IndyMike Posted September 5, 2014 Share Posted September 5, 2014 Hello Michel, My appolgies for the tardy response. I am repairing a number of PC's after a previous storm event (family members). This is good news. I have a new storm front approaching this evening. Will hopefully be able to confirm the updates by monring. As a side note - I may have been a bit harsh in my appraisal of HAM's service. While they have had issues, the version history you provided indicates that they are working to resolve them. I had not seen this previously, and am encouraged by it. IM Link to comment
jmed999 Posted September 5, 2014 Author Share Posted September 5, 2014 Hi IM, another update from HAM: This is good news! With kind regards, Michel I got 1.15" of rain yesterday and no rain today. In my admin console it shows 1.15" of "rain today". Still not fixed here. Link to comment
Michel Kohanim Posted September 6, 2014 Share Posted September 6, 2014 Hi jmed999, Back to HAM ... will report as soon as I hear back. With kind regards, Michel Link to comment
IndyMike Posted September 6, 2014 Share Posted September 6, 2014 Hello Michel, Things have definately changed on my end. Not for the better... KSBN is showing 1.09 inches yesterday and .05 inches since midnight. ISY is showing .07 inches yesterday and .04 inches since midnight. I suppose it's possible that HAM's updates were rolled out after the storm went through yesterday evening. I have no way of telling from this end. Link to comment
Michel Kohanim Posted September 8, 2014 Share Posted September 8, 2014 Hi IM, I am so very sorry. I am in contact with them as we speak. I'll keep you posted. With kind regards, Michel Link to comment
IndyMike Posted September 12, 2014 Share Posted September 12, 2014 The chart below is a comparison of Weatherunderground VS ISY/HAM reported rain for September. I can't even hazard a guess at how HAM is currently rolling up daily rain or when they implemented changes. I would normally try to offer some insight on where the errors might lie in this type of calculation. I have no explanation for the differences I see below. Link to comment
Teken Posted September 12, 2014 Share Posted September 12, 2014 I would normally try to offer some insight on where the errors might lie in this type of calculation. I have no explanation for the differences I see below. Well, its safe to say if IndyMike can't explain it. We are all screwed ! Link to comment
jmed999 Posted September 12, 2014 Author Share Posted September 12, 2014 My "Rain Today" has been all over the place as well, IndyMike and as I said early on in this thread, HAM is working on fixing the wrong issue. It's not a rolling 24 hour issue! Oh well...tired of babysitting the whole thing but I'm confident UD will find a way to get it straight at some point. Link to comment
Michel Kohanim Posted September 12, 2014 Share Posted September 12, 2014 Hi IndyMike, thank you. We are investigating other solutions. It's really not fair to you all to have to deal with this issue on a daily basis. With kind regards, Michel Link to comment
Techman Posted September 12, 2014 Share Posted September 12, 2014 Hi Michel The Inglewood station (East of LAX), MID_E0482, has been showing rain for the last several days. Other Los Angeles stations have incomplete reporting. Seems the only accurate stations are those located on an airport. HamWeather just isn't making the grade. Link to comment
Michel Kohanim Posted September 15, 2014 Share Posted September 15, 2014 Hello everyone, thanks so very much for the feedback. Here's an update from HAM While no code was pushed to production, over the weekend the team made several back end changes: • Updated the archival method of the observations for improved performance and improved speed to availability of latest data. • Split precip into a separate data set. To help improve precip parsing, this was broken into a separate data set for easier handling. • Internal Update to official stations o With METARs, the hourly precip can be added, but if there are multiple updates in an hour, then the last update should be utilized for the hours precip. o This change has been made internally and will be confirmed working and accurate today. If so this will be pushed to production later today or tomorrow. • Today the team is moving to improved processing of CWOP stations. o Starting with stations that report a running 24 hour precip. With such stations, to obtain an accurate precip total, we have finalized the formulas. o Once the above is working an implemented, we will move on to stations that are reporting only a running 1 hour precip. There are a couple methods the team came up with, but the simplest is to sum the last readings per hour. I will provide some additional updates this afternoon on progress. With kind regards, Michel Link to comment
Michel Kohanim Posted September 16, 2014 Share Posted September 16, 2014 Hello all, More updates from HAM: As of last night, updates for the daily summaries for official observations, when NOT using the filter=allstations parameter, were pushed to production. Changes included: • Improved total precip accuracy based on actual observations, for stations that report. • Daily summaries updated more frequently • A new flag, summary.precip.method has been added as an identifier of how the precip was calculated. For official metars this will normally be "sum" where the last observation in each hour will have the 1 hr precip summed. • The system will attempt to determine the best method to calculate the precip totals, though the system will allow setting a specific method on a per station basis if needed, which is useful with personal weather stations. Today, the team will be implementing these same changes to the personal weather stations "filter=allstations". By tomorrow, there should be improved precip totals for: • Official stations • Stations that submit 1 hour precip for the current hour only • Stations that submit a running 24 hour precip only. • Quality Control data will be added to the summaries. This data is included in the observations, but adding to the summaries will allow filtering out personal weather stations that are not passing quality control. Stations that report a running 1 hr precip may take another day or two. With these changes we are also adding support for observations/archive to work with filter=allstations. Specifically, this endpoint allows obtaining all observations from midnight to midnight station local time, that occurred on a specific day in a single request. Perfect for charting or creating custom processing of the observations. We will continue to update as we move forward. Please note that we use filter=allstations and therefore the actual updates for this will be published this evening. With kind regards, Michel Link to comment
