There are two climate sites at the Fairbanks International Airport. The main station is called Fairbanks International Airport. That is the official climate site. The secondary site is called Fairbanks Airport #2 and is closer the the terminal. When looking at the May data a few days ago, I noticed that the secondary site was actually below normal for the month wile every other station in the region was above normal (See Figure 1). As it turns out, the average daily temperature at the #2 site was 3°F less than at the primary station. Figure 2 shows the locations of the two stations as shown in the NCDC records and the May 2014 monthly temperature and departure from normal.
Figure 1. May 2014 daily average temperature at Fairbanks International Airport and Fairbanks Airport #2.
Figure 2. Station location for Fairbanks International Airport and Fairbanks Airport #2 along with the monthly average temperature and departure from the 1981-2010 normal. Note: the Airport #2 site uses an 11-year period of record for their normal calculation.
This begs the question of whether one station is reporting inaccurate temperatures. Since the primary site's temperature departure is consistent with the other regional stations, I am inclined to place much more weight on the primary station than the #2 station. Figure 3 shows the statewide departure from normal for the month of May 2014. With the exception of the area around Bettles and Nome, all of Alaska was above normal for May.
Figure 3. Statewide temperature departure from 1981-2010 normal.
Looking back of the Fairbanks Airport #2 period of record and comparing it to the Fairbanks International Airport station during the same time period should provide a good assessment of whether there is a consistent temperature bias. Figure 4 indicates that the Airport #2 site is consistently cooler than the main climate site. The average annual difference is 1.4°F.
Figure 4. Fairbanks International Airport and Fairbanks Airport #2 annual temperatures from 2000 through 2013.
Figure 5. Average monthly temperature difference between Fairbanks International Airport and Fairbanks Airport #2 from 2000 through 2013. Positive numbers indicate that the Fairbanks International Airport site was warmer.
Perhaps more interesting is the strong and consistent seasonal variation. As we can see in Figure 5, there is temperature parity during the months of February, March, and April. However, that transitions into a strong temperature differential for the remainder of the year – peaking in the high sun months.
So why is there a difference and does it matter? The answer to the first question is difficult to answer without inspecting the station equipment and its siting.Figure 6 shows the locations of the Fairbanks International Airport station since 1929. Interestingly, the location from 1952 to 1997 is very close to the current Fairbanks Airport #2 site. It is possible that they might actually be the same location.
In my opinion, the Fairbanks Airport #2 temperature readings are too low. As for the why does it matter question, the short answer is that better data always yields better results.
Figure 6. Station history for Fairbanks International Airport site from 1929 to present. Source: http://www.ncdc.noaa.gov/homr/
For comparison, I added three of the four main stations in Anchorage: Anchorage Forecast Office (PAFC), Anchorage International Airport (PAFC), and Anchorage Merrill Field (PAMR). I excluded Lake Hood (PALH) since it is located right next to a large body of water. The Forecast Office and the International Airport are pretty close to each other and are both quite close to the moderating influence of Knik Arm. Merrill Field is several miles inland and has a slightly more inland climate. That is particularly apparent in the warmer temperatures during the summer months. All-in-all, there is very little annual and monthly variability in the readings between the stations. The Forecast Office and the Airport have highly similar temperatures as one would expect – and as I would expect in Fairbanks. The Merrill Field temperatures are a little bit warmer on an annual basis. I attribute this to it being located in a more urbanized setting.
Figure 7. Anchorage, Alaska, annual temperatures from 2000 to 2013 for the Anchorage Forecast Office (PAFC), Anchorage International Airport (PAFC), and Anchorage Merrill Field (PAMR).
Figure 8. Anchorage, Alaska, monthly temperatures from 2000 to 2013 for the Anchorage Forecast Office (PAFC), Anchorage International Airport (PAFC), and Anchorage Merrill Field (PAMR).
Brian I believe #2 is co-located or nearby with the weather ballon launch site. There's an old white domed building very near that location. I assume it's for inflating and releasing the sensor.
ReplyDeleteThe 1952-97 site may be what was the WX observer's station in the old tower structure. Windows all around of course and probably shared the facilities with FAA staff until they moved the tower and flight service staff to the East Ramp. That structure probably went away when they remodeled the terminal, but I've not looked recently.
