Planetary Water Dynamics Lab | Global Hydrology & Remote Sensing

Our Work

Funded Projects

Innovative research supported by NASA, USGS, Google, and other leading agencies.

NASA 2024-2025

Improving Groundwater Withdrawal Estimation in Arizona

Enhancing machine learning and remote sensing-based model estimates to provide actionable withdrawal data amidst Colorado River reductions.

PI: Dr. Ryan Smith (CSU), Co-I: Dr. Sayantan Majumdar (DRI)

Google 2025-2026

Irrigation Status Mapping using Satellite Embeddings

Developing annual irrigation status maps (2017-2024) for the western U.S. at 30m resolution, integrating satellite embeddings into the IrrMapper framework.

PI: Dr. Sayantan Majumdar (DRI)

NMCG India 2026-2028

Digital Twin for the Ganga Basin

Developing an advanced Digital Twin and Water Cycle Atlas using AI, satellite remote sensing, and hydrological modeling for the Ganga River Basin.

Co-PI: Dr. Sayantan Majumdar

USGS / DOI Active

OpenET Planning & Development

Multiple phases of development for the OpenET platform, enhancing software tools, datasets, and integrating data into national hydrologic models.

Role: Faculty / Researcher

DOD-Army 2025-2027

Groundwater in Transboundary Watersheds

Characterizing groundwater resources in eight international watersheds using satellite remote sensing, hydroclimate, and land surface models.

Co-PI: Dr. Sayantan Majumdar

State of Nevada 2023-2026

Nevada Water Resources Initiative

Estimating water use across Nevada associated with agriculture and natural groundwater discharge areas to support state water planning.

Role: Faculty

Sayantan "Monty" Majumdar, Ph.D.

Lab Director & Assistant Research Professor

Desert Research Institute

sayantan.majumdar@dri.edu

DRI Profile

Education

Ph.D. in Geological Engineering

Missouri University of Science and Technology, USA

2019 - 2022

M.Sc. in Geoinformatics (Cum Laude)

Faculty ITC, University of Twente, Netherlands

2017 - 2019

M.Sc. in Computer Science

St. Xavier's College (Autonomous) Kolkata, India

2015 - 2017

B.Sc. in Computer Science (Hons.)

St. Xavier's College (Autonomous) Kolkata, India

2012 - 2015

Recent Experience

2024-

Adjunct Faculty

University of Nevada, Reno, NV
2023-

Assistant Research Professor

Desert Research Institute (DRI), Reno, NV
2022-23

Postdoctoral Fellow

Colorado State University, Fort Collins, CO
2022

Research Scientist Intern

Meta Platforms, Inc., Menlo Park, CA
2021

Analytics Modeling Intern

Planet Labs, Rolla, MO (Remote)

Knowledge Base

Selected Publications

Banerjee, S., Majumdar, S., Saha, J., Kukal, M. S., Thakur, P. K., Rathore, V. S., … Ndehedehe, C. (2025). Groundwater Potential Mapping in India: A Review of Approaches and Pathways for Sustainable Management. Cambridge Prisms: Drylands, 1–47. DOI

Parasar, P., Moral, P., Srivastava, A., Krishna, A. P., Majumdar, S., et al. (2025). Integrating genetic algorithms and machine learning for spatiotemporal groundwater potential zoning in fractured aquifers. Journal of Hydrology: Regional Studies, 62, 102800. DOI

Hasan, M. F., Smith, R. G., Majumdar, S., Huntington, J. L., Alves Meira Neto, A., & Minor, B. A. (2025). Satellite data and physics-constrained machine learning for estimating effective precipitation in the Western United States and application for monitoring groundwater irrigation. Agricultural Water Management, 319, 109821. DOI

Asfaw, D., Smith, R. G., Majumdar, S., Grote, K., Fang, B., Wilson, B. B., Lakshmi, V., & Butler, J. J. (2025). Predicting groundwater withdrawals using machine learning with limited metering data: Assessment of training data requirements. Agricultural Water Management, 318, 109691. DOI

Bailey, R. T., Abbas, S., Čerkasova, N., Arnold, J. G., White, M. J., Majumdar, S., & Smith, R. (2025). Quantifying agricultural groundwater abstraction using an integrated watershed modeling approach, Mississippi Delta, USA. Hydrogeology Journal, 33(5), 1429–1447. DOI

Ott, T. J., Majumdar, S., Huntington, J. L., Pearson, C., Bromley, M., Minor, B. A., ReVelle, P., Morton, C. G., Sueki, S., Beamer, J. P., & Jasoni, R. L. (2024). Toward field-scale groundwater pumping and improved groundwater management using remote sensing and climate data. Agricultural Water Management, 302, 109000. DOI

