FIELD REPORTS · SEP 3
Ghost Rain, Then the Real Thing.
Key points
- Wednesday afternoon the radar showed a broad shield of rain over central New York, sliding east toward Connecticut. It looked certain. Every cell died on approach: the airport gauges at Oxford and Bridgeport recorded no measurable rain from morning through the whole afternoon.
- This site got one line wrong and fixed it the same day: our Tonight summary said only "Overcast" while our own hourly numbers carried a 45 to 51 percent rain signal for 2 to 5 a.m. The summary let five cloudy evening hours outvote five rainy ones. It cannot do that anymore.
- The real rain kept its appointment. The Weather Service said after midnight at 80 percent, and the Oxford airport gauge went from light rain to heavy rain at 1:25 a.m., with Hartford following at about 1:55 a.m.
The first field report files when the sky makes news near home, and Wednesday the sky spent a whole afternoon making news it never delivered, then delivered the real thing while everyone slept. Every claim below was checked against the instruments twice: once against the radar we render ourselves, once against the ground-truth gauges the Weather Service operates at named airports.
The afternoon of ghosts
By lunchtime Wednesday the radar picture was the kind that makes a person cancel plans. A wide shield of returns covered central New York from Syracuse down through Binghamton and Ithaca, with heavy cores in it, and the whole mass was working east. Anyone watching that screen from central Connecticut concluded rain was coming. One of us said so out loud.
The forecasts disagreed with the screen. The Weather Service afternoon forecast for the area held the chance of rain at 4 percent, and this site's forecast read dry through the evening. That is an uncomfortable position, standing between a radar full of echoes and a forecast full of nothing, and the honest way through it is to wait for the gauges.
The gauges sided with the forecasts. The airport station at Oxford, which sits between Southington and the lower Naugatuck Valley, recorded no measurable precipitation in any observation from 9:55 a.m. through the afternoon. Bridgeport's airport, down on the Sound, recorded the same: nothing. The cells that looked so certain at noon dissipated as they crossed the state line. The radar was not lying about what was aloft. What was aloft simply never reached the ground here.
The line this site got wrong
Field reports check facts twice, and that includes ours. At midday Wednesday this site's Tonight summary read "Overcast. Lows around 60." At the same moment, our own hourly forecast carried rain codes and a 45 to 51 percent probability for the hours between 2 and 5 a.m. Both numbers lived on the same page. The summary was blind to its own data.
The cause was a counting rule. The night summary described the most common condition across the night's hours, and five cloudy evening hours outvoted five rainy ones after midnight. Rain lost a popularity contest to clouds. That rule was rewritten and shipped the same afternoon: when rain occupies a real share of the night, the night is about the rain, labeled by its strongest hour, with the start time and the percent stated. By Wednesday evening the line read "Rain after midnight, 51 percent." The receipts culture here is not a slogan; when the mistake is ours, it goes in the report.
The real thing, on schedule
The Weather Service overnight forecast said showers and thunderstorms after midnight, 80 percent. At 1:20 a.m. the Oxford gauge reported light rain. By 1:25 a.m. it reported heavy rain, and its 1:37 a.m. observation measured about a fifth of an inch in the hour. Hartford followed the same script half an hour behind: light rain at 1:40 a.m., heavy rain with mist by 1:55. More light rain brushed Hartford again near 7:40 in the morning.
So the after-midnight call verified nearly to the minute. Our corrected line said 51 percent where the Weather Service said 80, and the rain came, which means their number wore it better than ours. That gap, and the question of what our blend should learn from a radar full of dying cells, is now sitting with the forecast doctors, whose job is exactly this kind of homework. Their findings land where they always do, in the Receipts.
The day, by the numbers
| Item | Value |
|---|---|
| Afternoon rain chance, forecast | 4 percent (NWS), dry (this site) |
| Afternoon rain, measured | None at Oxford (KOXC) or Bridgeport (KBDR), 9:55 a.m. onward |
| Overnight forecast | 80 percent after midnight (NWS); "Rain after midnight, 51 percent" (this site, corrected midday) |
| First measurable rain | Oxford, light rain 1:20 a.m., heavy rain 1:25 a.m. September 3 |
| Hartford (KHFD) | Light rain 1:40 a.m., heavy rain about 1:55 a.m., light rain again near 7:40 a.m. |
Sources, as they read: National Weather Service station observations for KOXC, KHFD, and KBDR via api.weather.gov; the NWS Boston forecast for the area issued Wednesday; this site's own published hourly forecast and its correction, shipped September 2. The gauges are the referee. They usually are.
Field Reports file when the sky makes news near home. The forecast grades itself nightly in the Receipts, and the daily national picture is the Weather Beat.