This is the fourt hin a series dealing with anthropogenic CO2; in particular, why no measurable decline in atmospheric CO2 occurred during the most severe reduction in anthropogenic emissions that has ever occurred, during the 2020 SARS-Cov-2 pandemic lockdowns.
I compare measured CO2 monthly emissions at Mauna Loa Observatory (MLO) with published figures on global monthly anthropogenic emissions, for 2020, from the International Energy Agency (IEA). Apparently, IEA has not been tracking monthly anthropogenic emissions, only annual summations. However, probably because of the puzzlement and uncertainty regarding why 2020 was a tie with the average global temperature of the El Niño event of 2016, and there was no obvious decline in the annual growth of CO2, the IEA decided to look at monthly data.
I demonstrate that anthropogenic emissions are negligible in comparison to the net CO2 concentrations driven by natural sources and sinks.
The analysis and conclusions depend on the graphs following, where all published ‘carbon’ mass units are converted to equivalent Parts Per Million (PPM) atmospheric CO2 concentration instead of using unfamiliar gigatons or petagrams. Figure 1, below, shows the shape of the curve derived from the IEA monthly data, with the addition of an estimate for 2021 as a January ‘placeholder.’ Note the minimum global value (-14.5%) In April 2020. The IEA provides no estimate for the uncertainty of the monthly data. However, according to the Empirical Rule in statistics, the standard deviation of the anthropogenic time-series will be less than 1/4th the range, 0.065/4, or <0.02 PPM. The standard deviation of the de-trended anthropogenic series would be about 0.041/4, or <0.01 PPM. That is, the trend accounts for around an additional 0.01 PPM standard deviation during the year.
Fig. 1. Monthly anthropogenic CO2 flux with trend line.
The slope of the OLS regression line, which is plotted but obscured by the data, for monthly anthropogenic flux is statistically significant (p-value = 0.00272) at better than a 95% threshold, albeit, it only predicts a flux increase of 0.04 PPM anthropogenic CO2 per month over a year. The OLS linear fit predicts about 57% of the month-to-month variance, even with the significant, anomalous decline in April.
The sum of the monthly fluxes for the 12-months of 2020 is 4.4 PPM, which is in line with other estimates for recent annual anthropogenic emissions of CO2. That is, recent average annual anthropogenic emissions amount to ≈4.1 PPM (8.8 Pg * 1 PPM/2.13 Pg) out of a total of 101 PPM, or 4.1% of all sources.
The first thing that the reader will probably notice is that when the anthropogenic and net global data are combined (Fig. 2, below), and scaled to show the full net global range, the monthly anthropogenic line is almost indistinguishable from a straight line. The anthropogenic data have a nearly constant flux of about 0.37 PPM CO2 per month, with a range of 0.07 PPM CO2.
Fig. 2. Monthly anthropogenic and net global CO2 flux together.
Whereas, the net, monthly global-fluxes have a complex sinusoidal appearance, with a range of 3.71 PPM, and a slightly negative trend (-0.38). However, the trend is not statistically significant (p-value = 0.763, R2 = 0.0086) at the 95% threshold. Included error bars represent the average (±0.5) standard deviation of the MLO measurements for this time interval.
The apparent negative trend should not be surprising because, as has been pointed out by Moncton (2022), there has been no statistically significant change in the Global Mean Temperature for over 7 years. I have previously demonstrated a positive correlation between global temperatures and the slope and range of the seasonal ramp-up of atmospheric CO2. While correlation does not necessarily demonstrate cause and effect, the association of increased CO2 during the warm El Niño events implies that warmth is driving the CO2 because it is unlikely that the warmth is causing the El Niño events! The important point is that the net seasonal CO2 range is more than 57 times greater than the range of the anthropogenic CO2. That is, the annual range of the anthropogenic emissions in 2020 is 1.8% of the net global seasonal range. It is not the driver of annual changes; it is background ‘noise.’
The anthropogenic CO2 contribution has a range almost an order of magnitude smaller than the one sigma (1σ) uncertainty of the net seasonal MLO average, monthly measurement changes. Note in particular the relative position of the anthropogenic line, in February, March, May, and October in Figure 2, compared to the error bars. Even the nominal, average monthly anthropogenic flux (0.37 PPM) is smaller than the net global monthly uncertainty. The anthropogenic contribution is a constant background component whose estimated monthly flux is within the error envelope of the MLO change measurements.
Subtracting the monthly anthropogenic emissions from the monthly net global flux would, as a first-order approximation, shift the global net curve in Figure 2 downward about 0.37 PPM, not noticeably changing the shape. NASA claims that “about 45 percent” of the anthropogenic CO2 emissions has remained in the atmosphere. This is in line with other claims that about half of the anthropogenic emissions remain in the atmosphere. Based on that, to see what would happen if anthropogenic emissions were to cease suddenly, we should subtract 0.17 PPM (45% of 0.37) from the monthly, net global change flux. That is much smaller than the 1σ uncertainty of the monthly, net global CO2 flux.
