More observational evidence emerges showing CO2’s effects are too small to drive climate.

The oft-cited Feldman et al. (2015) paper proposes to isolate the CO2 impact by only considering a “clear-sky” atmosphere at two sites: the North Slope of Alaska and the Southern Great Plains (USA). In other words, that study can only claim to assess what occurs in an imaginary atmosphere where clouds do not exist – even though clouds are present about 60-70% of the time in the real-world atmosphere.

Results indicated that CO2 enhances the downwelling longwave radiative (DLRF) forcing trend, the greenhouse effect, by just 0.2 W/m² per decade (accompanying a 22 ppm CO2 increase from 2000 to 2010).

Image Source: Feldman et al., 2015

But a 2025 paper quietly published in Nature also assessed the contributors to the DLRF trend from 1996-2018 at the Southern Great Plains, even extending their measurements to all land surfaces across the globe from 60°S to 60°N.

But instead of employing only a simulated atmosphere where clouds do not exist, Liu and colleagues provided the all-sky values. So, unlike the Feldman paper, they employed the real-world atmosphere which have clouds contributing to (and dominating) the DLRF trend.

They found observed trends in downwelling longwave show the cloud DLRF contribution actually serves to reduce the greenhouse effect impact by -0.77 to -1.77 W/m² per decade.

Since the CO2 contribution enhances the trend by only 0.2 W/m² per decade, clouds therefore have been observed to easily outclass and even nullify (net) the effect of CO2 increases.

The tiny “finger printed” (FP) decadal effect of CO2 in a real-world, all-sky atmosphere is illustrated (red bar) in the DLRF trend (W/m² per decade) graph. Notice how insignificant the CO2 contribution is compared to the effect of clouds (gray bar).

Perhaps this is why Feldman and colleagues chose to use an imaginary-world, clear-sky-only atmosphere instead of a real-world, all-sky atmosphere.

Image Source: Liu et al. 2025