distance education. Method: A combination instructional scaffolding applied was used educator students how facilitate module. Results: Analyses student assignment scores, reflections, faculty...
Defnet et al., (2024). hf_hydrodata: A Python package for accessing hydrologic simulations and observations across the United States. Journal of Open Source Software, 9(99), 6623, https://doi.org/10.21105/joss.06623
Hydrologic models are an integral part of understanding and managing water supply.There countless hydrologic available that differ in their complexity, scale focus on different parts the cycle.ParFlow is a fully integrated, physics-based model simulates surface subsurface flow simultaneously (Ashby & Falgout, 1996;Jones Woodward, 2001;Kollet Maxwell, 2006;Maxwell, 2013).ParFlow also coupled with land which allows it to simulate full terrestrial cycle from bedrock treetops (Kollet...
Integrated hydrologic models can simulate coupled surface and subsurface processes but are computationally expensive to run at high resolutions over large domains. Here we develop a novel deep learning model emulate continental-scale flows simulated by the integrated ParFlow-CLM model. We compare convolutional neural networks like ResNet UNet autoregressively against our architecture called Forced SpatioTemporal RNN (FSTR). The FSTR incorporates separate encoding of initial conditions,...
Abstract. The Heihe River Basin in Northwestern China depends heavily on both manmade and natural storage (e.g., surface reservoirs, rivers, groundwater) to support economic environmental functions. Qilian Mountain cryosphere the upper basin is integral recharging these supplies. It well established that climate warming driving major shifts high elevation water through loss of glaciers permafrost. However, impacts groundwater-surface interactions supply corresponding lower reaches are less...