ASOS vs SJC vs Nearby Town

These three graphs are for, supposedly, nearly identical places. All come from Wunderground. The impetus for these graphs came from noticing that the temperatures were often a degree or two F higher on the ASOS than in nearby areas, and even higher than at the same Airport but on different equipment. We looked at this in this posting:

https://chiefio.wordpress.com/2010/10/01/sjc-san-jose-international/

Two of these stations are both at San Jose International Airport. One is the ASOS, the other is listed as San Jose Airport (why they have two is an interesting question, but does allow for a bit of A/B comparison…) ASOS is The Automated Surface Observing System.

Santa Clara essentially butts up against the West side of the Airport. So I’ve also got a (semi-randomly picked) Santa Clara site as well. These are the three graphs for conditions on October 1, 2010. They all are slightly different.

First up, Santa Clara. Away from the airport, but not so far for it to have different weather. And well inside the San Franciso Bay Area mountain ringed Urban Heat Island Bowl…

This is is a link to the Wunderground page for that site. The “history” button lets you get these graphs.

http://www.wunderground.com/cgi-bin/findweather/getForecast?query=95053&MR=1

Santa Clara Wunderground report for 01 October 2010

Santa Clara Wunderground report for 01 October 2010

Notice that temperatures start at about 61.5 F, then rise smoothly as the day begins to about 70F, followed by a nearly equally smooth drop down to just under the 16 C mark on the far right.

Next we have SJC. That is the airport proper at San Jose International:

This is is a link to the Wunderground page for that site. The “history” button lets you get these graphs.

http://www.wunderground.com/cgi-bin/findweather/getForecast?query=sjc&wuSelect=WEATHER

SJC San Jose International Airport - non ASOS system

SJC San Jose International Airport - non ASOS system

Notice that temperatures start at about 62 F, then rise smoothly as the day begins, but with a ‘hickup” about 10:45 AM, then on to about 72F, followed by a nearly equally smooth drop down to just under the 16 C mark on the far right.

So far, so good. A tad warmer in the middle of the hottest part of the day as the sun warms the tarmac and the traffic has picked up.

Finally, we have the ASOS station.

This is is a link to the Wunderground page for that site.

http://www.wunderground.com/weatherstation/WXDailyHistory.asp?ID=MKSJC

ASOS Recording for 1 October 2010 at San Jose International Airport

ASOS Recording for 1 October 2010 at San Jose International Airport

What can I say? Look at that wiggle between 10 AM and 1 PM. That sure looks to me like some odd kind of contamination problem. Jet wash? Solar heated electronics? Who knows. But given that all the other stations in the area show smooth curves, this is just kind of bizarre.

Then we look at the numbers. Starts at about 62F then takes a ‘step function’ to about 64 F. What’s Up With That?!

And if that little dinky “step” was not on todays graph but was on yesterdays, would we have started 2 F higher than the other stations in the area? On to mid day… Nearly a flat straight line to 10 AM at 64 F (with an odd “divot” just before 5 AM) while the others start warming at 8 AM (but in fairness, use a significant part of that time catching up to the ASOS 64 F reading… or maybe that’s a problem too…) We’ve already discussed the 10-1 lunchtime wiggle, but it is just bizarre, really. The peak is about 72 F from 1 to 4:30, but with a “jump” to about 72.5 for just a few minutes at 3:30 pm. (Perhaps a departure?…) That’s 2.5 F higher than “nearby” and about 1/2 F higher than “at the same place”. Also I note that the 72 F ‘run’ is longer. SJC is from 2-4 pm, significantly shorter time ‘at temperature’.

Finally, we come to the end. There is this odd “stair steps down’ effect to the chart. As though some electronics are hanging onto a value until it reaches a threshold transition before it will allow a lower reading through. Ending with the final temperature just a bit ABOVE the 16 C line on the right hand side.

Clearly the ASOS warms more, hangs higher, and ends warmer. Clearly also it has some significant impacts from odd temperature jumps of a couple of degrees during the “rush hour” of mid-morning flights. And just as clearly, it is reluctant to let go of a temperature one it has found one. It only cools in ‘fits and jumps’ and never as much as it’s neighbor.

To my eye, there is clear evidence for a significant Airport Heat Island, for air traffic and solar heat impacts on the ASOS, and for some “odd” non-climatology related behaviour from the equipment.

In fairness, were I designing an automated weather system for airports, I’d make it “upward sticky” with hysteresis to prevent accidentally reporting a wrong low temp. If a pilot gets a 2 F higher reading, his Density Altitude calculations still let him fly, but with more margin of safety. If I report a 2 F lower blip for 5 minutes, he calculates, then it warms 2 F; well, his Density Altitude may exceed the ability of the plane to perform properly. So the Precautionary Principle would tell me to hang on to peak highs and only slowly let them go. It would be interesting to know if that is designed into the ASOS (or if this odd behaviour is do to the position in the middle of a brown field next to black tarmac full of jet exhaust).