I'll drive by #2 today and have a look. The primary is in the Airport Secure Area and requires a security clearance.
Gary
Thanks Gary. I'll be interested in hearing your description of the station siting.
DeleteWell this is very interesting Brian. Great detective work! It seems rather unlikely that the #2 equipment has been in error with the same bias since 2000, so perhaps there really is a consistent difference in temperature between the two locations. The #2 location is immediately adjacent to the wooded area to the west, so perhaps that keeps the temperatures cooler there on average? It's not obvious why the difference would go away from February to April.
ReplyDeleteAn interesting implication is that if the primary site was at the #2 location prior to 1997, then the 1981-2010 normals are more than half based on temperatures that would have been cooler - unless an adjustment was included to account for the move.
Surely Rick will have some insight on this.
http://weather.uwyo.edu/cgi-bin/sounding?region=naconf&TYPE=TEXT%3ALIST&YEAR=2014&MONTH=06&FROM=1112&TO=1112&STNM=70261
ReplyDeleteThe station location Lat/Long coincides with the structure I noted above on Google Earth. They may use the surface temp data generated to calibrate the balloon's sensor prior to launch? Just a guess of course. There's trees around the structures except for a road facing north.
I'll have a look today as I walk the dog daily at my plane parking on the nearby float pond.
I just called the NWS this AM and visited with Ben. He helps launch balloons from that site and was kind enough to visit this Blog while we were talking. His preliminary thoughts are that there's trees nearby and a low slough/old water channel that could contribute to cooler temps. He said they use three sensors to cal the sonde prior to launch. If they're within tolerances (1 deg was mentioned) they're good to go.
He offered to join the fun here and maybe fill in some more info regarding Brian's obs on the #2 site at PAFA when he has the time.
Gary
For some reason #2's red dot on the map is placed over the airport's employee parking. The structure that Gary mentions is about 100 yards south. I also agree with Gary about the trees and land depression causing temp diffs.
ReplyDeleteVery curious. As for whether it matters... many studies, formal and informal, and many conclusions are drawn using airport weather data. However, the 1st intent and reason for the existence of airport weather stations is to support aircraft operations rather than to provide the quality of data useful for weather or climate research. So often the data is found to be fine for the former but frustrating for the latter. It could be better, but there does not seem to much priority in the system to make it and keep it so. This is a general rant...the temperatures from these two stations could be spot on. Even if they are, the issue of representativeness of airport stations is still pertinent.
ReplyDeleteAs to this case, I find the lack of difference between the two stations in Feb, Mar, Apr to be the biggest puzzle. Make the theory above (cold air draining out of the wooded area) less attractive.
When I first drove by the #2 in the mid-70's I don't recall the trees...the NWS facility and white dome were readily visible from the nearby road circling the airport. Old Guy memory of course, infallible.
ReplyDeleteAny terrestrial depression has been there since the airport was built. But shade and wind blockage from growing trees may have altered the sensor's output over time, if that's the major cause vs the open spaces on the airport proper.
Why the relative temperature parity during the months of February, March, and April? Insolation should be mimicked in the Fall, unless cloud cover during Spring is less (likely, as in this Spring particularly). Wind? It's probably windier in and around March. That's a good one to ponder.
Gary
I was intrigued enough to pull up the data for myself, and interestingly it seems the differences are mostly found in the minimum temperature in the summer and fall (over 3F difference from May through September) and maximum temperature in the dark season (2-3F difference from November through January).
ReplyDeleteRemarkably, June minimum temperatures average 4.5F warmer at the primary site - this seems like a legitimate microclimate difference, perhaps owing to the overnight heat capacity of runways etc compared to the sheltered #2 site, or perhaps higher nighttime wind speeds in the open area.
Minimum temperature differences are much smaller when snow is on the ground. Maximum temperature differences are smaller when there is significant insolation and/or clear skies. Feb-April have both?
Wow, this post has generated a lot of discussion. I appreciate everyone chiming in with their thoughts. Here are a couple of more from me.
