Here is why the best paper on UHI will eventually be withdrawn in disgrace due to bad methodology as so often happens in climate science...
The best authors used the MODIS 500 meter dataset to identify what is and isn't an urban area. Here is an arial photo from 2013 with the actual MODIS overlay.
Every one of the grid cells would be classified as "not built up" because less than 50% of any of the grid cells is built up area. Imagine, an airport is not an area where UHI might influence temperature readings...and you can bet that there is a temperature gathering station found at that airport.
Using the best methodology, the photo below, of the same area shows exactly how much of the pictured area could be covered with houses, roads, parking lots, etc and still not be classified as built up.
Their methodology is obviously and terribly flawed which is all to common in climate science. Perhaps that is because actual scientists familiar with the scientific method are quite rare in the field of climate science which seems to be topheavy with political activists playing scientist.
Little wonder that they found little evidence of UHI since using their methodology, they might have a hard time finding actual urban areas. The photo below is a modern (2013) view of the same area and is still classified as rural and not built up....the red squares come the closest to being classified as urban because 40% of the area is covered.
Here is a photo of the weather station found on the grounds of the airport.
And here is an example what happens to that station over 100 times a day..
Imagine...using the best method, no evidence of a heat island effect would be found anywhere near that airport.
The whole explanation of the terribly flawed methodology can be found
HERE