In Conclusion

It is my opinion that this kind of “study” needs to be done on ASOS (and the related automated systems) at airports around the world. They need to be characterized for time lags, calibration to standards and to neighbors, and for ‘representative nature’ of the actual nearby climate areas. Until that is done, the ASOS are clearly biasing the temperature record up by about the amount imputed to be “Global Warming”. Given that an ever higher percentage of the GHCN Global Historical Climate Network data come from such airport sited ASOS like systems (up to 92%+ in some countries), this is a significant potential for error.

Further, such a study ought to be looking at differential temperatures between ASOS and nearby climatologically clean stations (including a specific cross calibration to a LIG and MMTS). In particular, the degree of Airport Heat Island at the ASOS, the drift of Airport Heat Island to nearby non-airport stations if a wind is blowing their way, and the impact of wind, rain, snow, and other specific weather events on the disparity between ASOS readings and both LIG and MMTS at the airport and in “nearby climatologically clean” locations needs to be assessed.

We are about to commit $Billions, and perhaps even $Trillions of national treasure to mitigate “global warming” that might well just turn out to be “splice artifacts” from shifting our temperature history from Liquid In Glass equipment over grass fields to ASOS over tarmac near jet exhaust. At a minimum we ought to make sure we have an understanding of what that change has done to the temperature history. To say “nothing” or “we can adjust it out” is not enough, because it isn’t “nothing” and we can’t “adjust it out” if we don’t know what it really has been.

Postscript on Anomalies

Since some ill informed folks like to assert that “anomaly processing” would remove this impact as you would need a station to be constantly increasing it’s warming trend for an anomaly to show warming over time:

This notion fails on a couple of counts. First, it assumes that you are calculating an anomaly “self to self”. That is, the particular thermometer is compared to itself over time to find the “warming” anomaly. Unfortunately, the data series codes, such as GIStemp, do a “Grid / box” anomaly. They compare a location today to whatever was in that location in the past. So the ASOS at SJC today (complete with the international air terminalS; as there are 3 now) to the temperatures from 1950 when they had a small private plane field and no giant jet port. AND to the Liquid In Glass thermometer then.) Such “apples to frogs” comparisons will give an “anomaly”, but it will be one that “locks in” the offset between temperatures then and now (even if you try to use ‘magic sauce’ to ‘correct’ it out.) You can’t compare my Mercedes SLC now to my VW then and find that cars have doubled in speed.

The second mode of failure is related. “Splice Artifacts”. That form of making a temperature series effectively splices together different instruments and sometimes even different locations over time. That Santa Clara location in 1920 gets splice onto the San Jose Metro airport in 1960, then onto the SJC International ASOS in 2010. This will show a steady ‘warming’ of about 2.5 F (per the above comparison charts) just from splicing different instruments together. It does not matter if you call that splice “homogenizing”, The Reference Station Method, or “adjusting” and “combining locations”. It is still a splice. All that such data really measures is the degree of error in the spice, not the actual temperature trend of an instrument.

I’ve made a temperature data series that does do anomalies “self to self” and avoids some of the “apples to frogs” issues. It does still have the “spice artifact” issues (but does NOT smear out the impact via the various averaging and “homogenizing” steps of GIStemp… it leaves it crisp and highly visible). That data series shows a clear “hockey blade” in the 1987-1992 time span. The same time when the thermometers were being changed, automation was being brought in, and the “QA Process” was being changed to give a collection of “nearby” ASOS stations final word on what temperature would be allowed from non-ASOS instruments. That “pivot”, IMHO, is showing us that the ASOS, and related, stations “have issues” and that those issues make it into the GHCN in strong form, enough that GIStemp and related data codes are mislead into finding “Global Warming” which is really just a glorified “splice artifact” hidden by their “homogenizing”, “combining” and “adjusting” processes.

To see samples of that analysis:

https://chiefio.wordpress.com/2010/04/11/the-world-in-dtdt-graphs-of-temperature-anomalies/

There are more postings, and more detailed reports. under the dT/dt category at the right side.

About E.M.Smith

A technical managerial sort interested in things from Stonehenge to computer science. My present "hot buttons' are the mythology of Climate Change and ancient metrology; but things change...
This entry was posted in AGW Science and Background, NCDC - GHCN Issues. Bookmark the permalink.

22 Responses to ASOS vs SJC vs Nearby Town

  1. Phil says:

    Please take a look at wind speed (not wind gusts). The top chart, with the exception of one single spike, stays under 5 mph. The middle chart maxes out at about 12 mph. The bottom chart reaches almost 15 mph at one point. How far apart geographically are the three stations in miles? And, would sustained wind speeds really change that much over such a (presumably) short distance? Jet exhaust, on the other hand, ….

  2. E.M.Smith says:

    @Phil: Good Catch!

    The airport has large metal slatted blast barriers at each end (the wind tends to come from the end) so jet wash can’t hit the cars on the frontage roads.

    The top graph, Santa Clara, will be residential and treed. So I’d expect the houses and trees to cut the wind considerably.

