How not to measure temperature, part 84: "Pristine" Mohonk Lake USHCN station revisited

As WUWT readers may recall back in September of 2008, the New York Times ran an extensive first hand account of the Mohonk Lake, NY USHCN climate station of record. The Mohonk article was covered by WUWT guest contributors Dee Norris here and John Goetz here. Goetz shows that even the “pristine” station data gets adjusted by NASA GISS in their GISTEMP program.

NYT’s Anthony DePalma gave a very laudable account of the work done by the observers at the Mohonk Lake station:

Every day for the last 112 years, people have trekked up the same gray outcropping to dutifully record temperatures and weather conditions. In the process, they have compiled a remarkable data collection that has become a climatological treasure chest.

The problems that often haunt other weather records — the station is moved, buildings are constructed nearby or observers record data inconsistently — have not arisen here because so much of this place has been frozen in time. The weather has been taken in exactly the same place, in precisely the same way, by just a handful of the same dedicated people since Grover Cleveland was president.

DePalma also writes:

The record shows that on this ridge in the Shawangunk Mountains, about 20 miles south of the better-known Catskills, the average annual temperature has risen 2.7 degrees in 112 years. Of the top 10 warmest years in that time, 7 have come since 1990.

WUWT contributor Denise Norris looked into the temperature record at Mohonk and writes:

Here is what I found:

Since I live in the general area, I have previously used the data from a site in nearby Maryland, NY (42.52N, 74.97W; 363m) in a local lecture.  I was sure I remembered that the station in Maryland had not exhibited a trend like this.  Double checking the ol’grey matter, I got this graph:

Both sites are at the same altitude and in the same general vicinity.  I know that climate change can’t be that localized, so it has to be something else.

I should point out that the graphs Denise posted above use the RAW unadjusted USHCN data, (from CO2science.org ) rather than the adjusted GISTEMP record so many are familiar with. In the GISTEMP record, Maryland, NY is shown to be nearly flat while Mohonk seems to be the same.

My experience and the experience of many others in surveying the USHCN network of climate monitoring stations has shown that time and again, when such a significant divergence is seen with two nearby stations, localized siting issues and other microsite biases are often found at the weather station. On January 4th and 5th, 2009, WUWT contributor Evan Jones set out to find the answer. Due to other projects that required my attention, it has taken me until now to find the time to write this article.

First and foremost I should point out that I do not disagree with the NYT’s characterization of this station as being valuable due to its lack of station moves and long and meticulously recorded record of temperature and precipitation. The staff at Mohonk should be commended for their long and dedicated service. It is not my desire to impugn that in any way. However, it should be noted that what Evan Jones found at Mohonk mirrors some of the slipshod quality control that is under the purview of NOAA as the entity responsible for the equipment at COOP and USHCN stations around the USA. The responsibility for the quality control failures that I am about to highlight lie solely with NOAA, and not with the staff of Mohonk. It is NOAA’s responsibility to ensure that these stations adhere to specifications and to perform regular quality control checks.First some perspective as to where the Cotton Region Shelter (Stevenson Screen) and the Standard Rain Gauge are located on the grounds. Evan provided this annotated map:

The station’s Cotton Region Shelter (CRS)  is located on the public grounds, just a few feet off of a public path. On first approach, a couple of obvious problems appear right away.

Upon closer inspection we see the CRS is directly next to a tree stump.

And, the CRS is rather close to a building and chimney.

Looking at the photograph, here’s a few questions that come to mind. The CRS is located just a few feet from a building with an active chimney. The building is L shaped, and thus acts as a wind block for at least two directions. If the building was there 100 years ago was the chimney there then? When was central heating and A/C added? How much of the waste heat of the building is imparted to the thermometers by prevailing winds? Evan Jones reports in his site survey (PDF) that the “CRS is 5.8 m. from building to South”. That would make it a class4 station (considered as unacceptable siting) by the current criteria used to evaluate siting (PDF) for the NOAA Climate Reference Network. It also violates NOAA’s longstanding 100 foot and distance to 4 times height rules as documented in the NOAA/NWS COOP Observers Handbook (PDF available here) and on the COOP web page here.

