About

Ali Behrangi is a Professor of Hydrology and Atmospheric Sciences and an Associate Professor of Civil Engineering and Architectural Engineering and Mechanics. He is also an Associate Professor of Geosciences and a member of the Graduate Faculty at The University of Arizona. His research focuses on hydrology, atmospheric sciences, and related geosciences, contributing to the understanding of water and atmospheric processes. He is actively involved in academic and research activities within the Department of Civil & Architectural Engineering & Mechanics, engaging in teaching, research, and service to the university community.

Research topics

• Computer Science
• Meteorology
• Geology
• Environmental science
• Machine Learning
• Remote sensing
• Geography
• Data Mining
• Atmospheric sciences
• Physics
• Climatology
• Geodesy
• Mathematics
• Cartography
• Statistics
• Oceanography
• Algorithm

Selected publications

• Process-Aware AI for Rainfall-Runoff Modeling: A Mass-Conserving Neural Framework with Hydrological Process Constraints

  arXiv (Cornell University) · 2026-03-26

  preprintOpen access

  Machine learning models can achieve high predictive accuracy in hydrological applications but often lack physical interpretability. The Mass-Conserving Perceptron (MCP) provides a physics-aware artificial intelligence (AI) framework that enforces conservation principles while allowing hydrological process relationships to be learned from data. In this study, we investigate how progressively embedding physically meaningful representations of hydrological processes within a single MCP storage unit improves predictive skill and interpretability in rainfall-runoff modeling. Starting from a minimal MCP formulation, we sequentially introduce bounded soil storage, state-dependent conductivity, variable porosity, infiltration capacity, surface ponding, vertical drainage, and nonlinear water-table dynamics. The resulting hierarchy of process-aware MCP models is evaluated across 15 catchments spanning five hydroclimatic regions of the continental United States using daily streamflow prediction as the target. Results show that progressively augmenting the internal physical structure of the MCP unit generally improves predictive performance. The influence of these process representations is strongly hydroclimate dependent: vertical drainage substantially improves model skill in arid and snow-dominated basins but reduces performance in rainfall-dominated regions, while surface ponding has comparatively small effects. The best-performing MCP configurations approach the predictive skill of a Long Short-Term Memory benchmark while maintaining explicit physical interpretability. These results demonstrate that embedding hydrological process constraints within AI architectures provides a promising pathway toward interpretable and process-aware rainfall-runoff modeling.

  PublisherDOI

• LACC: A lightweight attention-conditional convolution network for long-term Wetland classification

  Environmental Research · 2026-04-30

  article
  PublisherDOI

• Urban rainfall trends in IMERG datasets

  Environmental Research Letters · 2026-05-21

  articleOpen accessSenior author

  Abstract Satellite precipitation records are increasingly used to assess whether urbanisation modifies local rainfall. However, it remains unclear whether reported urban signals reflect physical processes or artifacts from evolving satellite observing systems. The Integrated Multi-satellitE Retrievals for GPM (IMERG) merges observations from a growing constellation of microwave and infrared sensors, meaning that long-term trends derived from this product may partly reflect changes in the observing system rather than true precipitation changes. Here we analyse rainfall frequency and intensity over 15 major global cities spanning diverse climate regimes using IMERG Version 07B and show that urban areas exhibit a consistent increase in rainfall event frequency with a weaker enhancement in intensity relative to surrounding rural areas. Both signals are dominated by microwave-based retrievals, while infrared-dominated periods largely suppress the urban hotspot pattern, highlighting the sensitivity of detected urban signals to the underlying retrieval type. To isolate physical rainfall changes from observational artifacts, we develop a synthetic time series approach that quantifies the contribution of systematically increasing microwave sampling frequency to apparent long-term trends. We find that sampling artifacts explain up to 20% of observed long-term trends, with locally higher contributions in some cities. After accounting for these effects, the urban rainfall enhancement persists across cities, demonstrating that IMERG captures a robust and physically consistent urban signal in both rainfall frequency and intensity. These findings have direct implications for the reliability of satellite-based urban climate assessments, the interpretation of long-term precipitation trends, and the design of future observing systems.