    The two at the airport are, in theory, within about 10,000 feet of each other (two ends of the runway) modulo some added at the ends of the runways and the fact that the thermometers are not actually AT the ends… And the intervening turf is pretty much, well, turf with tarmac…

    If I pegged the ASOS right, it’s about 1/5 the way in from the N. end of the runway just to the W of the runway. It would get considerable “wake turbulence” there along with jet wash. The “spikey” nature of the winds is of interest too… Out there by the runway it’s hot, sunny, brown / dark grey, dry, and exposed to the winds and wash. These graphs look like that kind of place…

  3. Steven Mosher says:

    Nice.

    Now, Calculate tmax+timin/2

    I see the following

    santaclara ( 70.5,61) 131.5 = 60.75
    Airport ( 72, 61 ) 133 = 61.5
    Asos ( 72.5, 61) 133.5 = 61.75

    But you have to know the asos exactly because the figures get rounded up and round down.

    So, start by getting a couple more santa clara stations, and then at least a years worth of data so you have different sun/could/rain conditions and you have a first good estimate for the effect at this station.

  4. Steven Mosher says:

    august 24th, 2010 is a good one to show. nice clear difference in Tave

  5. Chuckles says:

    What a bizarre statement.

  6. E.M.Smith says:

    @Steven Mosher: Yes, a heck of a lot more work to be done in this “Dig Here!” spot.

    This posting was just to point out the garish makeup and the too tight pants… but yes, having determined the character of the ASOS, now we have to start counting the quarters to see how much she costs …

    Needs to be done globally (though likely some regional comparisions would give a decent idea of the impact) and it needs to be done for ALL days in a month, then compared to the official monthly values. Then for the whole year.

    The hard bit, though, will be ‘backing out’ the QA process where a daily value may SILENTLY be replaced by an average of ‘nearby’ ASOS stations. Basically, we don’t know how much of the “non-ASOS” stations are actually silently filled in from ASOS. More “dig here”…

    I did do a semi-random sample of a dozen or so locations in an informal “look at the WU listings” way prior to this post; and the effect is present globally and often. Though not 100%. (Chicago w/ thunderstorms being the only example I ran into, but one is enough to assure it’s not a constant thing). It’s “on my plate” now to look at more stations in more places, try to characterize the differences with latitude, altitude, cloud cover, wind, … and see if I can get similar offsets from the monthly records… and…

    So much to do for a volunteer with zero funding. I suspect I could do more with the rounding error in the “Climate Scientists” budget than they get done with the whole thing. Ah well, I’ll just have to make some more “stone soup”… A dish best served cold.

    The thing I find most intriguing is that ‘step function’ and ‘sticky’ nature of the ASOS. It looks very much like a design goal of some sort and under electronic control. I’m sure a person who knows the system could figure out what causes it in a heartbeat. From this side of the keyboard it’s going to be a bear to hunt down… (Polar? ;-)

    It will also be interesting to see if I can find more airports with both a traditional temperature station and an ASOS as at SJC. It would be very nice if that turns out to be a common things. Kept the old one when the ASOS was installed as a ‘backup’ of sorts? Or perhaps just as it was from a different funding source / agency? We’ll see.

    At any rate, I put this example up for other to see in the hope that more folks will pitch in and get some of the investigations done. This is a ‘target rich’ place to look; and the Wunderground angle makes it easy for anyone with a keyboard and a few minutes to take a look at THEIR nearby large city and see if there is an ASOS, AWOS, AWSS nearby, then check discrepancy.

    http://en.wikipedia.org/wiki/Automated_airport_weather_station

  7. E.M.Smith says:

    @Chuckles: Which statement is that? (I know, I make so many ;-) Or perhaps it wasn’t me? – in my best little kid voice..

  8. Steven Mosher says:

    Or you can look at monthly summaries.

    Comparing sept for all three sites is instructive.

    Things to be aware of:

    1. You cant compare temps, you have to normalize the stations because of differences in altitude. That actually works in your favor here since the PWS is at a lower altitude.

    2. Note the wind differences. you’ll see much higher wind velocity at the airport. Primarly because the PWS you selected looks to be wind sheltered. The higher winds at the airport ( anything above 10m/s) are going to mitigate the warmth from the tarmac which you should be able to see by comparing
    windy days with non windy days. That would be good work to recheck the science on that.

    3. Check the shading on your PWS, if the GPS is correct on that site it is in a shaded location. look for other locations as well since we care about trends and not absolute temps.

    So, to really nail this, you would need to collect a longer series of data to make sure you sample over a wide variety of weather conditions. You need to normalize the altitude ( anomaly is for this) and you need to include some PWS that are not in the shade part of the day.

    Then you have a chance at quantifying the effect it would have on the trend estimation.

  9. bruce says:

    its becoming increasingly clear to me why my success might be described as modest.

    I am in awe of your efforts, thank you.

    What a truly Herculean effort, no beyond that, beyond impossible to ever make hide or hair from this scat. How one even defines Earth’s temperature is a riddle.
    I never understood the reluctance to reveal method in developing models of climate prediction. In the few pages I’ve viewed here I see why.
    It would be easier to develop an accurate depiction of an offspring three generations hence. Given no more than a single persons appearance, their parent’s (same sex) fondness for a body type and a geographic location.

    But I digress. Add me to your list of interested readers.

