Anomalous DCA anomalies?

[Inverted_Trough]

Interesting article from CWG today about the divergence of DCA temperature anomalies compared to all other temperature anomalies in our region

 

http://www.washingtonpost.com/blogs/capital-weather-gang/wp/2015/08/05/is-washington-d-c-s-official-temperature-station-reading-way-too-hot/

 

Something new I learned is that the spread between temperature anomalies for IAD and DCA was similarly large back in the 1960s and 1970s.  Intuitively, this could make sense, as there has been more urban development around Dulles after 1980, and therefore an increase in UHI around Dulles would reduce its difference in anomalies with DCA.  (Although, what did IAD use for their "30-year normals" when the airport was less than 30 years old?)  

 

But why did this divergence resume 19 months ago?  Statistically speaking, this sustainced divergence is virtually impossible by random chance.  Did anything change at Dulles (or even BWI) recently?

[Eskimo Joe]

This again?

[MN Transplant]

I don't mind the topic.  We all know why DCA is warmer than IAD, but is there a good reason for that difference to change fairly rapidly?  The area around DCA is already dense, there isn't much reason that the UHI should have been more of a factor over the past two years.  Is it the sensor?  If the NWS says they've calibrated it, then it probably isn't.  River temps?  Weather patterns?

 

On question for Ian on the article, why bother with the anomalies?  A straight temp to temp comparison would make things easier. 

[Inverted_Trough]

To me, this topic is totally different than why is DCA warmer, is DCA representative of the area, etc, etc.  That's not the intent of my question.  It doesn't matter how warm DCA normally is -- the anomalies with its own averages shouldn't diverge much with anomalies for other airports in the same region.  Unless something changed in the environment (or equipment) at one of the airports in question.  CWG tends to be defenders of DCA's record, so if they're also questioning things - then the topic is worthy of scrutiny.

[AvantHiatus]

It's largely due to being near a large body of water. The same anomaly is seen along the whole coastal Mid-Atlantic which is primarily caused by very high mins due to enhanced ocean warming.

 

We've also seen higher elevated dew points during the summer in recent years aside from the prolific cold outbreaks in summer 2014.

[Ian]

I don't mind the topic. We all know why DCA is warmer than IAD, but is there a good reason for that difference to change fairly rapidly? The area around DCA is already dense, there isn't much reason that the UHI should have been more of a factor over the past two years. Is it the sensor? If the NWS says they've calibrated it, then it probably isn't. River temps? Weather patterns?

On question for Ian on the article, why bother with the anomalies? A straight temp to temp comparison would make things easier.

Yeah that was my thought too on anomalies. Jason wanted to run with it quick given other discussion that popped up yesterday and asked for anomalies. I provided a straight temp chart too and said I wasn't sure using anomalies was that helpful in the broader scope esp given the normalization process isn't that transparent. The charts look the same the y axis is just different. I mostly dug up the data and offered a bunch of thoughts and he merged it plus his stuff all together as the piece.

I've been less anti DCA than most party because I'm not sure what people expect for it to be representative of. At this point I wouldn't mind finding a specific issue but there aren't any smoking guns as far as I'm concerned.

[famartin]

DCA should represent downtown, the historic station location.  I think it does this reasonably well.  IAD is for the suburbanites.  I realize that most of the area is cooler and snowier than DCA since its on a ribbon of near sea level land surrounded on nearly all sides by significantly higher (relatively speaking) terrain, but that's how its always been, even when it was downtown.

[MN Transplant]

[MN Transplant]

DCA should represent downtown, the historic station location.  I think it does this reasonably well.  IAD is for the suburbanites.  I realize that most of the area is cooler and snowier than DCA since its on a ribbon of near sea level land surrounded on nearly all sides by significantly higher (relatively speaking) terrain, but that's how its always been, even when it was downtown.

 

I don't think anyone is arguing that, though I might argue that neither does their job particularly well (especially for low temps, DCA via the river and IAD via the cold pooling).  The question is why after almost 30 years of relative stability, especially in high temps, did DCA shoot up (or IAD down, I suppose) relative to the other in Jan. '14.

[famartin]

I don't think anyone is arguing that, though I might argue that neither does their job particularly well (especially for low temps, DCA via the river and IAD via the cold pooling). The question is why after almost 30 years of relative stability, especially in high temps, did DCA shoot up (or IAD down, I suppose) relative to the other in Jan. '14.

IAD and DCA represent the extremes of the area. You couldn't ssk for more.

