Nikolov & Zeller: Misrepresentation of Critical Satellite Data by IPCC

1. Introduction

The 6^th Assessment Report of the Intergovernmental Panel on Climate Change (IPCC AR6) concluded “It is very likely that well-mixed GHGs [greenhouse gases] were the main driver of tropospheric warming since 1979” (IPCC, 2021; p.5). This statement implies that all known climate forcings have properly been evaluated using the available data, and GHGs have been found to exert a disproportionally large radiative effect on the Global Surface Air Temperature (GSAT) over the past 45 years. However, a close examination of Chapter 7 of the Working Group I (WG1) Contribution to the IPCC AR6 (Forster et al. 2021), which discusses the Earth’s energy budget, climate feedbacks and climate sensitivity, reveals that the observed decrease of Earth’s albedo and the corresponding increase of absorbed shortwave radiation by the Planet for the past 20 years have not been taken into account as contributors to the recent warming. Section 7.2.2 of Chapter 7 entitled “Changes in Earth’s Energy Budget” acknowledges that there have been multidecadal periods of significant decreasing and increasing trends in surface solar radiation (SSR) called “global dimming” (i.e. from 1950s to 1980s) and “global brightening” (after 1980s), respectively. The report states: “There is high confidence that these [SSR] trends are widespread, and not localized phenomena or measurement artefacts.” Indeed, the existence of such dimming and brightening multidecadal periods has been acknowledged by science for more than 10 years (Stanhill et al. 2014; Yuan et al. 2021), but the IPCC AR6 provides no global estimate of the observed positive trend in SSR since 1980s and its impact on GSAT. Instead, the Report simply states “The origin of these trends is not fully understood”.

With respect to the top-of-the-atmosphere (TOA) solar fluxes, Section 7.2.2 of the IPCC AR WG1 offers no analysis of the substantial decrease in Earth’s shortwave reflectance since 2000 observed by the NASA’s Clouds and the Earth’s Radiant Energy System (CERES) project (Loeb et al. 2018, 2019, 2020 and 2021) and also reported by other research teams (e.g. Dübal & Vahrenholt 2021; Stephens et al. 2022). The Report does not discuss the observed 2.0 W m^-2 increase in solar-energy uptake by the Planet from 2000 to 2020 nor its contribution to the recent warming. What is even more puzzling, Subsection 7.2.2.1 of the IPCC AR6 WG1 Contribution features graphs in their Fig. 7.3 (on p. 936) showing an increasing reflected solar flux and decreasing outgoing thermal flux since 2000 that are supposedly based on CERES data. However, these trends are opposite of what CERES has actually measured and directly contradict results reported by prior studies. This article presents findings from our investigation of Fig. 7.3 in the IPCC AR6.

• Figure 7.3 in the IPCC AR6 WG1 Contribution

Page 936 of the IPCC Climate Change 2021: The Physical Science Basis (IPCC, 2021) contains the following figure:

Panel (a) of Fig. 7.3 shows anomalies of the reflected solar flux, panel (b) depicts anomalies of the emitted (outgoing) thermal (longwave) flux, and panel (c) displays net-flux anomalies calculated as a difference between absorbed solar shortwave (SW) and outgoing longwave (LW) anomalies. Since reflection is opposite of absorption, the absorbed SW anomalies are obtained by simply multiplying the reflected solar anomalies by -1. Figure 7.3 also portrays simulation results from 7 climate models that were forced by observed sea-surface temperatures (SST) and sea-ice boundary conditions. Note that the CERES data are depicted as thick black lines while the multi-model means are shown as thick red lines.

The problem with IPCC’s Fig. 7.3 is that the plots of reflected solar and outgoing thermal radiation show opposite temporal trends compared to trends of the same fluxes found in the actual CERES dataset. Figures 1 and 2 below illustrate this fact. At the same time, the net flux (a.k.a. the Earth’s Energy Imbalance or EEI) shows a correct trend (Fig. 3).

Figure 1. Trends of Earth’s reflected solar radiation in the CERES dataset (upper panel) and the IPCC AR6 WG1 Fig. 7.3(a) (lower panel). Note that the 13- vs. 12-month running means do not affect the long-term flux trends.

Figure 2. Trends of Earth’s outgoing thermal radiation in the CERES dataset (upper panel) and the IPCC AR6 WG1 Fig. 7.3(b) (lower panel). Note that the 13- vs. 12-month running means do not affect the long-term flux trends.

Figure 3. Trends of Earth’s net radiative flux (energy imbalance) in the CERES dataset (upper panel) and the IPCC AR6 WG1 Fig. 7.3(c) (lower panel). Note that the trend of the net flux has not been altered in the IPCC’s Fig. 7.3. Note that the 13- vs. 12-month running means do not affect the long-term flux trends.

