Reno's USHCN station

Last summer I attempted to do a survey of Reno’s USHCN official climate station. But I was thwarted by its placement at the Reno International Airport due to security and lack of accessible photographic vantage points. Reno’s USHCN station is particularly important due to it being part of the test cases of stations in the new USHCN2 scheme being implemented by NCDC. It’s also important due to it’s steep temperature trend which appears to be more of an urban heat island issue than a climate change issue. It shows up as a hot spot in USHCN contours done by Steve McIntyre.

reno-nv-station-plot.png

While there wasn’t good Google Earth photography online last summer, that has since been remedied, and high resolution photographs are now available at Google Earth and at Microsoft’s Live Earth. Having these, I was able to complete the station survey and determine that this station is a CRN4 rating due to proximity to the ILS building with a/c exhaust vents, less than 10 meters away. A CRN4 rating is unusual for an ASOS station.

reno-nv-asos.jpg
The Reno USHCN ASOS station undergoing repair, looking west.
(HO83 hygrothermometer repair perhaps?)

The Reno USHCN station is in the middle of the runway complex, between runways 16L and 16R. And what is interesting about that placement is that it seems the color/albedo of the surface where it is located is actually darker with a lower albedo than than of the nearby runways as shown below.

Reno NV ASOS aerial wide

You can see a complete photo collection of the Reno USHCN station here.

In wondering about just how this placement between runways on a darker surface environment might contribute to the upward trend in the GISS temperature graph shown above, I did some searching online and soon discovered that NOAA uses Reno’s placement problems as an example in a training manual for climate monitoring COOP managers. They’d already done all the work for me! More on that internal NOAA training manual later, as it has provided a wealth of information previously undisclosed.

What was amazing is that they’d already determined that there were significant problems with this USHCN station placement that contributed a significant warming bias to the record.

From that manual:

Reno’s busy urban airport has seen the growth of an urban heat bubble on its north end.
The corresponding graph of mean annual minimum temperature (average of 365 nighttime
minimums each year) has as a consequence been steadily rising. When the new
ASOS sensor was installed, the site was moved to the much cooler south end of the
runway. Nearby records indicate that the two cool post-ASOS years should have been
warmer rather than cooler. When air traffic controllers asked for a location not so close
to nearby trees (for better wind readings), the station was moved back. The first move
was documented, the second was not. The climate record shows both the steady warming
of the site, as well as the big difference in overnight temperature between one end of this
flat and seemingly homogeneous setting, an observation borne out by automobile
traverses around the airport at night.

They were also kind enough to provide a photo essay of their own as well as a graph. You can click the aerial photo to get a Google Earth interactive view of the area.

reno-nv-asos-relocation.jpg

This is NOAA’s graph showing the changes to the official climate record when they made station moves:

reno-nv-asos-station-moves-plot.png

Source for 24a and 24b: NOAA Internal Training manual, 2004-2007

What is striking about this is that here we have NOAA documenting the effects of an “urban heat bubble” something that Parker 2003 et al say “doesn’t exist“, plus we have inclusion a site with known issues, held up as a bad example for training the operational folks, being used in a case study for the new USHCN2 system.

It seems that Parker is looking more and more foolish with his attempts to make UHI “disappear” when NOAA references UHI problems with station placement in their own training manuals.

I’ll have more on revelations from this internal training manual from NOAA coming soon, plus I have my Paint Experiment Data collated, and I’ll be publishing results of that Latex -vs- Whitewash likely over the weekend.


44 Comments

  1. Alan Woods
    Posted Jan 10, 2008 at 3:04 AM | Permalink

    GISTEMP plot after homogeneity adjustment:

  2. Geoff Sherrington
    Posted Jan 10, 2008 at 3:44 AM | Permalink

    Anthony, has there been a steady improvement in documented UHI etc. corrections since you started your work?

    You might recall documenting the central Melbourne Bureau of Meteorology station. Dr Michael Coughlan, Head, National Climate Centre, BOM, Australia wrote to me on 26 Nov 2007 saying inter alia

    “I am sure you would agree that any allegation of scientific fraud is a very serious matter”. (I did not make one. I included the reference to Douglas Keenan).

    “The Melbourne site .. would not be used in scientific studies related to the monitoring of large scale climate change. Of course, when used in conjunction with climate records at sites well outside the city boundaries, this Melbourne site does provide a very good record for determining and monitoring the local changes in climate due to urban growth.”