ReplyDeleteFirst, I added a few charts of the Anchorage stations for comparison. In particular, I was interested in whether the Anchorage Forecast Office exhibited similar variability when compared to the Anchorage International Airport (they are very close to one another). The temperature were extremely closely related. Over the course of 14 years, they were identical. Even with micro-climate variability, the expectation is that they would even out over long periods of time given their close proximity – and they did. My house is farther inland than any of the stations and experiences far colder winter lows and warmer summer highs. However, after 6 years of data collection, it is only 0.15° different that the Forecast Office temperature over the same time period. By the way, the Anchorage Forecast Office and Anchorage International Airport stations are 1.6 miles apart vs. 0.93 miles apart between the two Fairbanks Airport stations.
Gary, your description of the station settings is very helpful. It certainly makes sense that it would be co-located with the RAOB launch site. Thank you for driving by.
Richard, the seasonal differences are what is especially intriguing. Thanks for digging deeper into the numbers and for the excellent summary of the findings.
Eric, the locations on the maps are what were provided by NCDC. They may or may not be accurate. I can only assume that they are. Over short time periods those micro climate difference can be significant.
Jim, I completely concur on the primary functions of airport weather/climate observations. Sometimes we have to make lemonade out of questionable lemons. The question is how much sugar needs to be added.
Another thought, several of the days in Fairbanks during May were cloudy, breezy, and wet. Those should be good days to compare temperatures since micro climate variations are minimal in those circumstances. Yet, there were 2° and 3° temperature differences during those days. That makes me wonder if it is more of a solar exposure difference with maybe a fan aspiration issue thrown in???
ReplyDeleteThe day-to-day consistency of the difference certainly supports your hypothesis. I see that the differences in May were again much larger for min temps than for max...
DeleteGary may be on to something with shading - if the #2 site is surrounded by trees, then it will have many more hours of shade at this time of year, i.e. more hours for radiative cooling?
Only one way to find out. Gary and Rick need to pull some strings to get permission, then go out there one night and run some transects with a handheld thermometer. A few passes will quickly reveal if the difference is real.
News from PAFA. I drove out to the radiosonde facility, AKA AP #2 and visited with Ben, NWS, great fellow BTW. Took some pics if someone wants to view. No gate. Like a shovel or rake, likely nobody's going to steal a work tool or WX instrument.
ReplyDeleteThe primary temp gauge lies next to a lower area (old slough bottom probably) and is near some vegetation. However there was a southerly breeze today and the anemometer on the unit was spinning away. He has a second temp probe on his facility, and also uses the ASOS at the south end of the airport. It's clearly visible from the road and in an open flat area.
Temps were 58-60F on my car therm both at the facility and driving around the airport. While the trees may offer shade at lower angles of sun, the gauge isn't buried in the pucker brush. Calibration is part of routine maintenance Ben offered.
I'm going back sometime to watch a balloon launch. Last time I did that deal was 1958 in Upper Michigan on a Geophysical Year school trip to the Weather Bureau's station in the Soo. We all signed the sonde and away it flew.
Gary
Excellent work, Gary. I'd certainly be interested in seeing the photos. If they're small enough to email, you could send them to my address: richard.james followed by @prescientweather.com (written that way to avoid spammers picking it up too readily).
DeletePics sent. Google Sat views. Zoom as required:
DeleteAP#2: https://www.google.com/maps/@64.816157,-147.8762664,489m/data=!3m1!1e3
PAFA ASOS: https://www.google.com/maps/@64.8029375,-147.8758176,490m/data=!3m1!1e3
Gary
Thank you Gary!
Deletehttp://www.nws.noaa.gov/ops2/ua/
ReplyDeleteHere's a link to the NWS Upper-air Observations Program co-located at AP#2 above in Fairbanks. There's lots of interesting info here available via the sub-links on the left.
Gary
Couldn't help but notice that AP#2 dropped to 31 F yesterday morning, while PAFA reported a 38 F minimum. Of course numerous other locations around the area were below freezing. Which is correct, or are both correct? Fascinating!
ReplyDeleteHere's my guess, providing the two airport sensors are similarly calibrated and provide accurate readings across the temperatures of interest .
