HydroTelliSense

Reading the water cycle insights from orbit.

HydroTelliSense team combines satellite remote sensing, physics-based hydrologic modeling, and machine intelligence to monitor, predict, and provides insights on water resources and early warning of hydroclimatic hazards.

A research group focused on intelligent decision making from satellite observations.

What we work on

We build tools that turn satellite observations into decisions — coupling land surface and hydrologic models with new data streams like SMAP and NISAR soil moisture, and applying deep learning to detect and forecast hazards before they escalate. Our work spans landslide early warning, flood forecasting, and the study of human-driven hydrological change in transboundary river basins including the Lower Mekong and the Ganges-Brahmaputra.

  • Landslide hazard assessment01
  • Flood forecasting02
  • Transboundary basin change03

News

Short, dated updates — talks, awards, new papers, field campaigns.

[Mon YYYY]

[Headline — e.g. "Paper accepted at ..."]

[One-sentence detail]
[Mon YYYY]

[Headline]

[One-sentence detail]

Latest innovation

ROSES-2025 ECIP-ES proposal

Coupling LIS–LHASA with SMAP & NISAR soil moisture

We're replacing LHASA's precipitation-based proxy for soil saturation with physically based, observationally constrained soil moisture — assimilating SMAP and NISAR retrievals into NASA's Land Information System via an Ensemble Kalman Filter, and adding 2D lateral groundwater flow. The goal: sharper landslide prediction over the Central and Southern Appalachians, evaluated against probability of detection, false alarm rate, and ROC-AUC.

LIS land surface model SMAP + NISAR soil moisture EnKF data assimilation LHASA + groundwater flow landslide hazard output

Team

A small group spanning hydrology, remote sensing, and machine learning.

Nishan Biswas

Team Lead · Associate Research Scientist

NASA Goddard Hydrological Sciences Laboratory, UMBC / GESTAR II. Works on landslide hazard assessment, flood forecasting, and transboundary hydrological change; mentors early-career and student researchers.

[Name]

[Role]

[Bio]

Portfolio

Prior and ongoing work across hazard types and regions.

Landslides · Colombia

Satellite-based landslide mapping, Cauca Valley

Multi-temporal mapping using U-Net deep learning and NASA's Prithvi geospatial foundation model, benchmarked against the SALaD (Semi-Automatic Landslide Detection) OBIA/Random Forest system, with temporal compositing to handle cloud cover.

Landslides · Appalachians

Landslide hazard assessment

Advanced hydrologic modeling coupling LIS and LHASA with soil moisture data assimilation.

Floods · Lower Mekong

Transboundary flood forecasting

Studying human-driven hydrological change and flood dynamics across the Lower Mekong basin.

Hydrology · Ganges-Brahmaputra & Latin America

Human-driven hydrological change in transboundary basins

Investigating how upstream infrastructure and land-use change reshape downstream water availability and risk across major transboundary river systems.

Publications

Peer-reviewed papers, proposals, and conference contributions.

2026

Multi-temporal satellite-based landslide mapping in Colombia's Cauca Valley

7th World Landslide Forum, Session 4.8 — extended abstract
N. Biswas, P. Amatya, T. A. Stanley, C. Chiesa, D. Morath
2025

Improving Landslide Hazard Assessment with Advanced Hydrologic Modeling

NASA ROSES-2025 A.11 — Early Career Investigator Program proposal
N. Biswas (PI)
[Year]

[Title]

[Venue / status]
[Authors]

Collaborators

Institutions

NASA Goddard Hydrological Sciences Laboratory UMBC GESTAR II SERVIR-Mekong Asian Disaster Preparedness Center Pacific Disaster Center

Researchers

Md. Rezaul Haider Protik bose Pranto Smita Sharma Maruf Ahmed Suchi Nandi Puja

Resources

Code, datasets, and tools this group has released.

[Code]

[Repository name]

[What it does — link to GitHub.]

[Dataset]

[Dataset name]

[What it contains, where it's hosted.]