  PublisherDOI

• The Newly Released Global Precipitation Climatology Project (GPCP) V3.3 Daily and Monthly Products and the Future Plans

  2026-03-13

  articleOpen access1st authorCorresponding

  The Global Precipitation Climatology Project (GPCP) provides a widely used satellite–gauge merged precipitation dataset designed to meet Climate Data Record (CDR) standards for long-term consistency and homogeneity. The latest release, Version 3.3 of the GPCP Daily (1998–2024) and Monthly (1983–2024) products, issued in February 2025, represents the final generation before the transition to GPCP Version 4. This presentation summarizes the V3.3 products and their satellite–gauge inputs, compares them with Version 3.2, and highlights major updates. It also includes evaluations over the global oceans using Passive Aquatic Listeners (PALs), buoys, and atolls, assessments over sea ice using snow-depth data from ICESat-2, CryoSat-2, and ERA5, and analyses over Antarctica using CloudSat, together with insights from GPM Version 07. Key upgrades in GPCP V3.3 include adoption of GPROF 2021 for passive microwave retrievals, a revised ocean climatology based on updated GPM and TRMM radar and microwave data, sensor-specific adjustments to GPROF-calibrated PERSIANN-CDR, and the introduction of a new absolute bias error variable. Relative to V3.2, V3.3 shows an approximately 11% increase in global ocean precipitation and a 9% global increase, driven mainly by ocean changes, while land precipitation changes are small (about 1%). Initial ocean evaluations using limited in situ data indicate a slight overestimation in V3.3, although energy-budget closure supports the overall increase. Interannual variability is also slightly larger, while regional and global precipitation trends remain largely unchanged. Enhancements in the GPCP V3.3 Daily product stem from updates to the Monthly analysis and incorporation of IMERG V07B Final Run, which uses GridSat to extend daily coverage back to January 1998 through May 2000. The presentation concludes with plans for GPCP V4, focusing on higher resolution, lower latency, and more advanced retrieval and gauge-analysis techniques.

  PublisherDOI

• Process-Aware AI for Rainfall-Runoff Modeling: A Mass-Conserving Neural Framework with Hydrological Process Constraints

  ArXiv.org · 2026-03-26

  articleOpen access

  Machine learning models can achieve high predictive accuracy in hydrological applications but often lack physical interpretability. The Mass-Conserving Perceptron (MCP) provides a physics-aware artificial intelligence (AI) framework that enforces conservation principles while allowing hydrological process relationships to be learned from data. In this study, we investigate how progressively embedding physically meaningful representations of hydrological processes within a single MCP storage unit improves predictive skill and interpretability in rainfall-runoff modeling. Starting from a minimal MCP formulation, we sequentially introduce bounded soil storage, state-dependent conductivity, variable porosity, infiltration capacity, surface ponding, vertical drainage, and nonlinear water-table dynamics. The resulting hierarchy of process-aware MCP models is evaluated across 15 catchments spanning five hydroclimatic regions of the continental United States using daily streamflow prediction as the target. Results show that progressively augmenting the internal physical structure of the MCP unit generally improves predictive performance. The influence of these process representations is strongly hydroclimate dependent: vertical drainage substantially improves model skill in arid and snow-dominated basins but reduces performance in rainfall-dominated regions, while surface ponding has comparatively small effects. The best-performing MCP configurations approach the predictive skill of a Long Short-Term Memory benchmark while maintaining explicit physical interpretability. These results demonstrate that embedding hydrological process constraints within AI architectures provides a promising pathway toward interpretable and process-aware rainfall-runoff modeling.

  PublisherOA PDF

• Assessment of Penman-Monteith and Priestley-Taylor Based Evapotranspiration Models Across the United States Aridity Gradients Using AmeriFlux Observations and Deep Learning