DeleteThere's a man-created heat island at PAFA due to the lack of surface vegetation, exposed soils, and the extensive paving/structures nearby exposed to insolation. The sensor at AP #2 is over or near slightly lower vegetated and somewhat insulated ground surrounded by trees (see SAT views above).
In FMA the albedo and ground insulation are similar at both sites due to snow cover, therefore similar temps are noted in Fig. 5. During ASO, and other low albedo/snow cover conditions, insolation contributes to the heat island effect at PAFA more than at AP #2.
Gary (my dog suggested all of this)
Not done.
DeleteMy dog the Oracle has spoken again. She suggests that without an examination of differential anemometer readings between the two points of observation, the preceding guess may be lacking an important component.
In addition, unless the calibrations of each instrument are consistent across expected temperatures, then instrument error may also be a factor.
BTW, she's still waiting for the record Howard Pass anemometer to return to the living (http://www.wrh.noaa.gov/mesowest/getobext.php?sid=HOWA2).
Gary
You have mostly puzzled out the issues, and here's what I can add. The upper air building is adjacent to an old slough, so cold air drainage is expected. The air field was built up when constructed in the late 1940s (Fairbanks Airport was an island above the water in the 1967 flood), and so all whole airfield is slightly elevated compared to surrounding terrain and there is no mirco-topography left.
DeleteAs Gary noted, the Brian's plotted pre 1997 position is for the old tower where the observers were located. The thermometer locations can be somewhat traced from instrument history published in older LCDs. From personal experience, I can say the HO-83 sensor was located near mid-field just east of the main runway for many years (at least since the 1970s) until 1991. In the record snowy winter of 1990-91 airport operations had to modify snow plowing because the sensor almost was buried in plowed snow. So in the summer of 1991 the HO-83 was moved to off the north end of the runway in an area not close to plowing (more or less on the field across from Pikes and the time and temp sign). This move does not appear in the LCD published instrument record, but I saw it with my own eyeballs. In Dec 1997 the ASOS was commissioned. For a couple years prior to commissioning the ASOS observations were transmitted with "TEST", so we could see both the ASOS "test" obs and the HO-83 temps in the official obs. The ASOS site generally was colder under clear skies, and there was often significant differences in winds in these situations, no doubt reflecting very small scale drainage. This did not seem unreasonable, given the two sensors were more than a mile apart and the ASOS was much closer to the Tanana River.
Finally, there is the difference in instrumentation. The ASOS is an aspirated shield that receives regular calibration and maintenance. FAUA2 (and virtually all other COOPs) have an MMTS electronic temp system that never receives any maintenance or calibration except in the event of total failure. So there is no way to "correct" for, or even usually notice sensor drift.
And to second Mr. Green's remarks, ASOS was designed to provide for aircraft operations. Other uses have been hung off this basic design. This is one of the great advantages of the CRN sites, that have multiple independent thermometers.
Hope this advances the discussion.
Rick
That you Rick. Your first-person description of the station settings and history is most helpful. The station relocation not listed in the HOMR database is quire interesting. I wonder how often that happens.
DeleteIf the temperature difference between the two sites is real, the AP#2 station is cold enough for permafrost to accrete and the runway station is right on the border of that threshold (-2°C annual temperature).
the climate of any area/country makes a direct impact on the flights, in the extreme cold situations the flights get affected in a great way. there remains extreme weather conditions in Alaska. Airport Car parking at Luton
ReplyDeleteThat is right, it really seems that hey may use the surface temp data generated to calibrate the balloon's sensor prior to launch? Just a guess of course. There's trees around the structures except for a road facing north. gatwick meet & greet
ReplyDeleteGet your car parked in a safe vicinity with valet parking gatwick while you are away and rest assured that your vehicle is well taken care of.
ReplyDeleteThe temperature info that you shared is quite elaborating. gatwick meet and greet is a convenient service to get your car parked at the airport while you are off to your amazing vacation.
ReplyDeleteThe balance,or more accurately the imbalance,of infrared radiation across the earths surface is the fundamental driver of climate and weather. Valet parking Luton
ReplyDeleteA weather radar is used to locate precipitation, calculate its motion, estimate its type and then plan for the next. seo services
ReplyDelete