    It is unstated as to whether other studies, including those overseas, use the raw or homegenised data, or none. “(other questions) are unanswerable by the Bureau as they relate to what to what others may have done with the data once supplied, and over which we have no control.”.

    “Obviously, it is better to have corrorobating evidence from a documented station’s history, but it is not always essential. Indeed, sometimes a change point is apparent even by a simple visual examination of the data”.

    I was given a 2003 paper on “Updating Australia’s high quality annual temperature dataset”, Della-Marta et al, Aust. Met. Mag. 53 (2004) 75-93.

    Next I was told that it could cost me to get more info, but that costing is expected to drop over the coming few years.

    Re your Reno story, I wonder aloud if the GISS homegenisation is accepted as official because it passes the test of “simple visual examination”. I have little confidence that the corrections are mathematically and physically complete and correct. What validation comparisons are presented? None?

    This, in turn, makes me worry about what other countries might be doing to Australian data, in comparative ignorance of “local rules”.

  3. Geoff Sherrington
    Posted Jan 10, 2008 at 3:53 AM | Permalink

    Re # 1 Alan Woods

    The first point on Anthony’s early graph is at about 10.1 degrees. On the later GISTEMP graph it is about 12.7 degrees. Although at first glance it seems that levelling the slope of the post-1960 temps is the main change, on second look the early years have been made somewhat hotter. Is there a credible reason for this?

  4. Alan Woods
    Posted Jan 10, 2008 at 5:02 AM | Permalink

    Re: #3 Geoff,

    Poking around the NCDC site I see that this weather station has only been in operation since 1931, so it would appear that GISS have combined it with another data set. Also, it appears to have been shifted around a lot since 1931, see:

    www4.ncdc.noaa.gov/cgi-win/wwcgi.dll?wwDI~StnSrch~StnID~20019168#GRAPHIC

  5. steven mosher
    Posted Jan 10, 2008 at 7:18 AM | Permalink

    re 3. note that anthony’s chart is daily min. the gisstemp is (max+min)/2

    The key would be to look atthe ushcn2 data ( min and max) and see what adjustment
    ushcn did in the SHAP file which is the adjustment for stations moves.
    then look at the gistemp adjustment.

    OR one could just eliminate non compliant stations from analysis and not fuss
    about with adhoc adjustments

  6. BarryW
    Posted Jan 10, 2008 at 8:22 AM | Permalink

    Don’t lose the main reason for ASOS: weather information related to aircraft flight operations. Placing it in the middle of the runway complex isn’t bad for that reason (although next to the ILS may not be such a good idea). You want the actual temps and winds that aircraft are going to experience taking off. For climate data, I wonder what the effect is of the aircraft exhaust washing across the temp sensor. I would think it would have a large impact on the max temp. I would expect to see it start in the 60’s when jets started to dominate and correlate with the number of flight operations.

  7. Severian
    Posted Jan 10, 2008 at 8:23 AM | Permalink

    Interesting “coincidence” that the difference between an eyeball mean of the temp between 1945 and 1975 is about 6-8 deg, which is the same difference they say they see between the north end of the runway in the heat bubble and the south end isn’t it?

    But the UHI effect doesn’t exist…

  8. Severian
    Posted Jan 10, 2008 at 8:23 AM | Permalink

    Interesting “coincidence” that the difference between an eyeball mean of the temp between 1945 and 1975 and the current high temp reading today is about 6-8 deg, which is the same difference they say they see between the north end of the runway in the heat bubble and the south end isn’t it?

    But the UHI effect doesn’t exist…

  9. D. Patterson
    Posted Jan 10, 2008 at 8:54 AM | Permalink

    Some airfields have a problem with dramtaic differences in weather between the opposite ends of a runway. Visibility can be 100 nautical miles plus, viewing distant mountain peaks over the 7 mile horizon, yet visibility down the other end of the runway can be less then 12 feet or four meters due to fog coming from a nearby body of water. Naturally, there are also comparable differences in sky condition, air temperatures, and relative humidity.

  10. GWA
    Posted Jan 10, 2008 at 9:10 AM | Permalink

    Basic question:

    Why is the mean temperature used to determine daily temps?

