Coyote Blog

Coyote Blog: I won't declare this to be a media trend, but this summer we were plagued with news reports about the drought "trend" when in fact no such trend exists in the US data (NOAA data from this article).

Averaging across the entire contiguous U.S. camouflages the actual trend.

"As Earth warms, the water cycle intensifies. While this means more total rainfall for some regions, it also means that more rainfall comes in the form of heavy and extreme events and periods between rainfall events become longer. Longer periods without rain and higher temperatures lead to losses of soil moisture; if drying of the soil continues for long enough, droughts occur. Since the 1950′s, some areas of the United States, including most of the Northeast and Midwest, have been experiencing drought conditions less frequently. Other areas, including most of the West and Southeast, have been experiencing drought conditions more often."
http://www.earthgauge.net/2009/climate-fact-drought-trends

The trends for almost all European countries look like this. A very few (like Greece) show some reduction of federal spending, but nothing even close to "austerity". Of course, the US is no better...

@Zachriel: Unfortunately, the links you point to are pretty useless. For example, the "hours per day over 100 degrees in Phoenix" makes no attempt to account for UHI. Phoenix is a lot bigger that it was 50 years ago, with a lot more asphalt. This very likely accounts for the entirety of the trend shown.

We'll have to agree to disagree.

From my perspective, the analysis is lacking. For example, yes, the graph says UHI - but makes to attempt at analysis, i.e., to say how much of the effect this causes. If it is the entire effect (entirely possible, since the trend is actually rather small), the graph is simply meaningless.

Meanwhile, Berkeley Earth says that UHI is no factor at all, which defies belief given the large number of weather stations in urban areas or at airports. The Berkeley analysis appears weak, because they are evaluating data that has already been "corrected" by the various research groups. The correction procedures themselves have been called into question. For example, some apparently perfectly sited stations are often corrected by lowering older temperatures and raising recent temperatures. The original raw data is generally not available; the correction procedures are not explained; the Berkeley makes no attempt to address these problems.

Zachriel:
Sorry to bust your bubble, but usually communication is difficult in climate discussions, and I do not think you have done enough critical examination of The Best project. You quote Muller as saying that the Best Project "went back to the raw data and used only that." Such a move would be amazing. Did they really use raw data before TOBS adjustment? Raw data in the United States shows a flat trends before TOBS and a couple of minor adjustments. I have not read anyplace that Best did an alternate adjustment for TOBS, and that would be an amazing move to do. Also, Muller is absolutely wrong to say that "examining only the data from rural stations" is "a clever and original way." Other have done it previously, and have shown that much of the rise is due to urban influences -- it is not the same answer.

Zachriel: Although I have not talked with Dr. Muller personally, I have read his description of his methodology. I have not found where he says that he uses pre-TOBS data. In fact, his reference to TOBS is this: reject the TOBS value when it falls outside of the T-max and T- min. That seems to an indication that post-TOBS data is used. Also, look at BEST's estimate of warming in the United States. If the raw data is flat, and BEST says 2.32 degrees of warming, that would seem further to suggest that BEST used post-TOBS data.
I won the bet with my spouse. I bet that you would bring up the spurious point that the US has only 2% of the earth's surface. Note that this discussion started with discussion of trends in the U.S. Also, it appears to be generally understood among those who have thought about the issue that the level of quality is higher in the U.S.
It is not only older studies but also newer studies that show a signficant influence of UHI. Six years ago, I took the data from a half-dozen rural stations about whose quality I was confident -- and added sites in Antarctica and the Arctic where UHI did not appear to be the factor. The result: no significant trend. I did not do anything with this study because of the obvious problems in data selection. But various other studies (more comprehensive) have been done, such as pairing rural and urban stations. Lat month, a paper was released on U.S. surface stations: http://wattsupwiththat.files.wordpress.com/2012/07/watts_et_al_2012-press-release-r1.pdf showing that surface trends have been signficantly overstated due to measurement issues.
In thinking about UHI and siting issues, it is important to recognize that UHI is not limited to urban areas. Rural areas also see build up of concrete and steel. Thermometers are often stationed at airports where concrete and increased exhaust from increased plane traffic can also affect results. Also, a couple of decades ago, a shift started in thermometers -- before people actually went outside to read the values but the new system included cables connecting new thermometers to inside displays. In the change-over, many thermometers were relocated closer to the building where they were more exposed to heat from the building and HVAC exhaust. In addition, some of the classifications as rural stations in ROW have been downright ludicrous. These are just some of the reasons why BEST might not have detected an influence of UHI.
Yet, in closing I want to affirm my belief that there has been an increase in temperature in the last 200 years. Such a trend is confirmed by more than thermometer readings. Yet two unresolved issues remain -- (1) how much of that trend is due to CO2 trends and (2) why has the trend been limited to T-min, and while T-max has been steady. CAGW theories and non-CAGW theories offer different explanations, but no one theory has won my heart.
BTW: increases in T-min has obvious advantages but there also are potential disadvantages.

David: I have not found where he says that he uses pre-TOBS data.

Perhaps this will help:

"We tested the method by applying it to the GHCN dataset created by the NOAA group, using the raw data without the homogenization procedures that were applied by NOAA (which included 39 adjustments for documented station moves, instrument changes, time of measurement bias, and urban heat island effects, for station moves). Instead, we simply cut the record at time series gaps and places that suggested shifts in the mean level. Nevertheless, the results that we obtained were very close to those obtained by prior groups, who used the same or similar data and full homogenization procedures."

Note that they compared their statistical 'scalpel' to conventional homogenization and they matched closely.