  Zenodo (CERN European Organization for Nuclear Research) · 2026-05-22

  datasetOpen access

  This archive contains the processed data, derived model outputs, metadata, and Python scripts required to reproduce all figures presented in the manuscript and supporting information associated with the manuscript "Assessment of Penman-Monteith and Priestley-Taylor Based Evapotranspiration Models Across the United States Aridity Gradients Using AmeriFlux Observations and Deep Learning". The package reproduces 10 main manuscript figures and 5 supporting information figures using archived figure-ready datasets and processed model outputs. Included materials consist of performance metrics, feature-importance summaries, explainability outputs, site metadata, and processed daily evapotranspiration data for representative Arizona AmeriFlux sites. The archive is intended for figure reproduction and verification of published results. It does not include raw AmeriFlux observations, AORC meteorological forcing data, MODIS products, ERA5-Land data, model training datasets, or the complete preprocessing and machine-learning training workflows. Instead, it provides the processed inputs necessary to regenerate all published figures exactly as presented in the manuscript. Figure reproduction can be performed by installing the required Python dependencies and running: python scripts/03_reproduce_all_figures.py Expected outputs are 10 manuscript figures and 5 supporting-information figures generated in the figures/reproduced directory.

  PublisherDOI

• A Review of the Past Half Century of Geostationary Satellite Thermal Observations for Global Precipitation Estimation: Developments, achievements, and future prospects

  IEEE Geoscience and Remote Sensing Magazine · 2026-01-01

  article

  Over the past half century since 1975, the development and improvement of meteorological geostationary (GEO) satellites have played a pivotal role in observing cloud dynamics and subsequently in advancing spaceborne precipitation estimation. Infrared (IR) observation from GEO satellites offers unique advantages, such as broad spatial coverage, high temporal resolution, and long-term consistency, motivating extensive research to unlock the potential of GEO IR-based data for capturing precipitation structure and dynamics. This article reviews the major developments and milestone achievements of GEO IR-based precipitation estimation over the past five decades and summarizes the future prospects, including potential directions and remaining challenges. By examining the history of global GEO satellites and more than 100 references in this domain, we categorize the development of GEO IR-based precipitation estimation methodology and technology into three distinct stages: 1) the Exploration Phase (1975 to ca. 1995), 2) the Growth Phase (ca. 1995 to ca. 2015), and 3) the Exploitation Phase (ca. 2015 to the present). As we transition into an emerging new stage, these efforts collectively point toward multispectral retrievals, lifecycle-aware machine learning (ML), smart sensing, and advanced multisource integration as key directions shaping the future of GEO IR-based precipitation estimation. In summary, GEO IR-based precipitation estimation has made substantial contributions over the past half century and will play an increasingly important role with great potential in future precipitation science.

  PublisherDOI

• Assessment of Penman-Monteith and Priestley-Taylor Based Evapotranspiration Models Across the United States Aridity Gradients Using AmeriFlux Observations and Deep Learning

  Zenodo (CERN European Organization for Nuclear Research) · 2026-05-22

  datasetOpen access

  This archive contains the processed data, derived model outputs, metadata, and Python scripts required to reproduce all figures presented in the manuscript and supporting information associated with the manuscript "Assessment of Penman-Monteith and Priestley-Taylor Based Evapotranspiration Models Across the United States Aridity Gradients Using AmeriFlux Observations and Deep Learning". The package reproduces 10 main manuscript figures and 5 supporting information figures using archived figure-ready datasets and processed model outputs. Included materials consist of performance metrics, feature-importance summaries, explainability outputs, site metadata, and processed daily evapotranspiration data for representative Arizona AmeriFlux sites. The archive is intended for figure reproduction and verification of published results. It does not include raw AmeriFlux observations, AORC meteorological forcing data, MODIS products, ERA5-Land data, model training datasets, or the complete preprocessing and machine-learning training workflows. Instead, it provides the processed inputs necessary to regenerate all published figures exactly as presented in the manuscript. Figure reproduction can be performed by installing the required Python dependencies and running: python scripts/03_reproduce_all_figures.py Expected outputs are 10 manuscript figures and 5 supporting-information figures generated in the figures/reproduced directory.

  PublisherDOI

• How Do IMERG V07, IMERG V06, and ERA5 Precipitation Products Perform over Snow–Ice-Free and Snow–Ice-Covered Surfaces at a Range of Near-Surface Temperatures?