    It seems that the average temperature would be more appropriate. I’ve often seen cases where the high temperature for the day was recorded at midnight, then the remainder of the day was much colder. It happened just this week as the warm temps of early this week were replaced when a cold front went through. It was 62 at midnight, then fell into the 40’s for most of the day. It seems that use of the mean temp would skew the actual ‘average’ of the day. Of course the opposite happens sometimes with the low temp being recorded for only a short time.

    Anyone have an explanation for this?

  11. yorick
    Posted Jan 10, 2008 at 9:21 AM | Permalink

    I was driving between Montpelier, VT and Montreal the other day when there were record high temps reported. Per my car thermometer is was 47 degrees in Montpelier area, 59 degrees in Burlington, VT then went back down to 47 degrees north of Burlington. Go Figure. It makes you wonder too if CO2 exagerates UHI on account of a lot of UHI must come from radiated infrared from buildings, engine blocks, high density of plowed, paved roads, etc.

  12. Steve McIntyre
    Posted Jan 10, 2008 at 9:26 AM | Permalink

    This is a very important contribution to the station surveys and, in a sense, completes the information required for testing the forthcoming USHCN v2 calculation.

    REno has been discussed in previous posts last year http://www.climateaudit.org/?p=1628 , http://www.climateaudit.org/?p=1669 , http://www.climateaudit.org/?p=1671 . As noted then, USHCN had identified Reno as a type case illustrating the supposed success of their version 2 adjustments, scheduled to have been released in July 2007 but still unreleased (to my knowledge). They said:

    In the original HCN, the regression-based approach of Karl et al. (1988) was employed to account for urban heat islands. In contrast, no specific urban correction is applied in HCN version 2 because the change-point detection algorithm effectively accounts for any “local” trend at any individual station. In other words, the impact of urbanization and other changes in land use is likely small in HCN version 2. Figure 1 - the minimum temperature time series for Reno, Nevada – provides anecdotal evidence in this regard. In brief, the red line represents TOB-adjusted data, and the green line represents fully adjusted data. The TOB-adjusted data clearly indicate that the station in Reno experienced both major step changes (e.g., a move from the city to the airport during the 1930s) and trend changes (e.g., a growing urban heat island beginning in the 1970s). In contrast, the fully adjusted (homogenized) data indicate that both the step-type changes and the trend changes have been effectively addressed through the change-point detection process used in HCN version

    The proposed USHCN changepoint methodology relied on nearest-neighbor analysis. Because one of the geographic centers of gravity of the surfacestations.org effort was in northern California, I’d been very hopeful that this effort would yield sufficient information that the Reno adjustment could be analyzed in detail. It looks like all of Reno’s neighbors have been surveyed and the adjustment method can be considered.

    In respect to #1, the adjusted version used by GISS removes the most grotesque aspects of this record. However, the version prior to GISS adjustment is usually the version used in other temperature composites (NOAA, CRU and, in particular, the Parker UHI article). This needs to be clarified in Anthony’s post; on the other hand, your comment implies that the matter is dealt with more than it really is.

    One of the real benefits of this analysis may also be sharpening analysis outside the U.S. We’ve noted that temperatures since about 1992 in the major temperature composites (CRU, GISS, NOAA) outside the US have predominantly been from MCDW stations (which seem to be heavily urban airports.) In the U.S., there are concerted efforts by USHCN to adjust airports by reference to rural stations and the effort does seem to result in major UHI catches (as arguably for Reno.)

    If it has been proven that such efforts are necessary – and the Reno example surely shows this – then it becomes incumbent on the temperature surveyors to show that corresponding due diligence and adjustments have been done in China, Africa, South America, Indonesia etc. I am unaware of any such demonstrations. Instead of this we get things like Parker 2006.

  13. yorick
    Posted Jan 10, 2008 at 9:34 AM | Permalink

    You wonder too what the CO2 concentration at an airport where huge jet engines are operating is locally, and whether that would be trapping radiation from the airport near the ground. Is there an effort to create climate quality stations worldwide underway? The UN could actually prove useful in an effort like that. Ten years of that could really settle a lot of these issues.

  14. Raven
    Posted Jan 10, 2008 at 9:39 AM | Permalink

    yorick says:

    Is there an effort to create climate quality stations worldwide underway?

    Of course not. The science is settled. Their is no need to collect data that might cast doubt on that conclusion….

  15. Steve McIntyre
    Posted Jan 10, 2008 at 9:42 AM | Permalink

    #10. Availability. My impression from looking at detailed records is that most stations spend more time near the min than the max. So if there is an increase in the min, the increase is slightly underestimated by using the min-max-average in the absence of daily averages on a 24 hour basis.

