Data guide
Daily rainfall data for rainwater tank sizing
Rainfall data is evidence, not decoration. A convenient global dataset is valuable for screening; a well-sited local gauge may be better for a costly property-specific decision. The useful question is what each source can support.
Use these records in the rainwater tank-sizing comparison. If you already have water stored, the rain barrel watering calculator checks its watering budget without assuming new rain.
Why daily records matter
Storage connects one day to the next. A tank fills, spills when capacity is exceeded, supplies demand, and carries the remainder forward. Monthly or annual totals cannot reconstruct that order. Daily data is not perfect—intra-day intensity can still affect gutters and overflow—but it preserves far more of the wet/dry sequence than one annual average.
NASA POWER: consistent global screening
The default planner request uses corrected daily precipitation, PRECTOTCORR, from the NASA POWER Daily API. It requests completed calendar years and rounds coordinates to the source meteorology grid. Nearby coordinates may therefore use the same grid cell. The result panel shows the actual source grid point, period, dataset label, and API version returned.
This consistency makes the source useful where local station data is difficult to obtain or format. It does not turn a grid cell into a roof gauge. Terrain, coastlines, convective storms, wind, gauge undercatch, and model limitations can all create local differences. NASA’s precipitation methodology documents validation against observations and explains why agreement changes with temporal aggregation.
Local gauge or station data
A local series can better represent the site when the gauge is close, exposed appropriately, maintained, quality-controlled, and long enough. “Closer” is not automatically “better”: a short private gauge with missed days may be less useful than a consistent regional source. Record the station location, instrument or provider, quality flags, units, time zone, day boundary, missing-value convention, and any corrections.
Prepare a CSV the planner can audit
The upload accepts comma, semicolon, or tab-separated text with at least 365 daily rows. Use one of these headers:
date,rain_mm 2025-01-01,4.2 2025-01-02,0
Use rain_in instead of rain_mm for inches. Dates must be real ISO YYYY-MM-DD calendar dates, with one observation for every day between the first and last date. Rows may arrive out of order; the planner sorts them before checking continuity. A zero means measured no rain; it is not the same as a blank or missing observation.
The planner rejects missing days, blank rainfall values, duplicate or invalid dates, and values outside its safety bounds. It does not insert zero-rain days or estimate missing observations. If a file has a gap, obtain checked observations from the provider or choose a complete continuous period with at least 365 days. The upload summary shows the period, consecutive row count and number of zero-rain entries; it cannot verify that a plausible observation is correct.
Quality checks before upload
- Sort and uniqueness: confirm one record per date and no duplicates created during spreadsheet imports.
- Continuity: compare the expected calendar sequence with the records. Investigate gaps instead of converting every blank to zero.
- Units: verify millimetres versus inches and whether accumulated readings were already converted.
- Impossible values: inspect negative numbers and extreme totals against source flags and known events.
- Day boundary: know whether a “day” is local midnight, a morning gauge observation, or UTC. Avoid silently combining incompatible series.
- Changes: note station moves, instruments, nearby obstructions, and provider processing changes.
- Coverage: use enough completed years to include varied wet and dry conditions, not only the most convenient recent period.
How long should the period be?
There is no universal number that guarantees representativeness. Longer records expose more sequences but can mix older climate conditions with the present. Shorter records may better reflect current conditions but miss rare dry periods. The planner offers 10–30 completed years for NASA POWER so you can compare windows rather than declare one unquestionably correct.
Historical simulation is not a climate projection. If drought resilience or a long-lived investment matters, include conservative rainfall and higher-demand cases and seek locally relevant climate and engineering guidance.
Compare sources without cherry-picking
Run the same catchment, efficiency, demand, start level, and tank ladder against both sources. Compare average annual rainfall, coverage, longest shortfall, and which capacities reach the plateau. If conclusions differ, do not simply choose the preferred answer. Check periods, gaps, elevation, grid location, gauge quality, and seasonal distribution.
With a checked rainfall series, estimate your roof collection and water demand, then follow the tank sizing workflow while keeping the source and period fixed.