[LINK TO THE HOME PAGE OF THIS SITE]

[LINK TO THE CLOUD PAGE OF THE CLIMATE-FOR-YOU BLOG]

THIS POST IS A CRITICAL REVIEW OF THE KAUPPINEN & MALMI 2019 HYPOTHESIS  [LINK] THAT THE THEORY OF AGW CLIMATE CHANGE HAS FALSELY ATTRIBUTED WARMING TO RISING ATMOSPHERIC CO2 CONCENTRATION WHEN MOST OF THE WARMING CAN BE EXPLAINED IN TERMS OF ALBEDO LOSS FROM A DECLINE IN LOW CLOUD COVER. 

FIGURE 1: LOW CLOUD COVER 1983-2008 [SOURCE]

FIGURE 2: HIGH CLOUD COVER 1983-2005 [SOURCE] 

FIGURE 3: TROPICAL LOW CLOUD & TEMPERATURE 1983-2011 [SOURCE] 

FIGURE 4A: GLOBAL WARMING ABOVE & BELOW LOW CLOUDS 1979-2019

FIGURE 4B: CORRESPONDENCE BETWEEN THESE TWO TEMPERATURE SOURCES

FIGURE 5: CORRELATIONS: CO2 FORCING AND TEMPERATURE 1979-2019 

FIGURE 6: KAUPPINEN & MALMI FIGURES 2&3

1. The Kauppinen & Malmi 2019 paper (KM2019) [LINK] with the provocative title “NO EXPERIMENTAL EVIDENCE FOR THE SIGNIFICANT ANTHROPOGENIC CLIMATE CHANGE” (SIC), uses recently published satellite data for low cloud cover (Figure 1) along with HadCRUT4 global surface temperature data to display a visual correlation between temperature and low cloud cover (Figure 6).  Obvious departures from the correlation are  explained in terms of the 1991 Mt Pinatubo eruption (temperature too low) and the strong 1998 El Nino (temperature too high) and concludes that declining low cloud cover, and not rising atmospheric CO2 concentration, explains global warming.
2. Based on the charts of overlaid temperature and low cloud cover data shown in Figure 6 (See Figure 2 and Figure 3 in the source document [LINK] ), the authors claim that the observed warming trend in surface temperature is explained by declining low cloud cover and not by rising atmospheric CO2. The rationale for the low cloud effect is that clouds reflect incoming solar radiation and that therefore rising cloud cover causes cooling and declining cloud cover causes warming. The decline in low cloud cover in the study period 1983 to 2008 is well established in the ISSCP datasets [LINK] and is displayed above in Figure 1. A relationship between low cloud cover over and temperature over the tropics is presented in the Climate4You blog [SOURCE] and is displayed in Figure 3 above. It shows rising temperature in the tropics coincident with declining low cloud cover and appears to be consistent with the KM2019 finding that the declining low cloud cover (LCC) causes warming globally and not just in the tropics. The causation rationale is that clouds reflect incoming solar radiation and thus lower surface temperature.
3. It is noted however, that correlation between time series source data do not always imply causation and that when they do, no information about the direction of the causation can be inferred from these data. For example, in {McCoy, Daniel T., et al. “The change in low cloud cover in a warmed climate inferred from AIRS, MODIS, and ERA-Interim.” Journal of Climate 30.10 (2017): 3609-3620} the authors find that warming surface temperatures have caused a decline in low cloud cover over sub-tropical regions in the same period of study as the KM2019 paper. The reverse causation is suggested in the Somerville 1985 paper in which he finds that CO2 induced global warming is self correcting because warming increases cloud formation and clouds reflect sunlight back into space. {Richard Somerville, Scripps Institute of Oceanography, UC San Diego}. 
4. An added complication is that in the instrumental record, global warming is found mostly in nighttime daily TMIN and not in daytime daily TMAX  {Related posts [LINK] [LINK] }. G. Kukla, PD Jones, and others (Kukla 1993) describe this apparent anomaly in terms of low cloud cover that reflects solar radiation upward and the earth’s long wave radiation downward. At night, with no solar radiation to reflect upward, the net effect of low clouds is warming by reflecting terrestrial radiation downward. The relationship between low cloud cover and warming is therefore more complicated than the reflection of solar radiation upward.
5. The significant claim of the KM2019 paper relating to AGW is that since the observed warming in surface temperature can be explained in terms of declining low cloud cover, no meaningful role for atmospheric CO2 concentration is possible and that therefore, the correlation between low cloud cover and surface temperature proves the falsehood of AGW theory. In this post, we test this KM2019 hypothesis that by comparing the relationship between atmospheric CO2 and global mean surface temperature below low clouds (HadCRUT4 surface temperature) with the corresponding relationship between atmospheric CO2 and global mean temperatures in the lower troposphere above low clouds (UAH). The study period is constrained to 1979-2018 by the availability of UAH satellite data for lower troposphere temperatures.
6. The results are displayed in Figure 4 and Figure 5 above. Figure 4 displays the rate of warming in the study period for each calendar month labeled 1=January to 12=December along with the strength of the regression in terms of the T-statistic for both the surface temperature below low clouds (HAD) and the lower troposphere temperature (UAH) above low clouds. The warming rate is seen to be much stronger under low clouds than above low clouds. This result is consistent with the low cloud effect assumed in the KM2019 paper.
7. The further conclusion in KM2019  that the observed warming during a time of decreasing low cloud cover proves the CO2 effect in AGW to be false is tested in Figure 5. If this KM2019 finding is correct, we would expect to find that temperatures below low clouds (HAD) would be relatively unrelated to atmospheric CO2 concentration but that temperatures above low clouds (UAH) would show a stronger correlation. What we see in Figure 5, however, is the exact opposite. Temperature below low clouds has a stronger correlation with Ln(CO2) {the natural logarithm of atmospheric CO2 concentration} than do lower troposphere temperatures above low clouds. In case of detrended correlations, though all values are generally very low, values above DETCORR=0.2 are statistically significant and are found only in the HAD surface temperatures below low clouds. These data are inconsistent with the KM2019 finding that the warming trend during a time of declining low cloud cover proves CO2 forcing of surface temperature to be false.
8. An additional consideration is that cloud albedo is not just in low clouds but also from high clouds that have the greater cooling effect along with stratospheric aerosols. No declining trend in high clouds is evident in the data (Figure 2) that appear to show only cyclical variations. In light of these considerations, the emphasis on low clouds in the KM2019 analysis appears to be a form of data selection bias.
9. CONCLUSION: The data show that the relationship between measures of AGW and low cloud cover is more complicated than implied by the KM2019  paper and that the correlation of warming above and below low clouds with atmospheric CO2 concentration are inconsistent with the interpretation of the data in the KM2019 study. As a reference, the cloud page in the climate4You blog [LINK] may contain more useful information on the interpretation of ISSCP cloud cover data than the relatively simplistic analysis contained in KM2019. Atmospheric water vapor content, low clouds, high clouds, and their combined surface temperature effects are more complicated and these effects vary regionally.

Update 3/25/2020): Figure 4B is added to show a relative correspondence in the warming rates implied by HAD and UAH temperatures 1979-2019 in support of using two different temperature data sources. With thanks to Kiri58 for his insightful comment which appears below.