  16. steven mosher
    Posted Jan 10, 2008 at 10:48 AM | Permalink

    RE 15. SteveMC I think if you look at Reno and Vegas you will see a similiar effect
    of the last few decades. Stable Max Temps and rising Tmins. Hallmarks of UHI.

    FWIW

  17. SteveSadlov
    Posted Jan 10, 2008 at 11:03 AM | Permalink

    Anthony as you well know, tumble weeds and dust are potential issues at that location. They use strong herbicides for tumble weed control but that leaves the dilemma of dust, since the ground cover has been abated. So, at Reno, they use a dust coagulant (maybe just waste oil, not sure the exact compound).

  18. Pat Frank
    Posted Jan 10, 2008 at 11:21 AM | Permalink

    #12 — “The proposed USHCN changepoint methodology relied on nearest-neighbor analysis.

    I’ve never entirely understood the logic of this. If the temperature records of urban centers are ‘corrected’ using the records of external rural sites to remove UHI biases, then the final information content of the urban site is entirely determined by the information content of the rural sites. The urban site is mere redundance of the rural sites. Why use it at all, then? It’s not an independent data set.

  19. Mike Davis
    Posted Jan 10, 2008 at 11:27 AM | Permalink

    Reno and Las Vegas are two completly different climates. Las Vegas has the sand, Tumble weeds and low humidity.

  20. Severian
    Posted Jan 10, 2008 at 12:02 PM | Permalink

    Steve McIntyre:

    My impression from looking at detailed records is that most stations spend more time near the min than the max. So if there is an increase in the min, the increase is slightly underestimated by using the min-max-average in the absence of daily averages on a 24 hour basis.

    This tends to reinforce what a completely meaningless metric global mean temperature is. Looking at the data for individual stations, we get the max, min, mean? Just the mean? Do they keep an account of the median and mode? It would seem those would be more informative metrics, and of course variance/standard deviation. If a site had the same mean after 10 years but the variance had gone way up, that would tell you something potentially a lot more interesting wouldn’t it?

    Even if they don’t compute the mean and mode due to temp being taken at only a couple of times a day, the variance would still be a very useful statistic, but do they even compute that or care?

    Yorick says:

    Ten years of that could really settle a lot of these issues.

    Ah, but don’t you know that we’re at a tipping point!? We have to act now now now, we can’t wait 10 years! :)

  21. steven mosher
    Posted Jan 10, 2008 at 12:15 PM | Permalink

    re 19. Mike D you dont understand. I am talking about the narrowing of the Diurnal
    range in both cities which has been rather pronounced. I’ll hunt around and post
    stuff up. OR you can go to ushcn and download max and min

  22. Mike Davis
    Posted Jan 10, 2008 at 12:19 PM | Permalink

    steven mo:
    I was only commenting on the tumble weeds and sand. I lived in Las Vegas for over 50 years. I know what happened there.

  23. steven mosher
    Posted Jan 10, 2008 at 12:20 PM | Permalink

    re 20. Severian. You can search the observer guidelines to understand the process,
    but I will give you the short version. At non ASOS stations ( which collect data at
    frequent intervals for airport ops ) the observer observes the instrument once a day
    and records the daily max and daily min, rounded to the nearest degree F ( in the US)

    One wonders if they round to the nearest C in the ROW.

    So that data exists at USHCN. the min and max are combined into a “mean” which is
    (max+min)/2 also rounded ( rounding rules unknown)

    Clear?

  24. Steve McIntyre
    Posted Jan 10, 2008 at 12:50 PM | Permalink

    #20. I disagree that it’s “meaningless”. There will be a very high correlation between the max+min temperature and the average temperature – and the max+min is a very reasonable “proxy” for the average temperature if you don’t have it.

  25. yorick
    Posted Jan 10, 2008 at 1:23 PM | Permalink

    The hard part was figuring out what has actually happened the past seven years, because it all depends on who’s doing the measuring, and whether it’s being done on the surface or by satellite.

    A Spot Check of Global Warming – NY Times

    It seems like only the GISS seems to be following the orders of the models. Roger Pielke is consulted. Commenter makes claim that clearly the GISS is the most accurate.

    For example, without being an expert, the accuracy of remote sensing as a methodology to evaluate worldwide temperature changes seems highly suspect. I would welcome a critique of the NCDC’s Global Summary of the Day dataset because the research that I have concluded showing mean changes since 1970 for 2,779 weather stations is significantly higher than what I am seeing published.

