When to report
- View
Occurrence time against reporting time, why a warning already being out is a reason to report rather than not to, and the tag ladder your report can push a storm up.
In this lesson you will
- Contrast real-time and delayed reports, and state the value of a late report.
- Distinguish the time of occurrence from the time of reporting, and encode time correctly.
- Explain why to report even when a warning is already in effect.
- Decide when to send an update and how to avoid duplicate reports.
As soon as it is safe, which is not the same as immediately
The office asks for it plainly: "please submit your report as soon as possible after a severe weather event", and "Always include the time the event happened and the location." Note the word after. Nothing in that sentence asks you to report during the event from an exposed position, and module 8 has a great deal more to say about that.
Occurrence time, not call time
This is the single most common defect in otherwise good reports. The report carries the time the thing happened. If hail fell at 4:52 and you called at 5:10 from somewhere with signal, the report is a 4:52 report.
The office reconciles your report against the radar timeline. A report stamped 5:10 gets checked against the 5:10 scan, which shows the storm somewhere else entirely, and what should have been a confirmation becomes an anomaly. Get the occurrence time right and your report corroborates. Get it wrong and it looks like a bad report.
For the archive, Storm Data has its own convention, and it is worth knowing because it trips people up: "Use local standard time in 24-hour clock throughout the year." Standard time, all year, including through the summer. That is an archive convention rather than something you need to do the conversion for.
| You are giving | Do this |
|---|---|
| A time to the office, live | Local clock time, and say which zone. "6:21 Central" removes all doubt |
| A time in writing | Add CDT or CST explicitly. A bare "6:21" in June is ambiguous |
| A time near midnight | Give the date as well. This is where the worst errors live |
| A time in UTC | Only if you are certain. A wrong Z conversion is worse than a correct local time |
The office can convert. It cannot recover a time you gave wrong, and a five-hour error near midnight moves the report to the wrong day.
There is already a warning out
The most common reason people do not report is that the warning is already issued and they assume they are too late to matter. Three separate things say otherwise.
- Verification. Whether the warning was right is decided by reports. A warned event with no report looks, in the record, like a false alarm.
- Extension and upgrade. Warnings get extended in time and space, and a severe thunderstorm warning becomes a tornado warning, on the strength of reports.
- Tags. A warning carries a damage threat tag, and your report can move it up the ladder. That is not a paperwork change; the top tag sets off phones.
The values are the NWS's own. Base is the unchanged severe criterion, 1.00 inch hail and/or 58 mph, and carries no tag and no wireless alert. Considerable is at least 1.75 inch golf ball hail and/or 70 mph, and also does not fire a WEA. Destructive is at least 2.75 inch baseball hail and/or 80 mph, and a warning carrying it automatically activates a Wireless Emergency Alert on phones inside the warned area.
The gap between a golf ball and a baseball is the gap between a warning people may not see and a warning that wakes them up. If the stones are large, that is the moment to get a ruler and a photograph rather than to estimate, because that report may be what moves the storm up a rung.
Late reports
A report made days later cannot help a warning that has already expired. It can still do two things, and they are not small: it feeds verification, and it goes into Storm Data, which is the record that still exists in thirty years and the one every long-term statistic is built from.
This matters most for exactly the events lesson 1 said were missing: the rural ones, the night-time ones, the ones nobody was awake for. If you found damage this morning that happened last night, report it. Say when you found it and when you believe it happened, and say that the second one is an estimate.
Updates, and not repeating yourself
- Send an update when
- Something material changed: the hail got bigger, the tornado formed or lifted, the water is now over the road rather than at the curb, the damage turned out to be worse in daylight. Say explicitly that it is an update to an earlier report.
- Do not send when
- Nothing has changed and you are anxious, or somebody else on the net has already reported the same event at the same place. Duplicate reports of one event cost the office time during the exact minutes it has least of it, and can inflate the record.
Check yourself
Golf ball hail fell at your house at 21:40. You lost power, and report it at 22:15 when your phone reconnects. What time goes in the report?
- 21:40, stated explicitly as the time it happened
- 22:15, because that is when the report was made
- Both, with 22:15 as the primary time
- Neither. Once 35 minutes have passed the report is too late to be useful
Hint
The office will check this against a radar scan. Which of these two times has a storm in the right place?
Answer
21:40, stated explicitly as the time it happened
The report carries the occurrence time. At 22:15 the storm is thirty miles away, so a report stamped with the call time is checked against the wrong scan and looks implausible. And a 35-minute-old report is entirely useful: golf ball hail is the considerable damage threat tier, and the warning may still be running.
Sources
- NWS Mobile/Pensacola, "Submit a Storm Report"
- NWS Instruction 10-1605, Storm Data preparation (current edition 26 July 2021)
- NWS, damage threat categories for severe thunderstorm warnings
- NWS Weather Spotter's Field Guide
- NWS, wireless emergency alerts