GOMD / PERSISTED WORLDWIDE DATA

Worldwide rainfall database ingestion

Native CHIRPS, IMERG, ERA5 and three-hour forecast cell values are persisted in the GOMD database. User maps, location series and CSV files read only the database. Only the background ingestion job contacts providers.

Actual storage and progress

Reading database status…

—Valid cell observations in DB
—Total stored native cell slots
—Periods with every cell imported
—Compressed numerical DB payload

An observation is one valid cell/period value physically stored in the database. Lossless compression preserves native resolution and numerical precision. NoData cells and source data not yet imported are excluded from valid observation counts.

Ingestion status of the four sources

Product / coverageDB coverageValid observationsPeriods / tilesJob queueOpen

Provider calendars and imported database periods are shown separately. Only persisted cells in a partial period can be queried; other cells return “not imported yet”. A complete period means every original native cell tile has been verified and stored.

Capacity and source gaps

Current hosting: 30 GB total disk, approximately 28.4 GB database quota. Ingestion safely pauses before total database storage reaches 23 GB; records are preserved. Every new global forecast run grows the archive. Full history and ongoing forecast retention may require more capacity.

CHIRPS: land from 60°S to 60°N; ocean and source gaps are retained. IMERG: the USF monthly distribution is separate from the earlier Late V07 selection. ERA5: mirror calendar gaps are tracked separately in the queue; unverified UTC wet-day counts are not fabricated. ICON: original triangular model cells, with three-hour totals derived by differencing consecutive accumulations from the same model run.