IndyMike Posted September 16, 2014 Share Posted September 16, 2014 Thanks Michel, No rain in the forcast until the weekend. Will report back when I have some data. Link to comment
jmed999 Posted September 16, 2014 Author Share Posted September 16, 2014 My "Rain Today" is 0.26" but my pws is reporting 0.06". The extra 0.20" is from yesterday. I have a good bit of rain predicted this week so I'll keep you posted! So far no fix for me though. Link to comment
Michel Kohanim Posted September 17, 2014 Share Posted September 17, 2014 Hi jmed999, The reason was that the first patch was only for national stations and not PWS. The second patch (7:00 PM CDT yesterday + one tomorrow) was/is for PWS stations. Here's an update from HAM: The team has pushed out the updates mentioned earlier as of 7pm CDT. You should notice most weather stations reporting much improved precipitation within the daily summaries. There may continue to be some minor issues to be resolved. Additionally, the team is adding "QC" support for both the summaries and the precip. This update will be in our Thursday API maintenance release. Again, while I believe most stations should be more accurate, let us know of any major discrepancies. With kind regards, Michel Link to comment
jmed999 Posted September 18, 2014 Author Share Posted September 18, 2014 We got a lot of rain yesterday. None today. The "Rain Today" is showing 0.0" of rain. This is a real good sign. Since it rained yesterday I was expecting it to incorrectly show in "Rain Today"...but it didn't. I'll keep you updated but this is good news. Link to comment
IndyMike Posted September 18, 2014 Share Posted September 18, 2014 I'm showing 0.05" rain today. Last recorded rain at KSBN was 0.07" on 9/15 <climate enabled="true" locationId="KSBN" rss="http://wx.hamweather.com/?pands=KSBN" lastUpdatedTS="2014/09/18 16:42:23"> <Temperature>72 F</Temperature> <Temperature_High>72 F</Temperature_High> <Temperature_Low>43 F</Temperature_Low> <Feels_Like>72 F</Feels_Like> <Temperature_Average>52.7 F</Temperature_Average> <Humidity>41 %</Humidity> <Pressure>30.11 psi</Pressure> <Dew_Point>46 F</Dew_Point> <Wind_Speed>7 mph</Wind_Speed> <Wind_Direction>ENE</Wind_Direction> <Gust_Speed>17 mph</Gust_Speed> <Total_Rain_Today>0.05 inches</Total_Rain_Today> <Light>93 %</Light> Link to comment
Michel Kohanim Posted September 18, 2014 Share Posted September 18, 2014 Hi IM, Does this mean that you had 0.05" even though it should have been 0? With kind regards, Michel Link to comment
IndyMike Posted September 19, 2014 Share Posted September 19, 2014 That's correct Michel. I'm thinking that something was "stuck" after the changeover. Rain today was showing 0.05 again this morning. I manually switched stations and then came back to KSBN. Now correctly showing 0". Link to comment
Mark H Posted September 19, 2014 Share Posted September 19, 2014 It appears that hamweather(?) is having issues. My weather station, and my two local airports (KMSP, KSTP) have correct data when checked with other sources (APRS, CWOP, findu.com), but ISY is showing either blank values for certain variables, or values that are way off. And as of about 30 minutes ago, all but three reporting stations near me are gone. Mine isn't showing, nor are the two airports, when I do a Find or a Locate Me. Anyone else having this happen today? EDIT: and naturally, as soon as I posted and checked yet again, the data is correct and all stations are listed again. This was not the case for several hours, but it's correct now. I'll keep an eye on this throughout the day. Link to comment
Michel Kohanim Posted September 19, 2014 Share Posted September 19, 2014 Hello IM/TheSidewinder, Are you both on 4.2.10? With kind regards, Michel Link to comment
Mark H Posted September 20, 2014 Share Posted September 20, 2014 I'm on 4.2.10, yes. (Updated that the day I installed the Z-wave board.) And after I posted about this, above, it has been correct since then, so about 6 hours now. If I had to guess, given the symptoms, I'd suspect ham weather was doing something, perhaps rolling out out an update? In any case, everything is spot on now. And...... it may be my imagination, but it seems the "delay" in data reporting is shorter than it had been, now less than a half-hour rather than 45 minutes. Link to comment
IndyMike Posted September 20, 2014 Share Posted September 20, 2014 Hello Michel, I'm on 4.2.8 (RC3). There was nothing related to HAM/Climate noted in the 4.2.10 (RC4) release notes. Did I miss something? Link to comment
Michel Kohanim Posted September 21, 2014 Share Posted September 21, 2014 Hi IM, How about your UI? We are going to have 4.2.15 tomorrow so would you please try it and let me know what you find? With kind regards, Michel Link to comment
IndyMike Posted September 22, 2014 Share Posted September 22, 2014 Hi IM, How about your UI? We are going to have 4.2.15 tomorrow so would you please try it and let me know what you find? With kind regards, Michel Hello Michel, Verified both firmware and UI at 4.2.8, Under 4.2.8: 1) KSBN received 0.06" of rain yesterday. The ISY reported 0.04" 2) No rain at KSBN after 16:00 yesterday and today. The ISY continued to report 0.04" until 4:50 PM when I loaded 4.2.15. 3) After loading 4.2.15 rain is at 0.0" No rain the the forcast for the forseeable future. Waiting on data on this end... Link to comment
Recommended Posts
Archived
This topic is now archived and is closed to further replies.