  10. E.M.Smith says:

    @Steven: I think you are being a bit too enamored of theoretical angels and heads of pins. This is an incredibly flat flood plain. Any difference of “altitude” is just irrelevant. (There is more difference between the freeway overpass and the airport than between the dirt in Santa Clara and the dirt at the Airport. Santa Clara says 12 feet, SJC 56 ft. So 44 ft difference? The wake turbulence vortex off a 747 will be mixing to that depth all on it’s own. Then the air flow around the 30 to 100 foot buildings will further disrupt the boundary layer effects. The slatted ‘wash fence’ at the perimeter is about 30 feet tall IIRC and induces turbulent mixing. Differential clouds, rain, and local impacts like sprinklers and passing trucks (or 747’s) ought to matter more by a mile.

    The wind is NOT going to mitigate the warmth of the tarmac (though it might move it around a little) as everything for miles around upwind and toward SJ is either more tarmac, concrete, freeways, car parks, stone works, concrete slab buildings, etc. etc. With a strong enough wind the cool bay air will start to intrude toward the airport and after going over several miles of industrial park get further warmed AT the airport (thus the occasional warmth in Willow Glen down wind, IMHO). You see this best by comparing Palo Alto / Moffett with SJC (as the first two are directly facing the bay). That’s why I put them in the first comparison, to see how much warming there was right at waters edge v.s. a bit away. Moffett still warms more than Palo Alto, but both have cooler temps from having less ‘warmup run’ between the bay and them. For SJC / Santa Clara the air is clearly warmed a couple of degrees before it gets there. The lateral difference between them is nothing compared to the upwind component.

    The typical wind pattern here is “off the bay down the valley toward Gillroy”. There is a bit of an eastward drift to it (wind from the west), so you can get SJC having the air arrive from Moffett rather than on the shallower angle from the bay directly. But it’s basically a channel between two mountain ranges. So “mixing” over the runway is less of an issue than “distance to bay upwind”

    Per the Santa Clara site, I essentially took whatever 95050 gave me so as to AVOID a “cherry pick”. It’s the “default”. IMHO, it is more representative (even, or especially, if shaded) as that is more like what the place would have been like in 1800. Part of the AIRPORT Heat Island Effect is, IMHO, the removal of the transpiration cooling from normal foliage. It would not have been a concrete jungle. So the comparison of a station under trees over grass would be a BETTER comparison to the airport for the purpose of showing just how much of the ‘warming signature’ is really ‘land use’ from a treeless airport being spliced onto the end of the data series.

    So no, I won’t be cherry picking another station in a field of concrete. In theory we already have that with the SJC station. I did look around at the collection of ‘nearby’ stations as far out as Sunnyvale and San Jose South. The pattern is pretty clear. Cooler away from the airport and cooler closer to the bay (during summers). I expect it to be warmer near the bay during winters, but we’ll see.

    Finally, the “trend” can not be found. ASOS is a new technology (too short a life to find it’s ‘trend’) and the airport has had massive growth over the last 50 years (so land use pollutes the history and you can’t separate land use from 60 year weather trend). It’s not possible to disambiguate any of those. What can be found is that the ASOS has a warm bias and that the “climate data codes” make a splice prior to doing their grid/box anomaly step. That alone will invalidate them for “global warming” use.

    Finally, it would be interesting to pick a nearby station that has had less land use change and find the trend for that station (if one can be found). If a clean station CAN be found, then the trend at that station would say what is likely to be the actual ’60 year weather’ trend ( I expect the trend to be zero with a ‘flat roll’ of about 60 years to it, with 1930’s and 1990 high, and 1900 and 1960 low; which would place us as headed back to low: which we have been doing.)

    Comparing that clean history to the temps at the stations at the airport and in Santa Clara would let us tease out a UHI / land use signature and an equipment signature. Then comparing all THAT to the GISS version would let us show how much of their ‘history’ is computer fantasy, how much equipment change, how much land use, etc.

    So you may ‘feel free’ to go chasing “trend estimation”. I’m going hunting for error numbers. (For the simple reason that the errors are real and demonstrable while the ‘trend’ – especially of a ‘grid / box’ – is a construction based on questionable assumptions.)

  11. Wayne Job says:

    I have a question for those on this site including Mr Smith, being as I am some what lacking in mathematical system analysis. My question is given the fortuitous warming of the thermometers by UHI after dropping or losing thousands of rural temperature measurements. The powers that be can find no real warming for the last decade or so. Thus with all the fudges, the homoginisation and pasteurisation of the temperature record, no warming. What then and by how much has been the cooling.

    I have a feeling as a practical engineer and an observer of my surroundings for much of my three score years and five that it is at least 1oC and maybe more. People in America are reporting very early bird migrations, this year ducks that normally start arriving for the southern Australian spring from Northern China do not have far to travel. They stayed the entire year in Australia and had two sets of young and are at it again. These ducks are now quite at home in our small towns as they are protected and accept people as part of their environment. Makes one wonder if they know more about the climate and its ups and downs than we do.
    I would like an answer from those with a better understanding than myself, for I fear that the northern winter coming may be some what harsh, if not fierce.