    –James Coyle

    http://www.terrafirmamapping.com

  26. SteveSadlov
    Posted Jan 10, 2008 at 1:32 PM | Permalink

    RE: #19 – Tumble weeds grow on the drier / more solar exposed parts of my ridge (Coastal Norcal mountains). Dust is a problem wherever there is no ground cover, all up in down the Cali and Northern Nevada vallies / lowlands.

  27. Severian
    Posted Jan 10, 2008 at 1:46 PM | Permalink

    Steven and Steve, thanks for the clarification. But, what I see reported and hawked is just a global mean, with no mention of max and min, that is the number that is reported for public consumption, and I think it’s pretty meaningless. This year has a global mean X more than that year type of thing. For individual stations or geographical areas, averaged max and min will be a useful metric, I can see that, but I don’t see most press releases giving that kind of info.

    But all the graphs I see and most of the raw data I see is in terms of a single mean number for a station for each year, I know the other info is there, the max/min, but I seldom see it in most plots.

  28. Roger Dueck
    Posted Jan 10, 2008 at 1:57 PM | Permalink

    I agree with Severian that the “global mean” metric is meaningless. The entire system was designed for purposes other than monitoring, to hundredths of a degree, global mean temp. The pervasive circular “corrections” are a symptom of this fact. It is comprable to using the old NORAD system for GPS satellite tracking, with the appropriate corrections, of course!

    Steve: global average has nothing to do with this post.

  29. Bruce
    Posted Jan 10, 2008 at 1:58 PM | Permalink

    Steve McIntyre

    There will be a very high correlation between the max+min temperature and the average temperature

    But in the mins go up and maxes do not, that will make the average go up.

    But … does CO2/GHG/AWG only work on mins? Or does UHI work on mins?

    I’m guessing UHI works on min temps more than max and CO2/GHG/AWG should work on both.

  30. Sam Urbinto
    Posted Jan 10, 2008 at 2:04 PM | Permalink

    How about this. The Max is 20C and the min is 10C so we get 15. But what if 23:55 of a day is at 20C and the other 5 minutes at 10C? Or vice versa?

    Assuming in the first place that the readings aren’t contaminated and that the sensor is calibrated.

  31. Severian
    Posted Jan 10, 2008 at 2:37 PM | Permalink

    That’s one of the points I was trying to get across Sam. We seem to get a max and min, and thereby a mean, but no median, and as far as I can tell no indication as to what time the max and min occurred at. These things would seem to be of particular interest, but if we can’t even run a collection of temp monitoring stations correctly now, as per Anthony Watts work, and we can’t be guaranteed of the accuracy and location of stations over a history of decades, chances of getting that info is pretty slim. We could start trying to collect such things now, but have no way without the usual hand waving that goes on to connect that with past measurements.

    Off the cuff, when the max and min occur and how long it stays low or high, during a day, would seem to offer significant information that might help identify why the temps were what they are, UHI, weather, global warming, etc.

  32. Sam Urbinto
    Posted Jan 10, 2008 at 3:13 PM | Permalink

    Yes, that was my point also. What’s sufficient? Every second, minute, hour, how often to read, and how to combine the data. (Although realistically, one would expect the heating/cooling slopes between min/max would be about equal, so in the lack of sudden abnormal changes, one would expect (as Steve said) that min/max is probably a pretty good proxy, at least in most cases.)

    That said, if I take a very well-sited station with properly calibrated measuring devices to get both wet- and dry-bulb temperature readings every 10 minutes and continually update the stats to get a more accurate picture of the mean and median temperature, how often does that diverge from just getting the min/max every hour at any given station?

    And that said, another matter to consider; assuming the above well-sited station, how well does the air 5 feet above a surface indicative of the area really reflect the overall temperature/anomaly that’s derived? Does a station (no matter how good; an iffy proposition in and of itself) that covers a 1000 km radius (for example); does that station really tell us “the” weather/climate in that area, or is it just a ballpark estimation of basicially what it’s like? (I pick the second….)

    I would ask why we aren’t using the (what I would assume to be more accurate) statellites to tell us what the readings are of the troposphere and of the stratosphere? That’s about 80+19% of the entire atmosphere! I think I know why; it only goes back about 30 years. (Interestingly enough, of course, is that the bulk of the anomaly trend is in the same time period….) Can’t have only 30 years, no way to claim it’s warmer than in a milllyun years, right?