When time-series data have trends, it is common to observe spurious correlations between phenomena even when they do not have a causal relationship. It is usually recommended to de-trend the data and see if there is still a correlation. I do that below with this CO2 data.
There is negligible trend (slope = 0.04) in the anthropogenic data, and the slope for the net global CO2 is small enough (-0.38) that the appearance of the composite graph of residuals is essentially the same as Figure 2; therefore, I don’t show it. The 2020 de-trended net global monthly change is, however, very similar in shape and magnitudes to a NASA graph (See Figure 3, below) for earlier years of de-trended data of net monthly CO2 change. The curve shapes are surprisingly constant since at least 1959.
Fig. 3. De-trended monthly atmospheric CO2 flux in PPM. (https://www.earthobservatory.nasa.gov/features/CarbonCycle)
However, it is instructive to look at the scatterplot of the de-trended residual data. Figure 4, below, shows that there is essentially no correlation between the monthly anthropogenic emissions and the net global monthly flux; although the apparent correlation is negative, the regression line is not statistically significant (p-value = 0.854) at the 95% threshold. The R2 value shows that only 0.32% of the variance in the net monthly global CO2 is predicted or explained by the change in monthly anthropogenic emissions. In fact, the curves often head in different directions, as in April. (Compare Fig. 1 and Fig. 2 for April [2020.25]) If, there is any correlation, a time delay might be obscuring it.
However, looking at the bigger picture, the annual changes in anthropogenic emissions and the net annual global changes in atmospheric CO2, there is little to support the idea that a time delay of greater than a month is hiding the control of anthropogenic emissions over the total net source-flux change. The annual peak in northern hemisphere atmospheric CO2 occurs in May every year!
Similar to the analysis done on the monthly residual data, Figure 5, (below) shows the plotted de-trended residuals for annual anthropogenic and total net-flux. While there appears to be a slightly positive correlation, it is not statistically significant (p-value = 0.943, R2 = 0.0091). That is, the OLS regression-line trend is so close to zero that it is statistically indistinguishable from a zero trend, and the correlation coefficient squared only predicts or explains about 0.91% of the variance in the net inter-annual changes.
Fig. 4. Correlation of monthly residual global CO2 change
with respect to the monthly residual anthropogenic CO2.
Fig. 5. Correlation of annual residual global CO2 change
with respect to the annual residual anthropogenic CO2.
Anthropogenic emissions of CO2 amount to 4%, or less, of the total source fluxes. The sinks that are responsible for extracting CO2 from the atmosphere cannot differentiate natural from anthropogenic CO2. Therefore, CO2-source abundance determines the proportions removed. Tree respiration and biological decomposition dominate the Winter ramp-up phase growth; photosynthesis dominates the Summer draw-down decline. That means anthropogenic CO2 does not accumulate at the level of the nominal 2 PPM per annum increase. The annual, global atmospheric increase is the increase from all the source fluxes, minus the drawdown of all the sinks. The annual anthropogenic CO2 range (0.07 PPM) in 2020 was only about 3.3% of the nominal 2 PPM annual increase. The monthly anthropogenic flux was about 1.8% of the net, monthly global flux range. The fact that the annual atmospheric increase of 2 PPM is about half the estimated annual anthropogenic emissions is coincidence. Perhaps those who see more than that are exhibiting the common human trait of apophenia. I’m reminded of the great effort some have gone to finding special meaning in the measurements and ratios associated with the Great Pyramid.
Anthropogenic CO2 is virtually constant compared to the seasonal variations of the natural sources and sinks. The monthly anthropogenic flux change is much smaller than the uncertainty in the net global monthly flux changes. Therefore, there is no support for the claims about anthropogenic CO2 driving the annual changes. The seasonal natural source fluxes swamp the anthropogenic sources. Eliminating anthropogenic CO2 would have a negligible impact on annual increases, which is why the pandemic lockdowns had an imperceptible effect on the global atmospheric concentrations. I do not expect even draconian reductions in anthropogenic CO2 to have the kind of results claimed to justify eliminating fossil fuel use. The annual growth in CO2 is a result of increasing natural sources that is not compensated by commensurate increases in sinks.
Dr. Pieter Tans, NOAA/GML (gml.noaa.gov/ccgg/trends/) and Dr. Ralph Keeling, Scripps Institution of Oceanography (scrippsco2.ucsd.edu/).
IEA (2021), Global Energy Review: CO2 Emissions in 2020, IEA, Paris https://www.iea.org/articles/global-energy-review-co2-emissions-in-2020