Majumdar, S., Smith, R. G., Hasan, M. F., Wilson, J. L., White, V. E., Bristow, E. L., Rigby, J. R., Kress, W. H., & Painter, J. A. (2024). Improving crop-specific groundwater use estimation in the Mississippi Alluvial Plain: Implications for integrated remote sensing and machine learning approaches in data-scarce regions. Journal of Hydrology: Regional Studies, 52, 101674. DOI

Tolan, J., Yang, H.-I., ... Majumdar, S., et al. (2024). Very high resolution canopy height maps from RGB imagery using self-supervised vision transformer and convolutional decoder trained on aerial lidar. Remote Sensing of Environment, 300, 113888. DOI

M. F. Hasan, R. Smith, S. Vajedian, R. Pommerenke, & S. Majumdar. (2023). Global land subsidence mapping reveals widespread loss of aquifer storage capacity. Nature Communications, 14, 6180. DOI

Majumdar, S., Smith, R. G., Conway, B. D., & Lakshmi, V. (2022). Advancing remote sensing and machine learning-driven frameworks for groundwater withdrawal estimation in Arizona. Hydrological Processes, 36(11). DOI

Majumdar, S., Smith, R. G., Butler Jr, J. J., & Lakshmi, V. (2020). Groundwater withdrawal prediction using integrated multitemporal remote sensing data sets and machine learning. Water Resources Research, 56. DOI

Smith, R., & Majumdar, S. (2020). Groundwater storage loss associated with land subsidence in Western United States mapped using machine learning. Water Resources Research, 56. DOI

Deep Dive

Paper Highlights

A closer look at the methodology and results from our most impactful recent work.

Fig 1. Field-scale groundwater pumping estimates using remote sensing and climate data in Nevada and Oregon.

Satellite Remote Sensing for Irrigation Water Use Monitoring

Satellite-based Field-Scale Groundwater Pumping Estimates for Nevada and Oregon

Paper: Ott, T.J., Majumdar, S., et al. (2024). Agricultural Water Management.

To address the challenge of monitoring groundwater overdraft in the western United States where flow meters are scarce, this study evaluates the use of Landsat-based evapotranspiration data from OpenET to estimate groundwater pumping. By pairing OpenET estimates of consumptive use with metered data from Diamond Valley, Nevada, and Harney Basin, Oregon, researchers developed models to establish a robust relationship between net evapotranspiration and pumping. The results demonstrated high accuracy, with the OpenET ensemble mean achieving field-scale error rates of approximately 11–14% and explaining nearly 90% of the variance in pumping volumes. Ultimately, the study confirms that satellite-based estimates are effective tools for not only estimating current and historical pumping volumes but also for performing quality control on existing metered data.

Read Full Paper

Fig 2. Crop-specific groundwater extraction estimates across the MAP.

Machine Learning and Remote Sensing

Crop-Specific Groundwater Use in the Mississippi Alluvial Plain

Paper: Majumdar, S., et al. (2024). Journal of Hydrology: Regional Studies.

The Mississippi Alluvial Plain (MAP) is one of the most productive agricultural regions in the world. This study addressed the challenge of data scarcity by integrating flowmeter datasets with remote sensing predictors (ET, precipitation, crop type). The resulting model allows us to disentangle water use by crop type—showing distinct withdrawal patterns (e.g., rice versus soybeans)—which is vital for targeted water conservation policies in the region.

Read Full Paper

Global Hydrology

Widespread Loss of Aquifer Storage Capacity

Paper: Hasan, M. F., Smith, R. G., Vajedian, S., Pommerenke, R., and Majumdar, S. (2023). Nature Communications.

In this study, we utilized Sentinel-1 InSAR data and Random Forests to map land subsidence at a global scale. The key finding, illustrated in the figure, reveals that permanent compaction of aquifer systems is far more widespread than previously documented. We estimated the permanent loss of storage capacity—a critical metric for long-term water security that is often overlooked in standard groundwater assessments.

Read Full Paper

Fig 3. Groundwater withdrawal-induced global land subsidence predicted by the random forest model.

Physics-Informed AI

Physics-Constrained ML for Effective Precipitation

Paper: Hasan, M. F., Smith, R. G., Majumdar, S., et al. (2025). Agricultural Water Management.

While remote sensing tools like OpenET provide accurate evapotranspiration (ET) estimates, distinguishing the specific contributions of precipitation versus irrigation remains difficult due to runoff and deep percolation. We hypothesized that ET observed over rainfed vegetation could serve as a proxy for effective precipitation. We validated this approach using machine learning, demonstrating that we can now produce robust, scalable estimates of effective precipitation across the Western United States. Read Full Paper

Fig 4. Physics-constrained machine learning (PCML) for effective precipitation.