We discovered this trend issue while working on a paper evaluating the contribution of Earth’s observed albedo decrease since 2000 to recent warming using the CERES dataset.

Upon a close examination, it became clear to us that the trend inversion of reflected solar and emitted thermal flux anomalies appeared to have been achieved by multiplying the original data series by -1.0. The text of the IPCC AR6 referring to Fig. 7.3 does not mention anything about altering the trends of key global energy-budget parameters, nor does it provide a clear explanation of the reason behind it. However, the caption of Fig. 7.3 contains the following peculiar sentence “All flux anomalies are defined as positive downwards, consistent with the sign convention used throughout this chapter.” This statement is perplexing, because flux anomalies are independent of the flux direction. Anomalies are simply departures of timeseries data from an arbitrary reference value. In the case of the CERES dataset, flux anomalies are calculated with respect to the mean value of the deseasonalized timeseries over the entire observational period. Hence, the choice of a reference value solely determines, whether an anomaly is positive or negative irrespective of the flux direction. In other words, the sign of an anomaly is not subject to the definition of a positive flux direction as suggested by the caption of Fig. 7.3. Furthermore, calculating the anomalies of an environmental parameter (such as global temperature, ocean heat content, radiative flux etc.) does not change the temporal trend of the original data set. This is basic knowledge in the field of climate science. In this regard, the statement in the caption of Fig. 7.3 relating anomalies to a flux direction is physically meaningless and confusing.

Another problem with the IPCC Fig. 7.3 is that the net-flux anomalies shown in panel (c), whose trend has not been altered, cannot be calculated from the timeseries shown in panels (a) and (b) as expected. Since the net-flux anomaly is a difference between anomalies of the absorbed solar and emitted thermal fluxes, the data curves on the graph in panel (c) are numerically incompatible with those plotted in panels (a) and (b).

• Correspondence with IPCC Authors about Fig. 7.3.

In an effort to clarify the above situation, we reached out to the Coordinating Lead Authors of Chapter 7 in the IPCC AR6 WG1 Contribution, Prof. Piers Forster at the University of Leeds (UK) and Prof. Trude Storelvmo at the University of Oslo (Norway). We asked them, who the authors of Section 7.2.2 in Chapter 7 were, where Fig. 7.3 appears. Prof. Forster promptly replied informing us that Section 7.2.2 was written and edited by Dr. Matthew Palmer at the University of Bristol and UK’s Met Office (UK) and Dr. Chris Smith at the University of Leeds (UK). Dr. Palmer was the lead author of that Section. Prof. Forster also provided us with a link to the IPCC AR6 WG1 repository at GitHub.com, where all data files and Python processing scripts reside used to generate the figures appearing in the WG1 Report.

We searched the GitHub repository and found three text files containing the source data for Fig. 7.3 as well as a Python plotting script used to generate the actual figure panels. There was a TXT file for every panel of Fig. 7.3 providing both CERES observations and climate-model projections covering the period July 2000 – June 2017. Upon inspecting the data files and the Python script, we discovered that, indeed, the deseasonalized monthly anomalies of reflected solar and emitted thermal fluxes have been multiplied by -1 resulting in inversion of the original trends, and that this data manipulation occurred in the Python script generating the panels of Fig. 7.3. The text files contained unaltered data that exhibit correct trends. Figure 4 depicts the anomalies of reflected solar and outgoing thermal fluxes from CERES observations and climate-model simulations as found in the text data files on GitHub. Figure 5 shows a portion of the Python script, where the signs of shortwave (SW) and longwave (LW) anomalies (code parameters SW_dict and LW_dict) are inverted during the plotting process by placing a minus sign in front of the corresponding data arrays. The Python script was utilized to invert the trends of both the CERES data and climate-model projections.       

Figure 4. Plots of TOA reflected solar and outgoing thermal flux anomalies found in text data files at the IPCC GitHub repository and used to generate Fig. 7.3 in the IPCC AR6 WG1. These data show correct trends.

Figure 5. Portion of the Python script used by the IPCC authors of Section 7.2.2 to plot the TOA anomalies of reflected shortwave (SW), emitted longwave (LW) and net fluxes from both CERES observations and climate model projections. Note the minus sign in front of the arrays containing SW and LW flux data.  

It is worth noting that, although the climate models correctly reproduced the overall pattern of observed changes in planetary SW and LW fluxes, the measured trends of these variables are much steeper compared to the modeled ones. Most importantly, the climate models significantly underestimate the rate of decrease of Earth’s albedo and the associated increase of absorbed solar radiation by the Planet since 2000. The rate of planetary cooling (i.e. emitted thermal flux) was also significantly lower in the model simulations compared to observations. A key detail in this comparison is that all models have been forced with observed sea-surface temperatures (SST) and sea-ice boundary conditions. Had the models been only guided by their internal physical processes without periodic nudging by real world observations, the above model-data correlation would likely have been much worse or even of opposite sign.

