The flaw in the Colorado average temperature report

A Jan. 8 article about a new CSU climate report showed it contends that Colorado’s annual average temperatures increased by 2.3 F between 1980 and 2022. So how did they determine that?

Appendix A of the report states that NOAA’S latest nclimgrid dataset is the source of its observations. The nclimgrid dataset is a gridded dataset of 5-kilometer by 5-kilometer squares. NOAA uses observations from the Global Historical Climatology Network (GHCN) to fill in the squares. The area of Colorado is about 269,600 square kilometers, so there are about 10,800 such squares in the state. But there are only about 4,000 GHCN observation stations in Colorado. So the 4,000 GHCN observations are spatially interpolated into NOAA’S 10,800 data squares.

One notices an immediate problem with this process, the data squares are evenly distributed across the state but the GHCN stations are not. The stations are mostly clustered along the Front Range and other population centers. The interpolation process is further complicated by Colorado’s mountainous terrain over much of the state which results in significant temperature variations over very short distances.

But probably the most serious difficulty in determining the actual temperature change from 1980 to 2022 is the urbanization that has occurred due to the doubling of Colorado’s population during that time. Urbanization results in what have been termed “urban heat islands” which raise the observed temperatures in those areas.

A 2020 UNLV study “The Urban Heat Island Effect in Nevada” found those increases can be significant, ranging from 4.9 F in Denver to 5.9 F in Albuquerque to 7.3 F in Las Vegas. When all of the artificially inflated 2022 temperatures of stations in the new UHI areas formed since 1980 are interpolated into the gridded dataset that alone could account for much of the contended 2.3 F increase.