In Anthony’s recent post about the new snowfall records, the record high at Jal, New Mexico caught his eye. I live in New Hampshire, so the lonely red dot in Maine caught my eye. I’ve never been fond of this data, and several of the records I’ve looked into are from sites that are relatively new or places I’ve never heard of, and there’s little to justify faith in the data. For example, that Jal NM record, 96°F, may simply be 69°F transposed.
The Maine dot turned out to be in a town of 2100 people or so, Corinna. Never heard of it, but they probably never heard of the even smaller town I live in, so that’s okay. The record they set was 49°F on 2/16. That day the NWS station in Bangor reported 31°F and a private weather station in Corinna just a mile to the east that seems pretty reasonable reported 33°F. Clearly that 49 is wrong. Corinna is an important enough site so the NWS has older data for it, and several other data look just as bad, like the 59°F on 2/1. (Bangor 29°F, Wunderground 24°F.)
[Update: I’m always afraid when dealing with data that doesn’t make sense that I may have missed something. Indeed I did. I’ll have a further update tonight, in the meantime see my comment below for a very important point.]
Okay, small station in the woods, but important enough for the NWS to track it. That helps. So, exactly where is the station? Hey, here’s something with good resolution which leads me to:
Google shows an arrow close to a white smudge that could be a Stevenson Screen that’s too close to the building to the north. It’s also too close to lagoons surrounded by trees. Trees are okay, but lagoons are red flags. No nearby river for draining treated sewage, but a quick check with Google yields a page with an aerial photo of the same lagoons. It is a waste water treatment plant that doesn’t dump into the river, they spray the output on the trees! That’s okay, there has even been research looking into spraying untreated sewage on to plants. What does that spraying do to the temperature readings? That may not be okay, but I need to see if they do that in the winter.
I was curious about the station history, and from Anthony’s link for Jal, I found the official metadata for Corinna:
I had seen the “USHCN” tag on some other pages. This struck me as odd, but with data going back to 1948, I began to worry that someone might think this is a good rural site to add to the USHCN.
I found a user friendly USHCN interface which listed Corinna, then found a FTP list of USHCN V2.5 stations which clinched it. Anthony confirmed it’s not in the V1 or V2 lists, so Corinna is a new addition.
Wow, just wow. Excuse me for a moment.
Why the heck is this station in the USHCN? Hasn’t the NCDC learned anything since the start of the Surfacestations.org project? Yeah, I know the station has a long record (likely what attracted them in the first place) but why oh why didn’t anyone look at the siting before promoting a station with what looks like an awful, awful recent record to a position with global influence?
Thanks, I feel better now. There could still be a significant problem with an equipment malfunction, insert rant about using broken equipment in the official (or preliminary) record here.
Now that we’re talking about a world class station, some things I had brushed away deserve more attention. The BEST folks have data for Corinna going back to 1870, but they also seem to be conflating Orono Maine with Corinna. We would write a whole new post about that page. Suffice it to say they think the temperature data has gone down since the last move in 2005, and after adjusting for that they come up with a steeper temperature climb on a graph with the Y scale stretched by a factor of 3. Why did they do that to the Y axis?
At the very least, the BEST data reminds me that all this comes from one obviously bogus high temperature record. Going back to that, there are many questions to look into:
- Is that record a fluke, e.g. a data entry error?
Probably not, is there are several others, at least in Feb 2013. There are a lot of other months to check!
- Could false warming due to the winter effluent offset false cooling from summertime effluent evaporation?
Perhaps, it might explain why the BEST data may not show the problem.
- And what about the transpirational cooling in the summer from all those trees?
- And all the stations BEST uses as the standard.
We should vet those.
- What other surprises are in the USHCN list?
I don’t have time to do surfacestations.org V2.5….
- The MMS database says the last site change was in 1991, but the wastewater plant didn’t open until 2005.
I might know who to talk to about that.
- What impact does Corinna have on the global temperature?
Probably very little. What about the other new stations? An error here, an error there and pretty soon you have a worthless database.
- What has the NCDC learned from other ill-sited stations and the Surfacestations project?
Well, umm, nothing comes to mind.
I called out the honey dipper last winter to clean my septic system. It was easy to find the septic tank because it was the place in the yard where you could sink a shovel in the ground. The rest of the yard was solid due to ice.
For purposes of investigating climate it would be much better if ground temperatures were recorded in a standardized way.
