Mar 05 2008
Crater Lake Surface Station Looks Okay
I have an interesting case for you to look at with your method.
in H2001 Hansen mentions that Crater Lake station was deleted in it’s entirety. I spent a couple days looking at it and the surrounding stations and nothing struck me as being odd about this station. This was an utterly subjective assesment so it might be interested to see what you method suggests.
I’ve been extremely busy lately, so I decided to reply in a post so that it looks like I’m not abandoning the blog.
The short answer is not really. Nothing looks out of the ordinary for this station. Keep in mind I haven’t fixed the problem with my code yet. As I said before, this should not affect these results, but who knows. I won’t post any raw results until I get around to fixing it because I know those are affected.
|
The red line before the start of the temperature data is because the smoothing function doesn’t properly handle NaN values. That’s another thing that needs to be added to list o’ things to do. Because it’s easy to fix and I actually use it in my research, it gets bumped near the top.
There is a much larger season cycle seen in this analysis though. This could possibly be due to the non-standard height of the measurements during the winter, which was noted in the surface station survey. It’s also noted that this setup will bias the summer measurements upward. No mention is made of it biasing the winter temperatures downward.
The “negative slope” shows that this station is warming slower (or cooling faster) than the surrounding stations. I won’t speculate what that might be.
Because of the way the temperature is measured here in the winter, it makes sense to just remove it from any analysis.
(No Ratings Yet)
Loading ...Related Posts:
16 Responses to “Crater Lake Surface Station Looks Okay”
To reduce spam, comments are automatically closed 30 days after the last comment. If you would like to comment on any closed thread, please use the contact form at the top of this page.
Thanks, my suspicion before seeing the site survey was the great amount of snow they got there could be a factor in measurement anomalies, but i could find anything wildly divergent.
“Because of the way the temperature is measured here in the winter, it … ”
But isn’t that exactly why micro-site issues should be investigated?
Dan, bearing in mind the difference between “investigation” and “leaping to conclusions,” has anyone ever seriously argued that these issues should not be investigated at all? Indeed, that appears to be precisely what Atmoz is doing right now.
Craig,
Yes, there have been and remain to this day people, including scientists and climatologists, who will argue that such investigations are not necessary.
I guess you don’t read RealClimate, Eli Rabbet, tamino, among others. Google and take a look at the abuse that ClimateAudit, Watt’s up With That, SurfaceStations and others have had to put up with. It’s not a pretty picture.
Dan Hughes,
I read those blogs too, but I have pretty much the opposite conclusion. Eli Rabbet in particular has argued until he must be blue in the face that photos of sites, interesting as they may be, do not replace comprehesive metadata that spans time. I have never heard RC, Eli, or Tamino/”HB” argue that siting is irrelevant. I’m kind of amazed at your impression.
The photos are not _necessarily_ irrelevant, but by no means are they any kind of thorough investigation of siting issues, especially not for historical temp readings. On their own, the photos – even if there were a picture for all stations – can just as easily be misleading as corrective, and may add no info whatsoever for analysis of climate.
Title: Crator should be Crater
Not nitpicking, just trying to help.
Just a note on the “surface station looks okay” part, again not nitpicking but trying to be helpful.
Of the 502 stations I’ve looked at, this is the only one where the Stevenson Screen had these factors:
1) It is about 2 feet off the ground during the summer, making for a warmer measurement environment due to earth proximity. See this:
http://gallery.surfacestations.org/main.php?g2_itemId=32041
and
http://gallery.surfacestations.org/main.php?g2_itemId=32047
2) It changes height with the snow, being mounted on a movable track on a tower.
http://gallery.surfacestations.org/main.php?g2_itemId=32077
I’m not sure how one would go about adjusting for situations like this since one never would know what the observing height is.
The methodology was designed to look for “typical” station moves; that is, changes that occur once at a specific (unknown?) time such that there would be a step function (either up or down) when compared to surrounding stations. The changes in height that occur at this station are likely going to vary nearly with the yearly cycle, and tend to have a higher variance in the temperatures. Summer months would be warmer than surrounding stations, and winter months colder. I did attempt to apply a very simple correction for this by adjusting the variance of regional average to equal the variance of this station. The method produced nothing useful, so I didn’t originally mention it.
I too am unsure how to correct for the observing height, but that does not mean it cannot be done.