    Or even better, why aren’t we trying to track energy balance as a total?

    Either is certainly better than mean samplings of tiny parts of the surface to 5 feet up.

  33. Severian
    Posted Jan 10, 2008 at 5:01 PM | Permalink

    Sam, you mention things that have been of concern/interest to me as well. Often, it seems to me that historical temperature data, as well as paleoclimate data, are the Climate Science equivalent of Electronic Voice Phenomena. People so desperate to tease a signal out of noise that they manage to convince themselves there is a signal there, and making exaggerated claims for what they can learn out of such data. The signal to noise seems to be negative, there are so many problems with the data, and in the case of paleoclimate like tree rings, so many variables, that it’s possible to squeeze the data until you get whatever answer you want.

    And yet, we see the ground data being used to “correct” the satellites. And we see, from Steve McIntyre’s work and others, significant errors and problems with the ground data, which is then used to derive a global metric when the ground data not only has significant issues and problems, but as you point out only covers a very small amount of the global surface.

    Having followed Anthony Watts work, we see station after station that display a significant warming trend, and also coincidentally shows UHI and other biases. When questioned about this, NOAA and GISS and such pat us on the head and tell us “don’t worry, we’ve adjusted the numbers to take out all bias” and we’re expected to believe them, despite the fact that the work of Steve McIntyre and others are showing that the problems are not adequately addressed.

    Maybe, I’m surely not expert at it, min/max/average is enough to tell you something interesting and unambiguously enough to make determinations based on the data, but it seems remarkably loose and uncertain to me. Hopefully more discussion will follow on this thread that enlightens me more about this subject, but my gut feel is that all this data is interesting but not sufficient to make any hard determinations and it sure doesn’t seem to be “adjusted” correctly.

  34. Posted Jan 10, 2008 at 7:41 PM | Permalink

    This from a report CLIMATE OF RENO, NEVADA by Brian F. O’Hara National Weather Service Reno, Nevada Updated in July 200d

    Change in location of [Reno] weather observing sites:

    The climate record for Reno, Nevada extends back to 1870, but the official observation site has moved occasionally during that period. By referring to U. S. Weather Bureau / National Weather Service (NWS) Local Climatological Data (LCD) Annual Summaries for Reno, a chronology of weather observation station moves can be derived.

    From December 1, 1870 until December 31, 1887 daily rainfall data was collected by the Southern Pacific Railroad Company agent at the rail depot in Reno. Starting on January 1, 1888 and continuing through November 10, 1905, maximum and minimum temperature data, and precipitation data, were recorded at the Administration Building (Morrill Hall) on the University of Nevada campus north of town.

    The U.S. Weather Bureau took over official observations for Reno on November 11, 1905. The Weather Bureau’s first observation site was at the Thoma-Biglow Building, on the northwest corner of First and Virginia Streets, and continued here until February 28, 1910. Starting on March 1, 1910 and continuing through February 28, 1934 weather observations were taken at the I. O. O. F. (Independent Order of Odd Fellows) Building on the southeast corner of Second and North Center Streets. The final downtown location was at the Post Office Building on the northeast corner of South Virginia and Mill Streets. Weather observations were taken here from March 1, 1934 through August 31, 1942.

    In order to assist the growing aviation industry, weather observing stations were opened at airports along the nation’s transcontinental air routes (Whitnah, 1961). The Weather Bureau’s station at Reno’s airport (then called Hubbard Field) opened on January 8, 1931. Twenty-four-hour duty, along with hourly weather observations and upper air observations, started at this time. However, weather observations were now being taken at two locations concurrently (Hubbard Field and the downtown Post Office). This situation continued from January 8, 1931 through August 31, 1942 (when the city Weather Bureau office at the Post Office Building was closed). On September 1, 1942 the airport became the official observation site for Reno. This airport continues to be the official weather observation site to this day.