[Ian]

I do tend to think DCA shows what's happening downtown fairly well. When it comes to snow they are quite similar imo. Most people argue this issue emotionally without a lot to back up their thoughts. Part of the reason for the big discrepancy in 90 degree days this year is because most of our hot days have been 'near 90' so DCA ends up just above while IAD ends up just below.

Someone did post satellite imagery showing that much of the grassy area north of the sensor was ripped up in construction in recent times tho I don't know if that would have any impact or if it is still like that.

I've also heard some say Dulles normals are too high.

Either way whatever is going on has at least some historical precedent it seems. If it continues a long time maybe there's something there.

[MN Transplant]

DCA and IAD vs the Arboretum coop

 

[gymengineer]

I do tend to think DCA shows what's happening downtown fairly well. When it comes to snow they are quite similar imo. Most people argue this issue emotionally without a lot to back up their thoughts. Part of the reason for the big discrepancy in 90 degree days this year is because most of our hot days have been 'near 90' so DCA ends up just above while IAD ends up just below.

Someone did post satellite imagery showing that much of the grassy area north of the sensor was ripped up in construction in recent times tho I don't know if that would have any impact or if it is still like that.

I've also heard some say Dulles normals are too high.

Either way whatever is going on has at least some historical precedent it seems. If it continues a long time maybe there's something there.

You can see a large discrepancy in central NC as well in terms of 90+ days. 

 

As for IAD normals being too high-- how is that possible unless there's equipment issues? Like, what does it mean for a 30-year normal to be too high without there being measuring or calculation errors? 

 

I would love for CWG to send a team over to DCA and take the temperature at different locations near/around the airport. It would be informative to confirm that the air temp at the airport is indeed close to what's being reported. 

[RIC_WX]

I for one have commented about IAD 1981-2010 normals being high, and they are compared to previously used 1961-90 or 1971-00 averages.  This does not mean they are wrong, per se, it just means that it may be easier for temps to appear closer to normal or even below normal because the baseline for normal has increased enough over the past couple decades.

 

A simpler way of explaining this is the current 30 year period of record is established with a set of normals which are likely the warmest in the recorded history of the station.  These "new normals" seemed reasonable from Summer 2011-Winter 2012-13, but outside of that atypically mild period it has been difficult to sustain much of a materially + anomaly at IAD this decade.  Which is what I think many initially suspected would occur after comparing the current data sets vs. prior averages.

[gymengineer]

I for one have commented about IAD 1981-2010 normals being high, and they are compared to previously used 1961-90 or 1971-00 averages.  This does not mean they are wrong, per se, it just means that it may be easier for temps to appear closer to normal or even below normal because the baseline for normal has increased enough over the past couple decades.

 

A simpler way of explaining this is the current 30 year period of record is established with a set of normals which are likely the warmest in the recorded history of the station.  These "new normals" seemed reasonable from Summer 2011-Winter 2012-13, but outside of that atypically mild period it has been difficult to sustain much of a materially + anomaly at IAD this decade.  Which is what I think many initially suspected would occur after comparing the current data sets vs. prior averages.

Gotcha. Of course, we're going to have to wait and see if the second half of this decade ends up with some stretches of major + anomalies, and IAD just happens to be on an accelerated increase in averages. Taking out the lows from the 1980's, especially during winter, may have a sizable impact on averages. 

[WxUSAF]

Heard the on-air met with WTOP today say that the DCA temp sensor was recalibrated over the weekend. Can anyone confirm?  

[Ian]

Heard the on-air met with WTOP today say that the DCA temp sensor was recalibrated over the weekend. Can anyone confirm?  

They actually replaced it entirely. Though I'm not sure there are any new answers from the ordeal.

 

http://www.washingtonpost.com/blogs/capital-weather-gang/wp/2015/08/11/weather-service-installs-new-temperature-sensor-at-reagan-national-airport/

[EastCoast NPZ]

"Lee explained that the temperature measurement at an ASOS site is within technical specification as long as its temperature readings are within (plus or minus) 5 degrees of the reference measurement."

 

+/- 5F.  Really?

 

 Why not just count frequency of cricket chirps to determine temps?

[H2O]

Well I now expect DCA to be the coldest reading in the area

[Ian]

Well I now expect DCA to be the coldest reading in the area

Can't wait till winter when the snow sticks easier now that the gauge is fixed.

 

The QC process seems kind of laughable. Maybe they should run two sensors at each location all the time. 

 

Still don't believe the answer is it's always been 2 degrees warm.. doesn't explain why it was basically right last spot check. 

[WEATHER53]

Nice article by Jason S. in Washington Post about it.