Theres a new solar variability article out in physics today: Scafetta & West. I’d be really interested to hear what you think.
Hi Atmoz.
Did you not see my post above about the title being mispelled?
[Reply: Nope. I missed that part. Thanks for catching the error.]
thank you ian. Exactly what Eli has been trying to say
Ian said, “On their own, the photos – even if there were a picture for all stations – can just as easily be misleading as corrective, and may add no info whatsoever for analysis of climate.”
Can you provide a single example for which knowledge of the exact conditions under which measurements of data are made has led to ‘misleading’ conclusions?
Additionally, can you point me to a single post/comment by either of tamino, Eli ‘Who’s He’ Rabett, GS or any other GISS/NASA employee, in which they have been supportive of collection of station meta-data?
TIA
Hi Dan,
Perhaps I should start by stating that I definitely agree that having station meta-data is important, and that it’s good to have thorough meta-data. I don’t think anyone disagrees that meta-data are desirable in interpreting measurements from weather stations.
What I’m objecting to is the notion that siting photos can provide “knowledge of the exact conditions under which measurements of data are made,” to use your words. The photos might provide _some_ useful information, but even the best set of photos taken on a single site visit can’t tell us the “exact conditions.” I think that this is also the objection of the hosts of Eli Rabbet, Real Climate, Open Mind, etc.
For example, suppose I take photos of a site where there is an air conditioner within 5 feet of the instruments (not that distance is marked precisely on the photos, but let’s say that the photographer measured it). Does the location of the air conditioner bias the temperature recordings? Many people assume that it will create a warm bias, but this is a question that we can’t answer from the photos alone – and in fact, we don’t even know in which direction the ac unit might bias temperature readings. It might be that hot air from the ac unit blows over the temperature sensor, and makes it read artificially high, if the ac is on when the temp measurement is taken. But we can’t tell from the photos if this is really happening. Or, it might be that the ac unit happens to put the temp sensor in shade at the moment of measurement, creating spurious cooling. Again, we can’t tell from the photos. Perhaps the updraft from the hot ac draws air in from the irrigated lawn nearby, which blows over the temp sensor at the time of measurement, again creating spurious cooling. Etc., etc. Or, maybe the ac has no impact whatsoever on the temp sensor. Then, aside from noting the ac unit, you’d also want to note all other potential influences in the photos – tree and building shade, whether someone regularly burns trash under the sensor when you’re not there to catch him with a photo, etc. These possibilities are all equally plausible if the only metadata you have is the photos.
You get my point – these are all valuable pieces of information in analyzing the station’s data, but the photos can’t inform us about any of these possibilities. A set of photos, especially when taken at a single point in time, can only give us a bit of metadata, and the photos can’t tell us about other, equally important kinds of metadata that are often required before drawing conclusions about biased measurement. This is why a statistical procedure like jackknifing a large data set of measurements can be a better way to spot measurement error and bias than comparing the stations records with photos taken on a single day.
Creating an archive of photos of the sites is a nice idea, and my admiration to the commitment of the photographers. But drawing unwarranted conclusions about decades-long measurement bias from the photos alone is over-reaching the data. Making an archive of site photos in conjunction with collecting information that helps quantify a measurement bias would be a huge improvement over the photos alone.
Tamino says the same thing in comments to his “surface station” post:
http://tamino.wordpress.com/2007/07/30/surface-stations/
To paraphrase Tamino, he thinks attempts at improving the SS record are worthwhile but he just doubts the competency and objectivity of some of the individuals doing it.
Well you could try this and this or as I put it over at Climate Audit (you can google it they do enough business)
“You might go read the station history.”
and
“A quick addition. A single photo is a point in time. Since climate measurements are made as differences between local averages, what really would be needed for such an approach to work would be a series of photos. It is not that Anthony’s project is useless, it is that for climate science purposes it would not be very useful, of course for political purposes it would be very useful.”
Eli’s links are well worth (re)visiting - real-life examples of crucial meta-data that photos alone won’t catch: a perfect-looking station with no instruments inside the screen, a delapidated screen that doesn’t actually house any instruments, a history of location and measurement time changes over decades, etc. (The schoolboys and the rain guage is truly a classic!) And he throws in cites to the lit and GISS’s urban adjustment procedures, all at no extra charge…