    The transfer in official observation site to the airport constituted a change to a location a few miles outside of town (actually 3.5 miles south southeast of the Post Office). However, the airport was “farther away”, so to speak, then than it is now. In the 1930s and early 1940s this area was rural. The city of Reno has expanded in area over time and the Reno-Tahoe International Airport is now surrounded by an urban area. This difference between the two locations can be seen with precipitation, and especially temperature, data for Reno. There was a continual rise in average annual temperatures from the late 19th century through the 1930s (a gradual evolution from wood structures to brick and stone buildings and paved roads in town?). During the 1940s the average temperature was conspicuously less than it had been (since the readings were now being taken in a rural area). Average temperatures then started a gradual warm-up, with impressive rises during the 1980s and 1990s (urban heat island effects). In a strongly developed heat island, the temperature in the urban area can be up to 10°F greater than it is in the surrounding rural areas (Oliver and Hidore, 2002). Emphasis added.

  35. Posted Jan 10, 2008 at 7:41 PM | Permalink

    This from a report CLIMATE OF RENO, NEVADA by Brian F. O’Hara National Weather Service Reno, Nevada Updated in July 2007

    Change in location of [Reno] weather observing sites:

    The climate record for Reno, Nevada extends back to 1870, but the official observation site has moved occasionally during that period. By referring to U. S. Weather Bureau / National Weather Service (NWS) Local Climatological Data (LCD) Annual Summaries for Reno, a chronology of weather observation station moves can be derived.

    From December 1, 1870 until December 31, 1887 daily rainfall data was collected by the Southern Pacific Railroad Company agent at the rail depot in Reno. Starting on January 1, 1888 and continuing through November 10, 1905, maximum and minimum temperature data, and precipitation data, were recorded at the Administration Building (Morrill Hall) on the University of Nevada campus north of town.

    The U.S. Weather Bureau took over official observations for Reno on November 11, 1905. The Weather Bureau’s first observation site was at the Thoma-Biglow Building, on the northwest corner of First and Virginia Streets, and continued here until February 28, 1910. Starting on March 1, 1910 and continuing through February 28, 1934 weather observations were taken at the I. O. O. F. (Independent Order of Odd Fellows) Building on the southeast corner of Second and North Center Streets. The final downtown location was at the Post Office Building on the northeast corner of South Virginia and Mill Streets. Weather observations were taken here from March 1, 1934 through August 31, 1942.

    In order to assist the growing aviation industry, weather observing stations were opened at airports along the nation’s transcontinental air routes (Whitnah, 1961). The Weather Bureau’s station at Reno’s airport (then called Hubbard Field) opened on January 8, 1931. Twenty-four-hour duty, along with hourly weather observations and upper air observations, started at this time. However, weather observations were now being taken at two locations concurrently (Hubbard Field and the downtown Post Office). This situation continued from January 8, 1931 through August 31, 1942 (when the city Weather Bureau office at the Post Office Building was closed). On September 1, 1942 the airport became the official observation site for Reno. This airport continues to be the official weather observation site to this day.

    The transfer in official observation site to the airport constituted a change to a location a few miles outside of town (actually 3.5 miles south southeast of the Post Office). However, the airport was “farther away”, so to speak, then than it is now. In the 1930s and early 1940s this area was rural. The city of Reno has expanded in area over time and the Reno-Tahoe International Airport is now surrounded by an urban area. This difference between the two locations can be seen with precipitation, and especially temperature, data for Reno. There was a continual rise in average annual temperatures from the late 19th century through the 1930s (a gradual evolution from wood structures to brick and stone buildings and paved roads in town?). During the 1940s the average temperature was conspicuously less than it had been (since the readings were now being taken in a rural area). Average temperatures then started a gradual warm-up, with impressive rises during the 1980s and 1990s (urban heat island effects). In a strongly developed heat island, the temperature in the urban area can be up to 10°F greater than it is in the surrounding rural areas (Oliver and Hidore, 2002). Emphasis added.

  36. Posted Jan 10, 2008 at 8:35 PM | Permalink

    Sorry for the double post. Double clicks on a Mac can create problems.

  37. Geoff Sherrington
    Posted Jan 10, 2008 at 11:42 PM | Permalink

    Re # 3 Alan Woods

    My apologies to all for comparing a min temp graph to a mean temp graph. I slipped up. Geoff.

    p.s. What is more, I admitted it.

  38. Geoff Sherrington
    Posted Jan 10, 2008 at 11:47 PM | Permalink

    Re # 37 and # 5 Steven Mosher
    No, I did not slip up except just now. I was comparing the top graph of Anthony’s with GISTEMP which is now blanked out. Anthony’s top graph is “Annual Mean temperature” from year 1880.