The CRS also has brush around it, and most notably has a large tree stump directly adjacent to it. Obviously at one time, there was a significantly large tree there that was cut down. There is also a large tree limb missing from the tree to the right. The CRS thermometers obviously recorded temperatures in the shade of these tree canopies. Seems’ like a good idea right?

For those that don’t know, measuring temperature under a tree canopy at the base is one of the worst places to do so for a long term record. Here is why:

• The shade of the tree canopy will suppress the Tmax in the summer due to decreased solar insolation in the local area
• The lack of a shade canopy will enhance the Tmax in the winter due to increased solar insolation in the local area
• The shade of the tree canopy will elevate the Tmin in the summer, due to LW infrared being reflected back to the ground by the leaves, resulting in a  localized warming.
• The lack of  shade canopy will  the lower the Tmin in winter due to increased LW infrared emissions (as compared to summer when the canopy reflects LWIR).
• The tree grows, and  the effects grow with it over the long term. They don’t end until the tree is trimmed, dies, or is cut down.
• The growing tree trunk gradually blocks air flow at the rear of the screen and is a mild heat sink.

Then of course there’s the state of the bushes. Trimmed regularly, overgrown and then cutback? Were they always there? when was the tree trimmed?

It appears some recent trimming done, perhaps due to a storm taking down a branch:

Here’s another view:

Some of these problems associated with trees and placement (specific to the USHCN network) are documented in the paper published in March 2005 in the International Journal of Climatology titled: The GeoProfile metadata, exposure of instruments, and measurement bias in climatic record revisited by Rezaul Mahmood , Stuart A. Foster, David Logan. Abstract available here.

But wait, there’s more.

In this picture below, notice the base of the CRS, made of logs, possibly from the tree that was cut down. Other than the fact that shade from the building is hitting the CRS, notice anything peculiar?

There is a standard height that the CRS must be off the ground. In this case the homemade log legs are shorter than normal, and according to measurements made by Evan Jones,  the base of the CRS is only 29 inches from the ground surface. Normally it would be 48 inches. You can read about the design criteria in this CRS manual (PDF) by a company, Novalynx, that supplies them. There is also a reference at NOAA here in FAQ#17 that states the wooden leg length to be 48″.

They write:  The legs are designed according to the original criteria to give the shelter a height above the ground of 48 inches.

The lowering of the screen height will create a warm bias, and since it is likely that these were not always the legs for this shelter (which appears relatively new) there will be a slight step function upwards in the temperature record when this occurred. Why NOAA allowed this I have no idea.

Here’s another view of the legs and the ground underneath.

Other than the photos seen above, there is only one other publicly available photograph of the CRS at Mohonk Lake. I’ve shown it below. The present curator, Mr. Paul Huth is standing near the CRS preparing to open the door.

Source: http://www.mohonkpreserve.org/index.php?weatherdata

I had always wondered what the dark material under and around the CRS was. It appears to be mulch and/or wood chips. Certainly such a ground cover with a dark albedo is not a good thing near a weather station.

When Evan was there, there was snow cover. But he found the ground cover by the CRS to be a dark crushed rock, similar to “road base”, shown below.

He photographed it in his hotel room at the resort and returned it to the site.

Crushed rock is not a good surface cover near a weather station as it will absorb more solar radiation on sunny days, raising Tmax and the rock will radiate LWIR at night raising Tmin. How long has the crushed rock been there? Was it always that way, or was it grass or some other surface cover in the past?  The crushed rock base certainly doesn’t fit the “representative of the area” specification for siting a Cotton Region Shelter in the NOAA/NWS COOP Observers Handbook (PDF available here).

3.1 Shelter Placement. The ground over which the shelter is located should be typical of the surrounding area. A level, open clearing is desirable so the thermometers are freely ventilated by the flow of air. Do not install on a steep slope or in a sheltered hollow unless it is typical of the area, or unless data from that type of topographic location is desired. When possible, the shelter should be no closer than four times the estimated height of any obstruction (tree, fence, building, etc.). Optimally it should be at least 100 feet from any paved or concrete surface. Under no circumstances should a shelter be placed on the roof of a building as this may result in extreme temperature biases.

In fact it appears that little about the Mohonk Lake USHCN station fits that criteria. If you look at the photograph that NOAA supplies with their handbook for section 3.1, the differences are obvious. Note the grass under the shelter and lack of nearby obstructions.