To make matters worse, Ric, the metadata shows the site moved to the waste treatment plant in 1986. Before that it was 3 miles away, and at a higher altitude.
https://mi3.ncdc.noaa.gov/mi3qry/map.cfm?fid=10226&stnId=10226
https://mi3.ncdc.noaa.gov/mi3qry/search.cfm
[Reply: Oh, I never got around to clicking on other things on that page, thanks. Some of the changes listed on the locations page are siting refinements (thank you GPS). One thing that’s unclear is that the whole treatment plant moved from near the river east (note the longitude, ignore the latitude) to the lagoon/spray site in 2005. -Ric]
Find out what year the treatment plant was built and discard the data henceforth. There are/were (if memory serves me correctly) USDA (Potato) experiment stations in the area and they must have been collecting data as well. The USDA has a forest research station in somewhat nearby.
http://nrs.fs.fed.us/ef/locations/me/penobscot/
They may have collected data as well.
john from DB
@DesertYote – They wouldn’t necessarily be aerated – they look more like final settling ponds, or emergency retention ponds…
And I can confirm, having worked in WW treatment for 8 years: in the winter, if you were out and felt a little cold, you went near or out on ponds like these- there’s more than enough heat there to keep you warm.
While I didn’t spray the wastewater on the trees, I would imagine that it would release all that heat quite quickly, and raise the temp (and humidity) of the area by quite a significant amount. If the station records humidity, you should check for that as well.
not awful far from me here in Etna, 15 miles by air or so. lot longer driving though.
I was 28-30 or so here iirc, local weather stations ran around 30.
be very leery of the Bangor (KBGR) one, its often 3-4 deg F warmer especially in summer.
KMECORIN3 in Corinna reported around 33 for a high that day.
“You basically cant win.”
Perhaps that’s because you are trying to use devices that are not designed for what you want to use them for.
Andrew
I just received a reply to my Email to the NWS in Caribou – from a Climatologist with the Northeast Regional Climate Center in Ithaca NY. I’m always afraid when I write a post that I’ve missed something important, especially when I can’t explain the data like this.
On the major issue, he wrote:
I briefly thought of the TOBS issue, but I was assuming they would have gone to a MMTS system. I should have rethought that when I later saw the arrow point to a white smudge and possible Stevenson screen. It’s tempting, but I won’t have time to try to plot the differences between a 0700 report time for this site and the nearby Wunderground site.
Did anyone else notice the black lagoon liners in the picture ?
http://www.lagoonsonline.com/images/corinna-lagoon-system-aerial-2005.jpg
Regarding the reply from the Northeast Regional Climate Center on time of day observations, what’s the impact on homogenization? This seems to happily mix station data willy-nilly, and if the TOD observations are this far off, then this is another source of error. Maybe a significant source of error.
Careful about using Wunderground data.
I went to Mexico City in Jan 2012, and Wunderground said it would be 32 degrees c for the whole week. Great, so I did not take a jacket. Got there, and it was 15 degrees c for the whole week.
Cost me a new jacket.
Wunderground were kind enough to acknowledge their error, without saying why it happened. But my guess is that they had not subtracted the city’s altitude from their general forecast data.
.
Ric Werme says:
February 25, 2013 at 6:21 am
Sooo, Mr. Climatologist, why do the stations have different reporting times? Please answer that question and I will give you the next in a long list of questions.
The larger question that I want answered is “What are NWS’ criteria for treating two station readings as comparable?” We know that NWS ignores altitude, humidity, and all features of the local environment. That is shameful enough. Are you telling me now that you also ignore time of day? (You can take readings at different times but you must explain why.)
No other science or near-science is anywhere near this sloppy.
it WAS in the 40’s here on 2-15
TOD question aside, why is a temperature guage set withing a waste treatment envelope? How could that location NOT affect temp readings?
Ric – Pics on their way to the NH email.
Ripogenus Dam weather data would probably be much better.
http://www.climate-charts.com/USA-Stations/ME/ME177174.php
George wrote above in http://wattsupwiththat.com/2013/02/24/nws-station-at-corinna-maine-same-old-problems-now-with-global-impact/#comment-1231936 about a weather station he stumbled across in Farmington ME. He sent me some photos today, this show the equipment in use:
One of the challenges of the Surface stations project was to find the weather station. Most of the records had coordinates accurate to about 600′ and directions like “2.5 miles NNW from the post office. This station wasn’t reviewed, which is a pity as anyone driving into Farmington on state route 4/27 will go right by this station and owner’s sign for the station. You can’t miss it.
While George is horrified at the siting, especially with a busy road on one side, and another road this side of the house, the Surface stations project has found a lot worse. The trees are too close too, but Maine and New Hampshire are the most forested states by percentage of forested land. We can’t get away from trees!
I hope when they’re spraying they keep it away from the fan. Oops too late.
Steve T
Steven Mosher said “Leave the bad raw data in and people bitch. detect it and throw it out, people bitch.” Yup, everybody likes a different flavor of fish stew. But it’s still fish stew and not filet mignon.