  12. E.M.Smith says:

    @bruce: Thanks for the compliments. The “sounds suspicious” reaction is basically the reaction I was having about 3 years ago when I was a ‘newby’ to this area and started looking into “What was causing all the warming”. The more I looked, the more I felt like “WTF, they are just making this stuff up. You can’t know that.”

    So welcome to the first stage of the journey…

    A peek ahead: The more you look, and the more in depth and detailed, the worse it looks for “climate science”. There is a lot of history of my journey here. The posts that are in a category starting with “AGW” on the right margin are probably the better place to start (after the intro under the GIStemp tab up top), then proceed into the dT/dt graphs and the NCDC/ GHCN category. Enjoy.

    @Wayne Job: Well, I hold a math award (that came with money, not just a piece of paper…) from high school and have worked as a systems analyst for many years, so we’ll see if those skills can help answer your questions.

    The first problem you run into is a subtle one. “How much has been the cooling?” (or warming) depends entirely on 2 things. The easy one is “cherry picking the start date”. The hard one is “how big is your ruler”.

    Start Date: Depending on when you choose to start measuring your series you can create any warming trend you want. Start 20,000 years ago, you get about 10 C. Start 8000 years ago, about -3 C. Bottom of the Little Ice Age in the 1700’s? About +3 C. Top of the Medieval Warm Period? About -1 C. 1970 to date? +1/2 C to 1 C (depending on what ‘corrections’ you make).

    The bottom line of all that is that we started with a frozen Ice Age, jumped up to the start of the interglacial warm period and got so warm that even now we are finding things ‘under the ice’. Then started a very slow 10,000 year drift downward toward cold and a new ice age glacial. But it takes about 15,000 years. On top of that is a 5200 year wave (Utze the ice man has been under the ice since that last time this wave was at a peak, and we are finding plants from 5200 years ago under receding glaciers in Peru. We’ve been this warm before…). Then there is the 1470 year “Bond Event” cycle. There are also 210 or so year solar cycles and 60 year ocean cycles (PDO) and even 11 year sunspot cycles.

    And with all those cycles on cycles, you get a very wiggley wave. So despite the overall downward drift from 10,000 years ago, there are times it has spiked much warmer and times it has spiked much colder. Some recently. (Like the Ice Age Scare of the 1970’s and the Little Ice Age to the downside, and the very hot 1930-40 period along with the ‘global warming’ scare of today).

    But taking your question as meaning: “How much of the reported temperature is ‘puff’ from UHI and land use? I’d guess it’s about 1 C. Looking at things like the thermometer comparison above, that’s about what you find. But the very idea of that number ‘has issues”.

    How Big Is Your Ruler: This gets technical… but it’s a common issue to deal with. Fractals. It’s not as complicated as it sounds. Everyone who buys land deals with it. Take a chunk of mountainous land. Measure it 40 miles “on a side” and you get one number of acres. (Basically, your ruler is 40 miles long). Now go back and let your ruler follow the terrain up and down the hills. It’s now longer than 40 miles. How long? Depends on how bumpy the terrain is and how small you make the ruler. Do you measure the up and down of a boulder 20 feet high? A dirt hill 10 feet? A pebble? Each gives a different answer.

    So a common ‘game’ in real estate is to buy a large chunk of land based on ‘perimeter’, then subdivide it based on ‘surface area’ and sell more acres than you bought…

    (Think of a vertical cliff of 209 feet wide by 209 feet tall. It would be zero acres on a perimeter measure, but it has a surface area of 1 acre…. You never want to buy vertical acres ;-)

    So what does this have to do with temperatures? Easy. The whole notion of getting an ‘average temperature’ is slightly broken simply because temperatures are also fractal. (They depend on surface features that are fractal). So the answer will vary with the size ruler you use. There can be no one right answer.

    Do we measure downtown Chicago and call it all of the USA? All of Illinois? All of Chicago? Do we measure the sidewalk or the tarmac? The spot of snow near the sidewalk? The black stone or the white stone in the park? The tree or the dirt next to it? Every time you change the scale measured, you change the answer. This is a fundamental mathematical error in the very concept of a “global average temperature”. There is none and their can not be one; for the answer depends on the size ruler you use.

    By extension, trying to find the ‘trend’ of that number when you keep changing the size of the ruler each time you measure is a fools errand. Part of why I try not to indulge that past time too much…

    So at the very core of it, the question “How much have we warmed?” or “How much have we cooled?” is unanswerable. Yet clearly the planet does warm and cool…

    And that is why I prefer to look at large scale indicia of long term weather trends rather than trying to find some fictional number for our ‘average temperature’.

    Last year we had a lot of snow in both hemispheres. We had significant quantities of energy carried to space and radiated way in the making of that snow. (This touches on another problem, the difference between temperature and heat… Temperatures are like water pressure, while heat is like the quantity of water moved. A swimming pool has a lot of quantity, but no pressure, while a 2000 psi needle sized jet has a lot of pressure, but not much volume. We measure the temperature then try to treat it as a proxy for the volume (heat) and that just does not work. As water turns to ice or ice melts to water a great deal of heat moves, but the temperature does not change…) So all that snow means we’ve “cooled” a lot (in the sense of having lost a lot of heat) even if the temperatures (measured with whatever sized ruler…) have not dropped much.