On July 8, 2024, we sent an email message to Dr. Palmer and Dr. Smith informing them about the findings from our search of the GitHub data repository and asking them to explain the reason for the trend inversion of the SW and LW flux anomalies in Fig. 7.3. We also sought their advice about whether to use the timeseries shown in Fig. 7.3 or the data found in the source text files, if we decide to create customized graphs of TOA fluxes for a review paper we’ve been working on.

We received a reply from Dr. Palmer on July 10, 2024, where he acknowledged that the reflected solar and outgoing thermal flux anomalies had intentionally been multiplied by -1. However, his explanation for this data manipulation was simply an expansion of the justification stated in the caption of Fig. 7.3 that invoked flux direction. Specifically, Dr. Palmer wrote:

“… reflected SW and outgoing LW are both defined as positive in the upward/outward direction. Therefore, for those timeseries we multiply by -1 so that they are expressed in a way that is consistent with the rest of the chapter. This means, for example, that a decrease in reflected SW means a relative GAIN of energy in the Earth system. Similarly, an increase in outgoing LW means a relative LOSS of energy in the Earth system. Note that in the figure we label these as “global solar flux anomaly” and “global thermal flux anomaly” rather than “reflected SW flux” and “outgoing LW flux”.”

As discussed above, this explanation makes no physical sense, because anomalies are always defined with respect to a chosen reference value and, therefore, have nothing to do with flux direction. Also, expressing a timeseries in terms of anomalies is not supposed to change the trend of the original data. Dr. Palmer correctly pointed out that multiplying anomalies of the reflected solar flux by -1 produces a timeseries of a relative energy gain by the system. This new timeseries is called absorbed solar flux, because reflection is opposite of (and complementary to) absorption. Hence, panel (a) of the IPCC Fig. 7.3 essentially shows anomalies of the absorbed solar flux by Earth. The problem is that the caption of Fig. 7.3 labels this panel as “reflected solar”, which is misleading. Since Dr. Palmer mislabeled the flux in the figure caption while recognizing that Fig. 7.3 (a) depicts a relative gain of solar energy by the Earth system, this obscured a key natural driver of climate related to the Sun.

On the other hand, multiplying anomalies of the outgoing thermal flux by -1 does not produce anything meaningful, because (unlike the solar flux) Earth’s LW radiation is always directed outward and does not have a complementary flux directed inward. By showing a decreasing thermal emission from Earth over time as done in Fig. 7.3 (b), the authors of Section 7.2.2 (Dr. Palmer and Dr. Smith) suggest a “heat trapping” in the climate system by increasing concentrations of atmospheric greenhouse gases. However, the 2^nd Law of Thermodynamics makes it impossible for an open system with a rising surface temperature such as Earth to have a decreasing emission of outgoing thermal radiation. In other words, by inverting the trend of the TOA outgoing LW flux, the IPCC authors have misrepresented the physical reality! 

Interestingly, Dr. Palmer advised us to use the data in the text files found in the GitHub repository in case we wanted to create customized plots of CERES and modeled radiative fluxes. We interpreted this as an acknowledgement that the correct data were contained in the text files rather than Fig. 7.3.

In our reply to Dr. Palmer, we listed a series of specific concerns that the trend inversion of reflected solar and outgoing thermal fluxes made in Fig. 7.3 was methodologically inappropriate, because it fundamentally changes the observed behavior of the climate system over the past 20 years and creates a false impression about climate drivers in the minds of researchers and politicians reading the IPCC Report. Dr. Palmer did not address our concerns and instead directed us toward an official IPCC webpage, where we could further raise the issue. Although he did not recognize the misrepresentation of satellite data in Fig. 7.3, it is possible that he was genuinely confused about flux anomalies and how they are calculated, since he made the following odd statement in one of his replies: “I don’t think there is any fundamental issue here – just different choices about the sign convention used”.

The truth is that the flux-trend inversions in Fig. 7.3 have enormous implications for both IPCC’s main conclusions and the climate theory; thus, they must be exposed and explained to the Public.