I have been on this for decades with little to no support until a while ago when a former state chief met got involved. It cast me a friendship but the head in the sand had to stop.  Looks like at least 2 degrees during the day and sometimes up to 4 and no night values but those generally run 4 to in some more extreme radiational events 10 degrees.  The idea that a small patch of earth where the sensors are could overcome tons of atmosphere, winds, snow cover, rain, was always weak.

Some of the really nasty detractors and disclaimers can begin thier mea culpas right now.

[WEATHER53]

"Lee explained that the temperature measurement at an ASOS site is within technical specification as long as its temperature readings are within (plus or minus) 5 degrees of the reference measurement."

 

+/- 5F.  Really?

 

 Why not just count frequency of cricket chirps to determine temps?

At least the heads are out of the sand and the FAA versus NWS stuff got put aside long enough to replace the sensors with the bent fins.  It's probably a tip of the iceberg and really the daytime maxes have been off 2-4 and the nights 5-10+ for a long time so even the acknowledgment of the 2-2.5 during the daytime calibration comparisions is a step in the right direction.

Now those who said I wore a tin foil hat will be very reluctant to change their stances but the cats out of the bag, thank you Jason and Mr Lee for your efforts to get this fixed.

[Inverted_Trough]

Since they claim the old sensor was about 1.8 degrees too warm, that would greatly reduce the number of 90-degree days they've had there this year.  Most of them seemed to be within the 90-92 range.

[H2O]

Can't wait till winter when the snow sticks easier now that the gauge is fixed.

 

The QC process seems kind of laughable. Maybe they should run two sensors at each location all the time. 

 

Still don't believe the answer is it's always been 2 degrees warm.. doesn't explain why it was basically right last spot check. 

 

instead of 33 and rain it will be 32.1 and rain

[RodneyS]

Can't wait till winter when the snow sticks easier now that the gauge is fixed.

 

The QC process seems kind of laughable. Maybe they should run two sensors at each location all the time. 

 

Still don't believe the answer is it's always been 2 degrees warm.. doesn't explain why it was basically right last spot check. 

I think two sensors would be a great idea.  Bear in mind, however, that the all-time high average annual temperature difference between DCA and IAD was not recently, but rather in 1976, at 6.0 degrees (57.9 vs 51.9).  Further, the all-time low average annual temperature difference between DCA and IAD was not in the distant past, but rather in 2009, at 1.7 degrees (57.4 vs 55.7).

[Ian]

I think two sensors would be a great idea.  Bear in mind, however, that the all-time high average annual temperature difference between DCA and IAD was not recently, but rather in 1976, at 6.0 degrees (57.9 vs 51.9).  Further, the all-time low average annual temperature difference between DCA and IAD was not in the distant past, but rather in 2009, at 1.7 degrees (57.4 vs 55.7).

Might as well.. can't imagine it would cost a lot. A single point ob for comparison every three months doesn't necessarily tell you much.  And yes, the fact that whatever was happening recently fit into a space we've seen prior made it hard to say there was a problem for sure based on the numbers alone.  

 

But again we still have the problem that NWS QCs every 90 days and the last one was a much smaller error. Anyone saying it's as simple as redacting 2 degrees off all past readings is just dumb... honestly a lot of the meteorologist response to the article on twitter has been dismaying. I thought these folks were supposed to understand science and nuance. Instead it's OMG WE KNEW IT WAS TOO WARM FOREVER AND EVER!!! lol..

[famartin]

This is not the first time I've seen an ASOS run "hot" for a while, only to get cool again once the thermometer was replaced. Probably won't be the last time, either.

[wxmeddler]

It's a macro accuracy vs precision problem. We fixed the precision at DCA, it is now showing the actual temperature at that location. But the accuracy still sucks because it's DCA. I've sorta given up on the accuracy problem, it is what it is, as long as the location doesn't change everything is relative.

[famartin]

It's a macro accuracy vs precision problem. We fixed the precision at DCA, it is now showing the actual temperature at that location. But the accuracy still sucks because it's DCA. I've sorta given up on the accuracy problem, it is what it is, as long as the location doesn't change everything is relative.

So long as the thermometer is reading correctly, it's representative of downtown, as it is meant to.

[wxmeddler]

So long as the thermometer is reading correctly, it's representative of downtown, as it is meant to.

 

Lots of large cities have their official obs stations away from "downtown". DCA is one of the closer ones respectively, it's just the geography that makes it "odd". If someone wants to make a national push to re-locate all the official obs sites back to downtown areas as they were 75 years ago, I am all for it. But as it stands now, DCA isn't likely the worst offender out there.