    OK, that’s kind of a lot to pack in a small space, but the net-net of it all is pretty simple:

    We got about as warm in 1998 as we were in 1934. We’re now cooling and have headed back toward what it was like in 1974. We are heading to the cool side faster than we headed to the warm side (based on my experience in California, where I’ve lived for a few years shy of 60 years.) So my ‘guess’ would be that it’s about 1/4 of the way back to 1974 already. I’d make it about 2015-2020 to be back at that condition (having wiped out all “global warming” and headed back into a degree or so cooler than ‘average’ with a New Little Ice Age scare).

    Finally, The Sun:

    We’ve got a “sleepy sun” right now. Looks like it’s headed into a new Grand Minimum. (That, for purely puckish reasons having to do with Gilbert and Sullivan, I like to call a Major Minimum ;-)

    There is a great deal of controversy over the idea that this will cause more cooling. I just note that we are 5200 years from the last cycle of the warm to cold turn on that scale (when glaciers began advancing for about 5000 years…) AND that we are “due” for a cold going Bond Event. There are also some decent theoretical explanations for how a sleepy sun could make us quite cold.

    So we get to watch for the next 20 years or so and see just how cold our Major Minimum will get. I’m expecting a bit deeper than the last Little Ice Age based on looking at a 100,000 year chart of temperatures (see the AGW Climate Perspective category on the right hand edge) and seeing that the trend is slowly down and the ripples are at a peak about the same as the trend line connecting the older peaks.

    It will be interesting watching this to see what happens.

    Believe it or not, I desperately hope that CO2 as greenhouse gas is “true”. Why? Because it is the only real hope we have of avoiding a mile of ice on top of Canada and Sweden. If we are VERY lucky, we can preserve our Holocene Optimum like temperatures for the next few thousands of years. If we are not, and CO2 really does nothing, we are screwed. It’s just a question of when. (But the good news is that it will take a few thousand years to get really bad, even if the new ice age glacial is already underway…)

    So, a very long answer, but I hope it helps more than hinders. For now, enjoy the warmth. We’ve got a couple of decades of it left, I hope.

  13. mark albright says:

    I think I see what’s going on here. Both SJC sites are from the ASOS. The 1st one is rounded to whole degrees F, while the 2nd one is rounded to whole degrees C and reported at 10-min intervals. A temperature of 69 F, is likely to show up as 70 F when rounded to whole degrees C and then converted back to deg F. My metar decoder for 1 Oct gives the following hourly values in deg F starting at 11 AM (18 UTC) and going to 5 PM (00 UTC): 67, 69, 70, 72, 72, 72, 69. The 6-hr high temperature issued at 00 UTC was 73, 1 deg F warmer than the highest hourly value. I have refused to use the 10-min data so far since it is “incorrectly” rounded to the nearest whole deg C!

    This plot shows the 3 PM 1 Oct 2010 temperatures surrounding KSJC. Ignore the geographical outlines.

    The temperature appears accurate. SA142 is the Santa Clara site you compared to.

    While this one is probably fine there is a problem in San Diego (KSAN) right now with the ASOS.

    The past few days have seen minimum temperature records at San Diego like this one for Sunday morning 3 October 2010:

    RECORD EVENT REPORT
    NATIONAL WEATHER SERVICE SAN DIEGO CA
    525 PM PDT SUN OCT 3 2010
    …HIGHEST MINIMUM TEMPERATURE RECORDS BROKEN OR TIED ON OCTOBER 3…
    LOCATION NEW RECORD OLD RECORD PERIOD OF RECORD
    SAN DIEGO 69 TIED 69 IN 1997 SINCE 1875

    Here is the surface mesomap for Sunday morning at 12 UTC centered on KSAN:

    This shows one of the all-to-common warm anomalies at ASOS sites. In this case San Diego is indicating a temperature about 4 degrees warmer than nearby sites like the citizen weather observers in Mission Valley to the north of KSAN.

    It looks like this has been going on for about a week judging from the daily data shown below covering the past month. Sunday, 3 October, is shown as being +2 F above normal and yet the stratus barely broke open except for a short time in the late afternoon.