• Implications of the Data Manipulation in the IPCC Fig. 7.3

Figure 7.3 in the IPCC AR6 WG1 essentially shows an increasing planetary albedo (panel a) and a decreasing infrared cooling to Space (panel b) for the past 2 decades, which is diametrically opposite of satellite observations. While the text of the IPCC WG1 Chapter 7 does not discuss any long-term trends of the TOA reflected solar and emitted thermal fluxes in the 21^st-Century, Fig. 7.3 subconsciously suggests that the solar forcing played no role in recent warming and the rising concentrations of atmospheric greenhouse gases due to human industrial activity had increased the retention of heat in the climate system by impeding the outgoing LW radiation. These implications of Fig. 7.3 based on manipulated data align perfectly well with the radiative greenhouse theory of climate change, but contradict directly the physical reality as revealed by CERES measurements.

By inverting the trend of reflected solar flux, the IPCC authors effectively eliminated the need to analyze the cloud-controlled solar forcing and its contribution to recent tropospheric warming while reaffirming at the same time the a-priori assumed pivotal role of greenhouse gases in driving the global surface temperature since 2000. The trend inversion of the outgoing thermal flux further solidifies the false impression that the Earth had warmed in response to “heat trapping” by increasing atmospheric trace gases.

Had the IPCC AR6 WG1 acknowledged the ~2.0 W m^-2 increase of solar radiation absorption by the Planet between 2000 and 2020 as measured along the 13-month running mean curve of CERES anomalies (Fig. 6), the following conclusions/claims found in the Report’s “Summary for Policy Makers” would have been impossible to defend “It is unequivocal that human influence has warmed the atmosphere, ocean and land” and “It is very likely that well-mixed GHGs were the main driver of tropospheric warming since 1979” (IPCC, 2021; p.4-5). This is because a solar forcing of 2.0 W m^-2 is more than sufficient to explain the entire observed warming for the past 2 decades, thus eliminating the need to invoke any model-generated (i.e. theoretical) radiative forcing by greenhouse gases.

Figure 6. Monthly radiative anomalies of the Earth’s absorbed solar flux estimated from the CERES EBAF 4.2 dataset by multiplying the reported reflected all-sky shortwave anomalies by -1 (in accordance with the fact that absorption is diametrically opposite of reflection).

For example, using the lowest value of IPCC’s transient climate sensitivity estimated from data on GHG radiative forcing and temperature increase presented in Sections 7.3.5.2 and 7.3.5.3 of the IPCC AR6 WG1 Chapter 7 (Forster et al. 2021), i.e. 0.47 K/(W m^-2), the warming caused by solar forcing alone for the period 2000 – 2020 should have been 2.0*0.47 = 0.94 K. The observed global surface warming over this time interval (calculated as an average temperature rise from 6 datasets) is 0.62 K along the curve of 13-month running means (Fig. 7). Similar results are obtained, if one uses the linear trends depicted in Figures 6 and 7 instead. Assuming the above IPCC climate sensitivity estimate, the mean decadal increase of Earth’s sunlight absorption measured by CERES (Fig. 6) corresponds to a global surface warming of 0.797*0.47 = 0.37 K/decade, while the actual observed warming is only 0.23 K/decade (Fig. 7). Thus, the observed solar forcing, which is mostly induced by cloud-albedo variations (Loeb et al. 2019), leaves no room for any action by greenhouse-gas radiative forcing or amplifying (positive) feedbacks as speculated by the IPCC AR6. In addition, according to our analysis, the IPCC’s lowest climate sensitivity estimate of 0.47 K/(W m^-2) is still an overestimation of the Earth’s actual climate sensitivity, 0.29 K/(W m^-2) discussed elsewhere (Nikolov & Zeller, manuscript under review).

Figure 7. Monthly anomalies of the global surface air temperature calculated by averaging values from 6 datasets (HadCRUT5, GISTEMP4, NOAA GlobalTemp, BEST, RSS and NOAA STAR) and a 13-month running mean used to smooth the interannual variability.

In other words, the measured increase of Earth’s sunlight absorption by CERES in the 21^st Century, if acknowledged in the IPCC AR6, would have falsified the Report’s central assertion that human carbon emissions were the main driver of climate in recent decades. It is conceivable then that radiative flux anomalies in Fig. 7.3 were manipulated and discussions about long-term CERES trends in Section 7.2.2 were intentionally omitted, precisely because the actual CERES data present a significant empirical challenge (obstacle) to the UN’s political Agenda (established by Resolution A/RES/43/53 of the United Nations General Assembly in 1988) to promote Anthropogenic Climate Change.

• Conclusion

Considering the above facts and the enormous global socioeconomic impact of the IPCC’s conclusions and recommendations, we believe that it will be in the World’s best interest to launch an independent, critical reevaluation of fundamental premises in the climate theory from the standpoint of modern observations, and establish a new, objective peer-review system that ensures a complete and unbiased representation of all available data in the IPCC Reports. These efforts should be accompanied by a dedicated and decisive depoliticization of climate science through appropriate legislation (International Law) that also incentivizes the adoption of novel approaches to solving climate physics problems.