Fig 5. High-resolution canopy height predicted from RGB imagery using Vision Transformers.

GeoAI & Computer Vision

Vision Transformers for Global Canopy Height

Paper: Tolan, J. ... Majumdar, S., et al. (2024). Remote Sensing of Environment.

Estimating vegetation structure usually requires expensive LiDAR. In this collaboration, we demonstrated how Self-Supervised Vision Transformers (ViT) can reconstruct canopy height maps using only standard high-resolution RGB imagery. This breakthrough allows for scalable monitoring of biomass and vegetation health without the cost of continuous active sensor deployments.

Read Full Paper

Machine Learning for Groundwater Pumping

Assesing Training Data Requirements for ML-based Groundwater Pumping Prediction

Paper: Asfaw, D., Smith, R. G., Majumdar, S., et al. (2025). Agricultural Water Management.

While various methods exist to predict agricultural groundwater withdrawals, a machine learning framework integrating remote sensing and climate data has proven effective at bypassing traditional challenges like linking wells to specific fields. Validated across regions including Kansas and Arizona, this approach focuses on providing consistent gridded information for numerical hydrologic modeling. Addressing the global scarcity of metering data, the study determined that accurate pumping estimates are achievable at a 2-km scale using training data from only 10% of available wells. The results indicate that spatially aggregating data on a grid, driven by climate variables and evapotranspiration, substantially reduces the uncertainty common in point-scale predictions. Read Full Paper

Fig 6. Cover figure showing the workflow for predicting groundwater withdrawals in Kansas (Groundwater Management District 4) using ML.

Policy & Review

Sustainable Groundwater Pathways for India

Paper: Banerjee, S., Majumdar, S., et al. (2025). Cambridge Prisms: Drylands.

India faces some of the world's most severe groundwater depletion challenges. In this comprehensive review (2025), we analyzed decades of groundwater potential mapping approaches. The paper identifies critical gaps in current methodologies and proposes a new framework integrating AI and social hydrology to move from simple "potential mapping" to actionable sustainable management strategies.

Read Full Paper

Fig 7. Framework for integrating physical and social factors in groundwater management.

Transparency & Impact

Open Science & FAIR Data

We are committed to making our research reproducible and accessible. Our lab adheres to the FAIR Principles (Findable, Accessible, Interoperable, Reusable).

💻

Open Source Code

We believe code is a primary research output. Our machine learning pipelines, preprocessing scripts, and hydrological models are published on GitHub under permissive licenses (MIT/Apache).

Visit Lab GitHub

📂

FAIR Data Stewardship

Datasets generated by our projects are hosted on long-term repositories (Zenodo, HydroShare, Google Earth Engine Catalog) with persistent DOIs to ensure citation and longevity.

Metadata Standards (ISO 19115)
Cloud-Native Formats (COG/Zarr)

🔄

Reproducibility

We prioritize "executable papers." Whenever possible, we provide Jupyter Notebooks and Colab environments allowing the community to reproduce our figures and results instantly.

Jupyter Docker Colab

Funding Compliance

Our data management strategies align with the NASA Open Science policy (SPD-41a), USGS Fundamental Science Practices, and NSF Public Access plans.

Request Data Access

Lab Assistant

Enter Gemini API Key

Planetary Water Dynamics

Laboratory

Remote Sensing

Planetary Water

Geospatial AI

Technology

Our Work

Funded Projects

Improving Groundwater Withdrawal Estimation in Arizona

Irrigation Status Mapping using Satellite Embeddings

Digital Twin for the Ganga Basin

OpenET Planning & Development

Groundwater in Transboundary Watersheds

Nevada Water Resources Initiative

Sayantan "Monty" Majumdar, Ph.D.

Education

Ph.D. in Geological Engineering

M.Sc. in Geoinformatics (Cum Laude)

M.Sc. in Computer Science

B.Sc. in Computer Science (Hons.)

Recent Experience

Adjunct Faculty

Assistant Research Professor

Postdoctoral Fellow

Research Scientist Intern

Analytics Modeling Intern

Knowledge Base

Selected Publications

Deep Dive

Paper Highlights

Satellite-based Field-Scale Groundwater Pumping Estimates for Nevada and Oregon

Crop-Specific Groundwater Use in the Mississippi Alluvial Plain

Widespread Loss of Aquifer Storage Capacity

Physics-Constrained ML for Effective Precipitation

Vision Transformers for Global Canopy Height

Assesing Training Data Requirements for ML-based Groundwater Pumping Prediction

Sustainable Groundwater Pathways for India

Transparency & Impact

Open Science & FAIR Data

Open Source Code

FAIR Data Stewardship

Reproducibility

Funding Compliance

News & Recognition

Awards & Honors

Invited Talks

Professional Service

Collaborate with the Lab