  39. Geoff Sherrington
    Posted Jan 11, 2008 at 5:12 AM | Permalink

    Re # 12 Steve

    Agreed Reno is a very important example because it demonstrates vividly a number of points. I admit to personal confusion between the various homogenised versions because I do not track them as closely as you do.

    You mention sharpening data collection in other countries. The Aust quality network now touted has about 100 stations selected from 1,200 or so on the basis of length of record, rural status, etc. This was improved by a “homogeneity assessment using a number of objective and semi-objective techniques..” whatever they are. Cut to the chase, the main stations are now typically at airports at small towns inland, it at remote coastal places, typically lighthouses. I am uneasy with this, because if a fatal flaw is found with either class of station, much of the coverage will disappear either from inland or from the coast.

    Ref with stats description is Aust. Met. Magazine 53:2 June 2004, Paul Della-Marta et al. Some interesting comments on corrections using distant stations. Main conclusion to 2002 is 100 years of increase, mainly in min temp of about 0.6 deg.C.

    Lastly, a quick comment on average daily versus mean daily temp. I’d sooner average, because for some types of instrument a transient effect will give a false max or min of greater moment than the average. But we don’t have much more to use than max and min in most places up to 20 years ago. I don’t know if we now have thermocouple errors here when changing from Hg in glass.

  40. Posted Jan 11, 2008 at 8:07 AM | Permalink

    Naturally, the Reno airport is on the CRU list of urban airport stations (#72488000 at 39.5/-119.8). See http://www.climateaudit.org/?p=2138. Other gems on this list have names like Laguardia and Newark.

    The dark surface looks like soil that has been oiled to keep down the dust. Since the station is directly on top of this artificial heat-generating surface, I think the station would be an easy 5 on either the CRN or Meteo France scales, not a 4 as rated by Anthony.

    What level of adjustment is incorporated into the graph?

  41. Mike B
    Posted Jan 11, 2008 at 8:10 AM | Permalink

    That’s one of the points I was trying to get across Sam. We seem to get a max and min, and thereby a mean, but no median, and as far as I can tell no indication as to what time the max and min occurred at. These things would seem to be of particular interest, but if we can’t even run a collection of temp monitoring stations correctly now, as per Anthony Watts work, and we can’t be guaranteed of the accuracy and location of stations over a history of decades, chances of getting that info is pretty slim. We could start trying to collect such things now, but have no way without the usual hand waving that goes on to connect that with past measurements.

    Off the cuff, when the max and min occur and how long it stays low or high, during a day, would seem to offer significant information that might help identify why the temps were what they are, UHI, weather, global warming, etc.

    Jerry B had linked to about 6 years worth of airport station data (warning: lots of data!) that had actual temperature recorded hourly. From that, you can reconstruct the Tmax and Tmin data that would be recorded by time of day and estimate those biases. USHCN has an adjustment for time of observation bias, and from what I’ve seen, the science behind it is sound. However, I am unconvinced that the execution of the adjustment regime appropriately accounts for the bias. I’ve been working on this in my spare time for a couple of months.

    There was an entire thread about TOBS last year. I will say this, time of observation bias is complex. It varies by latitude, season (i.e. proximity to soltices and equinoxes), local conditions (Minneapolis is very different than Seattle), and there are year-to-year variations.

    Correcting for it is even more complex, because you not only need to know all of the things above, you also need a reliable historical record of the time of observation for each station. That might be the most difficult part of all.

    Hope this helps.

  42. steven mosher
    Posted Jan 11, 2008 at 8:17 AM | Permalink

    32 & 33.

    somewhere I have a study or two on the issue of sampling interval. there is also jerry bs data
    and crn data.

  43. Mike B
    Posted Jan 11, 2008 at 8:22 AM | Permalink

    Re My #41, the link does have the Reno airport data from 1984-1992 (NV23185C.ZIP). Data format, once unzipped is YY MM DD HH TEMP.

  44. Severian
    Posted Jan 11, 2008 at 8:59 AM | Permalink

    Thanks for the links and info Mike B (#41). That helps a lot. From what I can see, this is analogous to a problem I once worked with a ballistic missile accuracy errors, which were dependent on everything from time of year, altitude of launch, azimuth of launch, location of target, location of launch, etc. The factors were complex and interdependent, and almost impossible to accurately model. Determining a range bias to correct for these effects seems somewhat similar to TOBS in complexity.

Follow

Get every new post delivered to your Inbox.

Join 3,304 other followers

%d bloggers like this: