“It was estimated that up to 89% of the stations were poorly located & were providing mediocre data results.”
You can go further than that. By now it’s up to around 91%. (Also that lovely location in Fairmont, CA, turns out to be — ugh — mounted on concrete!
Oddly enough, the wastewater plants have lower trends (the mass amount of waste heat masks the trends). But urban sites definitely have higher trends than rural. And, ironically, it turns out that urban sites are better placed (re. microsite issues), on average, than rural sites.
There’s an equipment bias as well. MMTS readings are a lot different from Stevenson screens (CRS). (And ASOS setups, typical of airports, have much higher trends than either MMTS or CRS.)
Not only that, but most sites have seen several station moves apiece over the decades, and that totally throws things out of whack.
BTW, I know what I’m saying. I personally made over 200 of those surveys (either direct or via satellite), supplemented another hundred, and evaluated the whole (1000+) shebang! it is a worms’ nest. A squirmy one.
And, yes, trends are affected by bad siting as well as absolute temperatures. Badly sited stations tend to run hotter faster in warming trends (and colder faster during cooling trends, as the effect “undoes” itself). And yes, I know what I’m saying here, too: I’ve run all those raw data trend numbers for all 1221 USHCN stations, as well, by both site quality and equipment type.
I would conclude that there has been warming, but it has been exaggerated. A straight 20th Century average of NOAA USHCN trends (to 2006, ungridded), is +0.14C/century for raw data — and +0.59C/century for “adjusted” data.
