Skip to content
SIH Buddyby Ganeev Singh
Dev
All problem statements
SIH26074Proceed with cautionacceptance 2/5

Downscaling of weather forecast from Block level to Panchayat level: Inferring high-resolution plots/ data/ information from low-resolution plot /data /information /variables for agro-meteorological advisory services.

Ministry of Earth Sciences (MoES) · Disaster Management · Software

You can certainly produce a higher-resolution forecast field, but with no village-level observations to verify against you cannot show it is more accurate than the coarse one — and adding detail without adding skill is exactly the criticism this approach attracts.

What it actually is

Farm advisories are issued at block level, covering areas large enough that the advice does not fit any particular village. The ask is to take those coarse forecasts and produce them at panchayat level instead, so the advice a farmer receives actually reflects conditions where their field is.

What to build

A statistical downscaling pipeline mapping coarse gridded forecast fields to a finer panchayat-scale grid using predictors that actually carry sub-grid information — terrain elevation and aspect, land cover, distance to water bodies and observed local station history — trained against whatever fine-scale observations can be assembled, producing downscaled temperature and rainfall fields with an explicit uncertainty estimate at each point, and crucially a verification component testing whether the downscaled field has more skill than simply applying the coarse block value everywhere, because added resolution without added skill is the failure mode this whole approach invites.

Smallest thing that wins the room

Hold back a set of village-level station observations, downscale to those locations, and show your error against the observations alongside the error you would have had by just using the coarse block forecast unchanged.

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.

Moderate140–330 teams expectedroughly 1 in 119–275 wins it

Quieter than 43% of the 226 · #129 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.

What you will be writing

  • quantile mapping and regression-based statistical downscaling
  • terrain elevation, slope and aspect as predictors
  • land cover and NDVI covariates
  • gridded IMD rainfall and temperature inputs
  • leave-station-out verification against coarse baseline
  • per-point uncertainty quantification
  • Agrometeorology
  • Statistical downscaling
  • Forecast verification

Prior art to read before you start

spatial downscaling of coarse forecast fields · terrain-informed local weather estimation · skill verification against coarse baseline

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.