How not to measure temperature, part 88 – Honolulu's Official Temperature ±2

People send me stuff, some days my email explodes. Today I got all sorts of things about Obama and John Holdren and the new NCDC climate spincycle which Steve McIntyre has dubbed Chucky Returns Part IV. That one made me laugh out loud.

Something else that made me laugh today was this well done story (h/t to Andrew Walden) from reporter Tina Chau of KGMB-TV in Honolulu, HI. The NWS spokesman Tom Birchard was clearly flummoxed, and at the end of the interview said exactly what I’ve always said about the ASOS system and measurement of climate data at airports. Comedian George Carlin was right in his “Hippy dippy weatherman routine”: Why do they always give the temperature for the airport? Nobody LIVES there!

“ASOS…placed for aviation purposes…not necessarily for  climate purposes.”

Yet, ASOS weather stations at airports worldwide are in fact used for climate, and are part of the official climate record. In the US alone, there are 64 ASOS stations (that I’ve found so far) in the official USHCN climate record, plus there are hundreds in GHCN worldwide.  In my studies of the USHCN temperature network, I’ve found dozens of such poor siting examples even at non-airports. See my report here (PDF, 4 MB).

Our old friend the ASOS and HO-83 temperature sensors may be up to tricks again in Honolulu. It seems the temperature is a wee bit off and new records are being set by the ASOS weather station at the airport. I’m reminded of the similar situation in Tucson years ago that went on a long time before anybody caught it. I’ve found the HNL station, seen below. It is located at lat/lon 21.32403 -157.939467 There’s more than meets the eye. More after the news story.

Honolulu Temperature Records Questioned

Written by Tina Chau

The high in Honolulu Monday was 92 degrees. It was the hottest June 15 since the National Weather Service started keeping track and the 8th straight day we’ve broken or tied a record. But was it really that hot?

That’s what the experts at the NWS have been wondering. They settled their suspicions with a trip to the airport to check Honolulu’s official temperature sensor.

“We had one of our technicians visit the site and they did a side-by-side calibration and found the thermometer at the Honolulu International Airport was reading a little warmer than what his caliberation thermometer was reading,” said Tom Birchard, a meteorologist at the NWS.

It was two degrees warmer. There’s some wiggle room with the accuracy of the temperature sensor.

“Which means, if the reading is 90, the thermometer is only accurate to read within about two degrees so it could be anywhere between 88 degrees and 92 degrees.

Which means our records these past eight days may not be records after all.

“If it turns out, after further investigation of the thermometer the data were skewed,” said Birchard, “they could be stricken.”

Now see these ground level photos I’ve found of the HNL ASOS courtesy of NCDC’s photo library of ASOS weather stations.

Not only is there a faulty sensor at HNL, as indicated by the NWS meteorologist in the news story, take a good look at the site photo below.  The HO-83 temperature sensor is the little white mushroom shaped device.

There’s an asphalt access road directly adjacent to the ASOS temperature sensor. Plus a heat generating power transformer, and the requisite air conditioner exhaust for the ILS electronics building.

As we know from common experience, temperatures are almost always warmer near asphalt than natural soil or ground cover. From a scientific perspective we can cite either Yilmaz et al (PDF 2008 ) and the measurements of temperature differences over such surfaces, or we can reference NOAA’s Climate Reference Network Handbook which rates the likely temperature error of such placements.

See the Climate Reference Network Site Handbook (PDF 2002) including explanation of the CRN 1-5 rating system used by surfacestations.org

How close is the temperature sensor to the asphalt? I decided to use Google Earth’s measurement tool:

The answer: 1.6 meters, about 5 feet.

But what’s a little asphalt when the temperature sensor is surrounded by a sea of it at HNL?

The 2 degree temperature error in the ASOS record was only found because it exceeded the June 15th high temperature record for Honolulu. It seemed “odd” enough for somebody to check to be sure.

So the question we have now is, how much of the data in the plot above from NASA GISS is from influences such as urban growth, airport expansion, etc.  I’ve confirmed that GISS uses the Honolulu Airport for climate data since their ID of 425911820000:

Matches that of the National Climatic Data Center description for Honolulu Airport’s WMO code (91182)

Note the big step change in 1960 in the HNL airport data plot. Hawaii became the 50th US state in 1959. So it would stand to reason that 1960 would be a period of growth at the airport, or perhaps a station move to a warmer location. According to the National Climate Data Center, the ASOS station was installed on February 1st, 1998. Since then, the temperature swings appear to have been larger than in prior years.

But the nearby Honolulu Observatory temperature record doesn’t seem to have much of a trend, though it no longer measures temperature for climate records, a pity:

When dealing with climate change, it is generally accepted that the amount of temperature rise attributable is about 0.7°C or about 1.3°F. With errors of 2 degrees or more creeping into the climate record due to faulty equipment and poor station placement, are we so certain?