    STATION: SAN DIEGO/LINDBERGH FIELD CA
    YEAR: 2010
    ELEVATION: 13 ft
    LATITUDE: 32 43 N
    LONGITUDE: 117 10 W
    ===================================================================
    12Z AVG MX 2MIN PK 3SEC
    DY MON MIN MAX AVG DEP HDD CDD WTR SNW DPTH SPD SPD DIR SPD DIR
    ===================================================================
    4 Sep 62 74 68 -4 0 3 0 0 0 6 16 290 20 290
    5 Sep 60 70 65 -7 0 0 0 0 0 6 13 170 16 180
    6 Sep 60 66 63 -9 2 0 0 0 0 9 14 190 18 190
    7 Sep 62 69 66 -6 0 1 0 0 0 7 12 230 16 230
    8 Sep 64 70 67 -5 0 2 0 0 0 6 14 260 18 250
    9 Sep 62 70 66 -6 0 1 Tr 0 0 6 14 280 21 290
    10 Sep 60 72 66 -6 0 1 0 0 0 4 13 290 17 300
    11 Sep 61 71 66 -6 0 1 0 0 0 6 16 290 20 290
    12 Sep 60 70 65 -7 0 0 0 0 0 8 16 290 20 290
    13 Sep 62 74 68 -4 0 3 0 0 0 7 15 300 20 300
    14 Sep 64 75 70 -2 0 5 0 0 0 6 15 290 18 300
    15 Sep 63 74 68 -4 0 4 0 0 0 6 17 290 21 290
    16 Sep 62 72 67 -5 0 2 0 0 0 4 12 280 15 290
    17 Sep 61 72 66 -6 0 2 0 0 0 5 12 310 16 280
    18 Sep 62 72 67 -5 0 2 0 0 0 7 16 290 20 290
    19 Sep 63 72 68 -4 0 3 0 0 0 7 16 290 20 300
    20 Sep 63 71 67 -4 0 2 0 0 0 5 12 280 16 290
    21 Sep 61 66 64 -8 1 0 Tr 0 0 7 10 190 18 190
    22 Sep 63 71 67 -4 0 2 0 0 0 5 13 250 16 250
    23 Sep 62 73 68 -4 0 3 0 0 0 4 12 290 M M
    24 Sep 63 76 70 -2 0 5 0 0 0 4 12 310 14 300
    25 Sep 64 84 74 3 0 9 0 0 0 3 9 300 14 290
    26 Sep 67 89 78 7 0 13 0 0 0 4 13 290 17 310
    27 Sep 70 95 82 12 0 18 0 0 0 4 12 180 16 210
    28 Sep 73 82 78 8 0 13 0 0 0 4 16 170 21 170
    29 Sep 72 85 78 8 0 14 0 0 0 3 10 320 13 300
    30 Sep 69 78 74 4 0 9 .03 0 0 3 13 290 15 290
    1 Oct 70 80 75 5 0 10 0 0 0 3 10 220 13 220
    2 Oct 70 79 74 4 0 10 0 0 0 4 10 220 15 250
    3 Oct 69 75 72 2 0 7 0 0 0 4 10 270 14 300
    ===================================================================
    Temperature – Mean Min: 64.1 Mean Max: 74.9 Range: 60 to 95
    Temperature – Mean: 69.5 Dptr from Normal: -1.8 F
    Total Precip: 0.03 Total Snowfall: 0.0
    Mean Wind Speed: 5.2
    Total HDD: 3 Total CDD: 145

    Units: Temperature-deg F, Precipitation-inches, Wind-mph

  14. Steven Mosher says:

    I think you missed the point. The goal of an airtight analysis is to foresee and forestall all possible objections. Hence, the concern about altitude, however, slight you imagine it to be in this case, may not be so slight in others. Hence, the concern about wind and the characterization about that. hence the concern about the siting of the Santa Clara site.

    The point is this. Given the nature of your opponents, you would be well advised to anticipate every objection and eliminate them. That, I found, is a good practice. Like so, when a smarty pants says “what about the altitude difference you have already prepared for that by

    1. calculating an anomaly
    2. calculating the difference in altitude
    3. showing what that amounts to given standard lapse rates.

    That’s done with numbers and charts.

    My experience trying to convince government officials that they were wrong, suggests this is a fruitful way to proceed.

  15. E.M.Smith says:

    @Mark Albright: That would explain a few things… I’d rather not think that we had ANOTHER case of NASA getting screwed up because it could not keep metric vs traditional units straight though…

    @Steven Mosher: We have different goals. I didn’t miss your point, it’s just not my goal right now. My goal is to find where are “the big lumps” that are causing the errors. By definition when you are fast and rough ploughing you are not doing detailed finish work. If I slow down to do everything to 6 sigma precision then I will get far less done. Once the ‘rough pass’ is done, I can come back to the biggest issues and do more detail if desired. (And until then, other folks more interested in that type of work can take a topic and run with it…)

    From my ground school days (fixed wing private pilot) I remember the weather section quite well, including the use of lapse rate. I can also tell you that given the air flow at ground level here it simply does not matter. So I will not be wasting my time on a 44 foot lapse rate theoretical calculation when that’s just not the issue. As I said before, you are free to do so. IIRC it’s about 3 1/2 F / 1000 feet, so you are looking at something like 0.152 F of impact. Though to do it properly you need to allow for humidity too. Then look up the impact of turbulent flow on lapse rate and realize that it’s kind of silly at 44 feet in a maze of buildings, streets, cars, airplanes…

    But I am pretty darned sure that a 1/10 place F value is not the problem since they are off in whole degrees of F.

