Saurav Bhattarai Hydrologist
Software

Open tools, free to run.

Two projects carry the dissertation: the GeoClimate Intelligence Platform opens Earth data to people who do not program, and MRRpy turns those data into floods. Both are free, documented, and already used in teaching. The code behind my published papers is public too.

Python package · lead developer

MRRpy

The Multi-Mechanism Runoff and Routing model: rainfall–runoff and flood routing for any basin on Earth, from a DEM and a storm. It traces the drainage network of a DEM, works out how much rain runs off, routes the water through every grid cell and sub-grid river channel, and reports the outlet hydrograph, flood maps, and a water-balance check.

Animation of water depth and river discharge spreading over the Kathmandu Valley terrain during one storm, with the outlet hydrograph
Figure 1One storm over the whole Kathmandu Valley DEM: water depth; river discharge, drawn wider where it is larger; hydrograph.

Install

pip install MRRpy            # core (CPU)
pip install "MRRpy[gpu]"     # + CuPy/CUDA acceleration
pip install "MRRpy[gee]"     # + Google Earth Engine forcing

Run, given only an outlet

from MRRpy import Config, run_pipeline

cfg = Config(DEM_PATH="dem.tif", OUTPUT_DIR="results/",
             OUTPUT_POINT=(27.632, 85.293),   # (lat, lon) of the outlet
             TARGET_CRS_EPSG="EPSG:32645")    # a projection in metres
run_pipeline(cfg)                             # results/hydrograph.csv
MRRpy wizard            # answer a few questions, get run.yaml
MRRpy run -c run.yaml   # process the DEM and route the storm
Documentation GitHub PyPI Interactive course

What it offers

  • DEM to watershed. Reproject, pit-fill, D8 flow direction and accumulation, and delineation, or route the whole DEM with no outlet.
  • Runoff generation. Coefficient, raster, SCS curve number, or physical runoff: variable source area saturation excess, Green–Ampt infiltration with recovery, and impervious surfaces as composable mechanisms.
  • Flood routing. Kinematic wave, Muskingum–Cunge, a semi-implicit diffusion wave solved exactly on the drainage tree, and a local-inertial dynamic wave, with sub-grid river channels sized from the 2-year flood and an always-on water-balance check.
  • Satellite forcing. Optional DEM download, IMERG rainfall, curve numbers, SERVES soil deficit, SoilGrids, HiHydroSoil, and land cover from Google Earth Engine; every option also works offline with your own files.
  • Flood maps. Depth and discharge maps over time, animations, and peak-depth GeoTIFFs.
  • CPU or GPU. One code path in NumPy or CuPy, with automatic CPU fallback; 34 times faster on GPU at 1.38 million cells.
  • Interfaces for every taste. Python API, a command line with a question-and-answer wizard, a Jupyter form, and a QGIS plugin.
License

MIT

Release

v0.3.0, October 2026

Tests

1,893 automated tests, Python 3.9 to 3.12

Paper

Drafted for the Journal of Open Source Software

Figure 2Documentation with a user manual and 11 worked examples, each shown as notebook, command line, Python, and QGIS.
Figure 3Eight Bagmati floods at Khokana without event-wise fitting: median Nash–Sutcliffe efficiency 0.81. How it was verified.
Web platform · lead developer

GeoClimate Intelligence Platform

A Streamlit web platform on Google Earth Engine that lets students, engineers, and researchers search, analyze, and export environmental data in a browser: 1,042 datasets and 23 climate indices, with no code. Free for research and teaching.

See it working

Every screenshot here comes from a real run of the platform in October 2026. Click a step, or let it play.

geeclimate.streamlit.app · Hydrology AnalyzerOpen app
Step 1 of 7 · Pick a tool

The nine tools

  1. Tool 1

    GeoData Explorer

    Search and download from the STAC API and Earth Engine catalog, with map preview, flexible areas, and GeoTIFF, NetCDF, or CSV export.

  2. Tool 2

    Climate Intelligence Hub

    ETCCDI climate indices such as TXx, RX1day, and CDD, with Mann–Kendall trend tests, time series, and maps.

  3. Tool 3

    Hydrology Analyzer

    Precipitation analysis with return periods, frequency analysis, IDF curves, and drought indices.

  4. Tool 4

    Optimal Product Selector

    Compares station data with gridded datasets to find the best source, from daily values to extremes.

  5. Tool 5

    Multi-Geometry Export

    Time series for many areas in one file, such as every county in a state.

  6. Tool 6

    Data Visualizer

    Upload results from any tool and explore time series and rasters with charts and summaries.

  7. Tool 7

    Raster Calculator

    Band math on any Earth Engine dataset: NDVI, NDWI, EVI, or custom expressions, with temporal aggregation.

  8. Tool 8

    GLOBAL-SAM

    Monthly snow depth from MODIS and GLDAS with an ERDC algorithm, compared against reference data.

  9. Tool 9

    SERVES

    Soil moisture, snow water equivalent, and land-surface analysis with the method of Dr. Nawa Raj Pradhan (ERDC).

More tools, real outputs

Climate Intelligence Hub
Climate Intelligence HubA guided seven-step workflow for ETCCDI climate indices, with progress shown at the top.
Climate Intelligence Hub
Climate Intelligence HubIndex time series with linear and Sen’s slope trends: Daymet over Mississippi, 1994–2024.
Climate Intelligence Hub
Climate Intelligence HubA map of any index for any year at the dataset’s native resolution: TXx for 2023.
GeoData Explorer
GeoData ExplorerFive steps from area to download, with keyword search across the Earth Engine catalog.
SERVES
SERVES30 m root-zone soil moisture from Landsat for Mississippi Delta farmland, 15 July 2023.
GLOBAL-SAM
GLOBAL-SAMMonthly snow depth from MODIS and GLDAS for the Nepal Himalaya, February 2023.

Papers and talks

In classrooms and workshops

  • 2025Used by 16 undergraduates in Advanced Water Resources Engineering at Jackson State University
  • 2025AUC Data Science Initiative faculty workshop, "Harnessing GeoClimate Intelligence for Research Innovation" (28 registered faculty)
  • 2025Smartphones4water workshop on crowdsourced and accessible environmental data analysis (lead instructor)
  • 2025Best in Show and first place, Microsoft × AUC Data Science Initiative Data Challenge
Launch the platform Source on GitHub
Apps, courses, and code

Smaller tools and the code behind the papers.

Earth Engine app

Cooling Efficiency Analyzer

Estimates the cooling potential of urban green and blue spaces from Landsat land surface temperature (Urban Climate, 2025).

Open the app Interactive course

Learn hydrology with MRRpy

A free six-module course that follows the model's pipeline in the browser, from terrain to routing.

Start the course Course platform

Advanced Water Resources Engineering

Interactive Streamlit materials for CIV 465: big-data analytics, machine learning, extreme-value analysis, and climate projections.

Open the platform

Code on GitHub

All repositories