WEATHER COMMENTS WB4APR ----------------------------------------------------------------------------- Updated 11 April 2006 This file attempts to clarify many details about APRS weather formats. Curt Mills was very helpful in identifying some of these omissions. On Tue, 17 Jul 2001, Curt Mills, WE7U wrote: >Page 64 of the APRS spec talks about weather report format: > > "s = sustained one-minute wind speed in mph. > p = rainfall (in hundredths of an inch) in the last 24 hours. > P = rainfall (in hundredths of an inch) since midnight > ... > Other parameters that are available on some weather station units include: > L = luminosity (in watts per square meter) 999 and below. > l (lower-case "L") = luminosity (in watts per square meter) 1000 and above. > s = snowfall (in inches) in the last 24 hours. > # = raw rain counter" > > > 1) Why are there two 's' fields defined, wind speed _and_ snow? > Is one a typo and should have been a capital 'S'? Probably snow? In the RECOMMENDED complete WX format, the WIND SPEED does not have a "s" format identifier since it is in a FIXED position. Thus if you see "s", then it is snow. Only WinAPRS uses the verbose positionless Weather format which does have this ambiguity. Thus, it cannot be used to send Snow values. But the "positionless" weather format is not recommended for further implementation. > 2) Rainfall in last 24 hours. I assume this is intended to be a > rolling count going backwards from the current time 24 hours, and not > the amount of rain that was seen yesterday. "last 24 hours" implies > a rolling count to me. I'm fairly sure on this one, just checking. Yes. I think only the original APRSdos and Xastir do this properly. Newer programs took the simpler approach of starting the count each day at midnight... But this is why we further defined three different rain identifiers. "p" is as above, "P" is since midnight, "r" is last hour. > 3) The "Other parameters" _letters_ are defined, but the number of > bytes for each field are not. The are all three bytes unless otherwise specified. > 'l' is in the spec also under the "Other parameters" heading, > supposed to be luminosity greater than 1000. Not knowing the > range for luminosity, do I assume that a 1000 gets added to a > three digit value to give the final > result for the 'l' field? No. if the byte is "L" then it is a luminosity from 0 to 999. If the luminosity is above 1000, then the lower case "l" is used. Thus l123 is a luminosity value of 1123. Kinda cute isnt it.. > Snow 's' has ?? bytes? Format? (s1.25, s100.99) Three bytes. S001 is one inch. S1.5 is 1.5 inches. S010 is 10 inches. Notice that fractions of an inch above 9 inches are not possible... > What is the "raw rain counter"? Is that the "total rain" counter > that some weather stations provide, or the raw counts from the > rain bucket over some time period? How many bytes for that field > and in what format? 1/100 inch? Inches? Counts? It is the three least significant bytes of the counter no matter what units. The assumption here is that it is just a raw bucket count and the users of that station have to understand the specifics of that counter. Mostly it lets users see if the value has CHANGED from the last reading. This is not recommended. These days it is better for all raw WX stations to have a pic processor to put the data into the original "COMPLETE" format. > So, are the \W and /W icons to be used for stations sending Complete > Weather Reports on APRS? If so, the spec text is wrong. No, they are just alternate symbols for a weather station. They were intended for use at National Weather Service sites to distinguish real "calilbrated" weather reports from "amateur" home stations. > I just tweaked Xastir to only allow \_ and /_ for weather stations. > If the NWS icons are intended for use at NWS _as_ weather stations, > then I need to tweak the code some more, and the spec needs to be > corrected as well. Yes, I have always accepted /W and \W interchangeably with /_ and \_. thanks for finding the bug in the spec... > How do you differentiate "No X" from "No sensor for X". > Is your windspeed 0 because it really is, or because you don't have a sensor? The spec allows "..." or " " for the fields. It also allows any field to be absent, as long as it's not one of the few required fields at the beginning. I prefer "r..." so that the columns line up nicely even if a station has no rain gage. The number of '.' or ' ' characters matches the field width. > Note that the spec does not fully specify the fields for the > additional weather fields possible (details from Bob in an earlier > message to the aprsspec list)... Yes, they are variable. Also, two new software type identifiers have been assigned: "O" for Opentracker and "K" for Kenwood. de WB4APR, Bob