    Now I already knew that without “numbers and charts” because the right side of my brain does a great job of that kind of thing ( I have matching IQ scores for both sides and can swap between right and left side thinking at will) but I’m also well aware that most folks can’t and that many folks think the only way to think is left brain. But it isn’t. Look at the history of things done by left handed (typically right brained) folks and those few of us who are ambidexters. Then please stop telling us how to think. All it does is frustrate you and annoy us. ;-)

    (If the last line doesn’t ring a bell, I’ll post the line it comes from… having to do with singing…)

    Just to make it clear: I am not saying that the approach you would like to take is bad in any way. I am saying that I have different goals, a different process, and a different path. Like Johnny Appleseed, I’m not going to be picking and polishing all the fruit, but I’m looking for places to plant trees that others can tend. Someday when I come back to a given area, I may enjoy polishing a few apples (and eating a few)…

    And if I’m ever in a position of needing to convince government clerics of things, somebody will need to pay me for that level of torture. Not going to do it voluntarily. So until the day that paycheck shows up, I’ve got no interest in fixing those particular broken processes of thought. I’ll leave that to others more interested in it. (Yes, I’ve done it before. When paid enough to put up with the slow and wasteful process. When it’s on my dime, I’m going to employ my time much more efficiently.) Might I suggest that since you seem very interested in this, you take on that particular task?

  16. E.M.Smith says:

    From: http://www.weather.gov/om/brochures/asosbook.shtml

    TEMPERATURE/DEW-POINT: Each is reported in whole degrees Celsius using two digits; values are separated by a solidus(/); sub-zero values are prefixed with an M (minus).

    So they are reporting in C from the ASOS and it is in whole degrees. So there must be significant offset from the F degrees report, especially for stations reporting in 1/10 F or finer precision.

    Oh God, I’ve got that sinking feeling…

  17. Chuckles says:

    Might be useful –

    Click to access 71791.pdf

  18. E.M.Smith says:

    And, they “round up”…

    From: http://www.crh.noaa.gov/crh/?n=arp17-05a

    Figure 3. ASOS and standard equipment field locations. Area shown is enlarged from the rectangle outlined in upper right section of Figure 2.

    The methods used by ASOS and observers to measure maximum and minimum temperatures are somewhat similar. ASOS software uses an algorithm that samples the ambient air temperature nominally every 30 seconds and computes a one minute average based on this reading. It then averages five consecutive one minute values to compute a five minute ambient air temperature. This temperature is updated every minute. The highest and lowest five minute average temperatures of the day are stored as the maximum and the minimum temperatures, respectively. All values are rounded up (NOAA et al. 1992).

    Well, no problem then… just change to a system that measures using a nearly twice as large increment, then round up across the board. That will give you about a 1/2 increment jump. Then compare that with different instruments in the past and Ta Da!! You get 1/2 C of “Global Warming!!!”

    And a very nice fit to the 1 F of offset between the ASOS and other stations seen above. (And I wonder if they then apply a TOBS adjustment to that carefully computed maximum of the day?…)

    Where did I leave that single malt…

  19. There is no TOBS adjustment necessary since the ASOS creates a calendar day max/min temperature. This is one reason I like using ASOS daily max/min temperature data.

    There is a fundamental problem in the coding of metars which you have documented in the above comments. I believe it should be a requirement to transmit the data to tenths of a deg C precision. We are losing information because much of the metar data arrives in whole deg C precision. In the USA an additive group with temperature and dew point given to tenths of deg C precision is included for the primary hourly observations and also for the max/min temperatures. But the specials issued between hours don’t include the additive group, so they only report temperatures to whole deg C. To further complicate the issue, the 10-min ASOS data archived at NCDC appears to only have data to whole deg C precision, a situation which I find unacceptable.

    Cut and paste this lat/lon into google maps to see location of the ASOS at KSJC:
    37.35917,-121.92417
    It lies midway between the two runways over grass.
    I determined this using the guest login to MMS:
    https://mi3.ncdc.noaa.gov//mi3qry/login.cfm

  20. E.M.Smith says:

    @Mark Albright: Thanks for the info. Yeah, my TOBS comment was just some sniditude about putting a TOBS generic adjustment onto stations that don’t need it and was probably unwarranted. Then again, every time I’ve thought “They would never do THAT” I’ve discovered that they do it… so my wondering if they TOBS what doesn’t need it will remain, but it really is just a wee bit carping on my part…

    That location is also the most ‘downwind’ end of the runways and just about the point that most departures are dumping on the max fuel. (While on rare occasion we get backwards winds and they ‘turn the airport around’, it’s almost always winds off the bay and ‘launch’ from the downwind end.) So that ASOS is ideally situated to give how hot is is over the part of the runway where rotation and liftoff happen. Not so well for climate research…

  21. Summer 2010 was very cool along the California coast. To put this in historical perspective I have plotted the time series of summer (May-Aug) mean temperature back to 1964 for the Monterey CA climate station maintained by Prof Renard at his home SW of downtown Monterey. The graph shows a pronounced cooling trend since the mid 90s in summer temperature after a long warming trend over the previous 30 years:

    Here are the 5 coldest summers in Monterey since 1964, a span of 47 summers:

    2010 56.5
    1999 57.1
    1965 57.1
    1964 57.1
    2002 57.5

  22. E.M.Smith says:

    Yeah, near the coast has been very cold this year. Didn’t get any tomatoes to ripen until nearly August, and then only a few. We’ve got some now (thanks to the couple of warm days in September) but no new fruit set. Some years in California you can pick tomatoes all the way to November. Not this year…

Comments are closed.