AI/ML-Based Integrated heavy rainfall Early Warning and Inundation Prediction System using Satellite, Radar, observational Weather and numerical weather prediction model data.
Ministry of Earth Sciences (MoES) · Disaster Management · Software
Two research problems in one sentence, and the inundation half needs metre-scale terrain and drainage data that simply is not public in India — if you take it, pick one catchment where you can get real elevation data and hindcast a single documented event honestly.
What it actually is
Warning that heavy rain is coming is only half of what a city needs; what people actually want to know is which streets will go under water. The ask joins the two together — forecast the rainfall from satellite, radar and model data, then translate that rainfall into where the flooding will actually be.
What to build
Two chained components: a heavy rainfall warning stage blending satellite precipitation estimates, radar reflectivity where available and numerical model output into a probabilistic rainfall forecast over a defined area and lead time, calibrated against gauge observations; and an inundation stage translating that forecast rainfall into flood extent over a terrain model with a drainage network and infiltration assumptions, producing a map of which areas flood to what depth and when — the whole chain demonstrated by hindcasting a real past flood event and comparing your predicted extent against the flooding that was actually observed.
Smallest thing that wins the room
Hindcast a well-documented past urban flood: feed in the rainfall that actually fell, produce the inundation map, and overlay it against satellite-observed flood extent from that day.
How crowded this one gets
A guess, projected from the 2025 statements — the last year where both the submission counts and the winners were published.
Quieter than 44% of the 226 · #126 of 226 by expected field
A normal-sized field. Your idea has to be good, not miraculous.
Why: central ministry statements sat below the average.
This is a guess, not a fact
Nobody has published 2026’s numbers yet. This is an analysed estimate from last year’s pattern, so please do not take it as the truth — check the live counter on the SIH portal before you decide anything. The range covers the middle half of likely outcomes, so one statement in two lands outside it. Entry closes at 500 ideas per statement, so no range goes past that — a statement that reaches the cap fills and shuts rather than drawing an unlimited crowd. The model reads only three things a team can see before choosing — software or hardware, the theme, and what kind of body posted it — and those explain about a quarter of the variation in last year’s field sizes (R² 0.25 on held-out statements). Trust the band more than the number, and the ordering more than either. It cannot see how good your idea is, which is the part that actually decides it.
The scores
The number is the shorthand. The line under it is the reason.
Acceptance potential
2/5The inundation half is the differentiator and it rests on high-resolution terrain and drainage data that is not publicly available in India, so most submissions will run a flood model over coarse elevation data and produce maps that look authoritative while being unable to resolve the street-level question the system exists to answer.
Feasibility
3/5The rainfall half is well supported with free satellite precipitation products and gridded gauge data, but the inundation half needs terrain at metre-scale resolution together with drainage and storm-sewer network data, and neither exists publicly for Indian cities — the coarse elevation models that are free cannot resolve which street floods.
Innovation scope
4/5The description repeats the title, so no method, region, lead time, resolution or model type is prescribed anywhere and the entire approach is yours to define.
Clarity
1/5The description is the title reproduced verbatim, giving no background, no requirement, no lead time, no spatial scale and no definition of what a correct inundation prediction would be.
Effort
MassiveMulti-source rainfall blending and calibration is a full project, hydrological and hydraulic inundation modelling is another entirely different discipline, and chaining them with honest uncertainty propagation is a third — this is two research problems joined by a sentence.
Demo-ability
MediumA flood extent map over a city is compelling and hindcasting against a documented event gives you real evidence, but the result is only as credible as the terrain model beneath it and a judge who knows flood modelling will ask about resolution first.
In its favour
- Green flag: Satellite radar imagery of past floods gives you genuinely observed inundation extent for validation, which is a rare opportunity to check a flood model against reality rather than against another model
- Green flag: Free global satellite precipitation products and gridded gauge data make the rainfall half entirely tractable and calibratable
- Green flag: Because the description is only the title, you can scope to a single well-documented catchment and a single past event and be fully responsive while remaining achievable
- Green flag: Established open hydraulic modelling engines exist, so you can build on validated flood physics rather than inventing a routing scheme
Against it
- Red flag: Urban inundation depends on terrain at metre resolution plus the storm drain network, and neither is publicly available for Indian cities — the freely available elevation data is far too coarse to say which road floods, which is the entire question
- Red flag: Two distinct disciplines are joined in one sentence, and a team that splits its time between rainfall forecasting and hydraulic modelling will produce a weak version of each
- Red flag: Errors compound through the chain: a modest rainfall forecast error becomes a large inundation error, so propagating uncertainty rather than presenting a single crisp flood map is both more honest and more defensible
- Red flag: A confident flood extent map is exactly the kind of output that gets used for decisions, so overstating the resolution of your terrain model is a real hazard rather than a presentational flaw
What you will be writing
- GPM IMERG satellite precipitation blending
- gridded gauge calibration of forecast rainfall
- LISFLOOD-FP or HEC-RAS 2D hydraulic modelling
- SRTM / Cartosat DEM hydrological conditioning
- Sentinel-1 SAR flood extent for validation
- probabilistic rainfall threshold warning logic
- Hydrometeorology
- Flood modelling
- Disaster early warning
Prior art to read before you start
heavy rainfall forecasting and warning · rainfall to inundation extent modelling · hindcast validation against observed flood extent
Analysed by Claude Opus. Every score above is a judgment call with its reasoning attached — kindly cross-check this against the official statement on the SIH portal before your team commits to it.