Rain shortcut
274 of 565 comparisons
15 Pacific places · 565 year-to-year comparisons · 1980-2024
Wet year.
Bigger coconut harvest?
Try the shortcut: wetter means yield rises; drier means it falls. Make the call before each record is revealed.
This tests one place-level annual shortcut—not whether rainfall matters to coconut farming.
01 · Make the call
What can “wetter” or “drier” tell you on its own?
Eight selected cases let you try the rule before seeing the reported yield direction.
Ready
0 of 8 calls matched
Choose “Make the first call” above.
—
Annual rain category relative to the 1991-2020 reference.
Why these eight cases?
They are deliberately balanced: four hits and four misses, spanning eight places and both wet and dry years. They make the interaction fair; they are not a random sample. The full comparison below carries the evidence.
02 · Change the setup
Move the rain. Raise the bar.
Use rain from the same year, one year earlier or two years earlier; then ignore small yield moves. Four of 12 setups beat the common-direction guess, but none gains five percentage points.
trails the common-direction guess
beats the common-direction guess
On narrow screens, scroll sideways to reach all four yield thresholds.
| Rain used | Any move | Over 1% | Over 5% | Over 10% |
|---|
Same year · any reported yield move
Always predict a yield rise
311 of 565 comparisons
6.5 percentage points behind.
03 · Look place by place
The pooled rate hides 15 different records.
Each line compares the rain shortcut with that place's more common yield direction. Sample sizes run from 16 to 44 comparisons, so small differences should stay small claims.
rain shortcut
common-direction guess
Before drawing a conclusion
A failed shortcut is not a failed climate relationship.
Scale. Annual place-level rainfall is not farm weather, the growing season, rainfall timing or extremes.
Other forces. Technology, crop mix, hazards, pests and reporting can also change reported yield.
Coverage. Record lengths differ and place-years repeat, so the pooled rate is not a Pacific population estimate or a causal test.
Method and sources
Annual precipitation anomaly (RAIN_ANOM, mm, 1991-2020 reference) is joined to reported coconut yield (COCON, CROP_YIELD, kg/ha) by place and year. Only consecutive yield years form comparisons. The headline excludes unchanged yield and rain exactly at reference; no missing value is imputed.
Sources: Pacific Data Hub climate-change indicators SPC:DF_CLIMATE_CHANGE(1.0) and agricultural production SPC:DF_AGRICULTURAL_PRODUCTION(1.0), snapshots frozen 25 August 2026.
Technical limits
The comparison is descriptive. Repeated place-years are not independent, the benchmark is in-sample and both official dataflows may be revised. No model is fitted and no causal or Pacific population inference is made.