  Journal of Hydrometeorology · 2025-05-19 · 4 citations

  articleOpen accessSenior author

  Abstract Accurate precipitation estimation is essential for hydrological research and applications. This study assesses the performance of Integrated Multi-satellitE Retrievals for Global Precipitation Measurement (GPM), version 7 (IMERG V07), IMERG V06, and the fifth major global reanalysis produced by the European Centre for Medium-Range Weather Forecasts (ERA5) over snow–ice-covered and snow–ice-free surfaces, using 3 years (2018–20) of Multi-Radar Multi-Sensor (MRMS) system gauge-adjusted data as a reference. Mountainous regions were largely excluded from the analysis due to known uncertainties in precipitation estimates over complex terrain. Besides the IMERG final products, the study compares precipitation estimates from the infrared (IR) and passive microwave (PMW) components of IMERG V07 under different surface and environmental conditions. This is particularly relevant as both calibrated IR and PMW products, which rely on the Combined Radar–Radiometer Algorithm (CORRA) for calibration, face significant limitations over cold regions, especially over snow–ice-covered surfaces, complicating the choice of products for integration in the merged products like IMERG. We found that IMERG V07 offers notable improvements over IMERG V06 and generally outperforms ERA5 over snow–ice-free regions, demonstrating enhanced accuracy in precipitation intensity and spatial coverage. Conversely, ERA5 outperforms IMERG V07 over snow–ice surfaces, highlighting remaining challenges in satellite-based precipitation products over cold regions. An evaluation of PMW precipitation products indicates that while they generally perform better than IR precipitation products in warmer conditions, IR precipitation is still invaluable in cold regions with snow–ice cover. Among the PMW products and over snow–ice surfaces, Advanced Microwave Scanning Radiometer, version 2 (AMSR2), underperforms other PMW precipitation products for most statistical metrics, while GPM Microwave Imager (GMI), Special Sensor Microwave Imager/Sounder (SSMIS), and Microwave Humidity Sounder (MHS) products perform relatively better than others. The results emphasize the need for improving spaceborne sensors and algorithms to improve their accuracy across diverse environmental conditions, especially over cold regions in the presence of snow or ice on the surface.

  PublisherOA PDFDOI

• Do land models miss key soil hydrological processes controlling soil moisture memory?

  2025-08-19

  reportOpen access

  Soil moisture memory is critical for understanding climatic, hydrological, and ecosystem interactions. Most land surface models overestimate surface soil moisture and its persistency, sustaining spuriously large soil surface evaporation during dry-down periods. Do LSMs miss or misrepresent key hydrological processes controlling SMM? We used Noah-MP with advanced hydrology that represents preferential flow and surface ponding and provides optional schemes of soil hydraulics. Effects were tested, which are generally missed by LSMs in SMM. We compare SMMs computed from various Noah-MP configurations against that derived from the Soil Moisture Active Passive L₃ soil moisture and in situ measurements from the International Soil Moisture Network between 2015 to 2019 over the contiguous US. Results suggest soil hydraulics plays a dominant role and the Van Genuchten hydraulic scheme reduces overestimation of the long-term surface SMM produced by the Brooks–Corey scheme; explicitly representing surface ponding enhances SMM for the surface layer and the root zone; and representing preferential flow improves overall representation of soil moisture dynamics. The combination of these missing schemes can significantly improve the longterm memory overestimation and short-term memory underestimation issues in LSMs. LSMs for use in seasonal-to-subseasonal climate prediction should, at least, adopt the Van Genuchten hydraulic scheme.

  PublisherOA PDFDOI

Frequent coauthors

• Kuolin Hsu

  University of California, Irvine

  54 shared

• Soroosh Sorooshian

  University of California, Irvine

  41 shared

• Yang Hong

  University of Oklahoma

  36 shared

• Bjorn Lambrigtsen

  Jet Propulsion Laboratory

  33 shared

• Robert J. Kuligowski

  29 shared

• B. Imam

  29 shared

• George J. Huffman

  Goddard Space Flight Center

  28 shared

• Mohammad Reza Ehsani

  University of Arizona

  22 shared

Education

• Postdoctoral scholar

  California Institute of Technology

  2012

• PhD, Civil Eng., Remote sensing, hydrology, water resources

  University of California Irvine

  2009

• MSc, Civil and Environmental Engineering

  Sharif University of Technology

  2004

• Bsc, Civil and Environmental Engineering

  Sharif University of Technology

  2003