Finally we have the rain gauge, at the end of the public dock protruding into the lake. In such an open public area where fishing, swimming, boating, and picnic coolers abound, could it be assured that somebody didn’t pour their coke, coffee, or take a whiz in the gauge? Would the Mohonk staff catch it if somebody did, or would they just dutifully record the level on the dipstick, write it down, and empty the gauge into the lake?

Here is the closeup photo Evan took of the rain gauge at the end of the dock:

I found this photo taken by a tourist and placed  on Flicker photo service:

And since there was a 10+ megapixel version of the photo available, I was able to isolate and identify the rain gauge using Evan’s photo as a guide.

Going back to NOAA’s COOP website I found this criteria for rain gauge siting:

Precipitation gauge siting: The exposure of a rain gauge is very important for obtaining accurate measurements. Gauges should not be located close to isolated obstructions such as trees and buildings, which may deflect precipitation due to erratic turbulence. To avoid wind and resulting turbulence problems, do not locate gauges in wide-open spaces or on elevated sites, such as the tops of buildings. The best site for a gauge is one in which it is protected in all directions, such as in an opening in a grove of trees. The height of the protection should not exceed twice its distance from the gauge. As a general rule, the windier the gauge location is, the greater the precipitation error will be.

Looking at the boats stacked there, it made me wonder how they might affect the rain gauge with wind turbulence. I did some additional digging and found this photo, also from a tourist on another photo sharing site called Webshots.

I wonder what the boats stacked around the rain gauge does to the precipitation record?

Clearly the Mohonk station has a value in it’s long and uninterrupted record. But it pains me to say this, as I always feel bad for the observers when NOAA doesn’t step up and help them with quality control, but the value of the Mohonk Lake record has been tarnished due to the significant number of local site issues related to placement and maintenance of the thermometer and the rain gauge. When you look closely with NOAA’s own criteria (something the NYT reporter didn’t do) you find that the station isn’t as good as the story leads us to believe. Even NOAA’s spokesman quoted in the article made no mention of these issues:

“The quality of their observations is second to none on a number of counts,” said Raymond G. O’Keefe, a meteorologist at the National Weather Service office in Albany. “They’re very precise, they keep great records and they’ve done it for a very long time.”

But I wonder, has Mr. O’Keefe ever visited the station, or is he just assuming the data quality is there because the Mohonk staff does such a good job with the monthly B91 form they send in? I agree, they do a great job of record keeping. For example here’s the December 2008 B91 form completely filled out and with extra detail.

mohonk_lake_b91 (PDF) Source: NCDC http://www7.ncdc.noaa.gov/IPS/coop/coop.html

December is normally a problem month for records, yet to their credit, they have every day.

But even with such attention to detail in record keeping, how valuable is such a record when there appears to be no attention to detail in the exposure and placement of the instruments gathering the data?

When scientists look for the “global warming signal” of a few tenths of degrees, how would they know if the data they are getting is accurate if they don’t know the provenance of the station quality? Even if they did know the things we know now about Mohonk Lake, how would they go about sorting out the local site biases that have cropped up over time and apply a correction?

Then there’s the resort itself. The NYT article  claims it “has hardly changed over time”.

“Mohonk’s data stands apart from that of most other cooperative weather observers in other respects as well. The station has never been moved, and the resort, along with the area immediately surrounding the box, has hardly changed over time.”

Yet if we were to examine things like: when was air conditioning installed for guest rooms? When were the grounds walkways and pathways switched from dirt or stone paths to concrete or asphalt? How much has bus and automobile traffic increased in the last 100 years? How many additional outbuildings have been built? How much of the ground cover, trees, shrubbery, and landscaping has changed? I wonder by what basis the NYT reported made that sentence? Given how much energy is used for heating/cooling cooking, laundry, landscaping and trasnportation, surely the resort has it’s own localized UHI signature compared to surrounding areas.

A tourist resort is hardly a static place. They are always being revised and improved to keep the customer experience fresh and rewarding, so I have a little trouble believing that Mohonk Lake “has hardly changed over time”, especially since the NYT reporter missed everything else about the station’s measurement environment.

I feel for the staff at Mohonk, and I have contempt for NOAA and the NWS for allowing this station to escape basic quality control.