Halogenated Greenhouse Gases Made Global Warming Primarily

Time-series observations of global lower stratospheric temperature (GLST), global land surface air temperature (LSAT), global mean surface temperature (GMST), sea ice extent (SIE) and snow cover extent (SCE), together with observations reported in Paper I, combined with theoretical calculations of GLSTs and GMSTs, have provided strong evidence that ozone depletion and global climate changes are dominantly caused by human-made halogen-containing ozonedepleting substances (ODSs) and greenhouse gases (GHGs) respectively. Both GLST and SCE have become constant since the mid-1990s and GMST/LSAT has reached a peak since the mid2000s, while regional continued warmings at the Arctic coasts (particularly Russia and Alaska) in winter and spring and at some areas of Antarctica are observed and can be well explained by a seaice-loss warming amplification mechanism. The calculated GMSTs by the parameter-free warming theory of halogenated GHGs show an excellent agreement with the observed GMSTs after the natural El Niño southern oscillation (ENSO) and volcanic effects are removed. These results provide a convincing mechanism of global climate change and will make profound changes in our understanding of atmospheric processes. This study also emphasizes the critical importance of continued international efforts in phasing out all anthropogenic halogenated ODSs and GHGs. Introduction It is generally agreed in the literature that the measured global mean surface temperature (GMST) had a clear rise of approximately 0.6 K between 1950 (more precisely 1975) and around 2000, coinciding with a rise in atmospheric CO2 level. There were several reports of global warming stopping for the period between 2000 and around 20151-9, whereas the GMST appears to have risen again in recent years. It is also generally agreed that El Niño southern oscillation (ENSO) is one of the largest sources of year-to-year variability, but it is not likely the main cause. This paper is devoted to understanding global climate change since the late half of the 20th century when the emission of human-made halogenated gases (mainly, chlorofluorocarbons—CFCs) into the atmosphere became significant. In the preceding paper (Paper I)10, the author reported the discovery of a large, deep and allseason ozone hole over the tropics and the data showed the formation of three ‘temperature holes’ corresponding to the ozone holes over the Antarctic, tropical and Arctic respectively. The lower stratospheric temperatures (LSTs) in the three “holes” are fairly well reproduced by the cosmic-


Introduction
It is generally agreed in the literature that the measured global mean surface temperature (GMST) had a clear rise of approximately 0.6 K between 1950 (more precisely 1975) and around 2000, coinciding with a rise in atmospheric CO2 level. There were several reports of global warming stopping for the period between 2000 and around 2015 [1][2][3][4][5][6][7][8][9] , whereas the GMST appears to have risen again in recent years. It is also generally agreed that El Niño southern oscillation (ENSO) is one of the largest sources of year-to-year variability, but it is not likely the main cause. This paper is devoted to understanding global climate change since the late half of the 20 th century when the emission of human-made halogenated gases (mainly, chlorofluorocarbons-CFCs) into the atmosphere became significant.
In the preceding paper (Paper I) 10 , the author reported the discovery of a large, deep and allseason ozone hole over the tropics and the data showed the formation of three 'temperature holes' corresponding to the ozone holes over the Antarctic, tropical and Arctic respectively. The lower stratospheric temperatures (LSTs) in the three "holes" are fairly well reproduced by the cosmic-ray-driven electron reaction (CRE) equation with the total concentration of anthropogenic halogenated ozone-depleting substances (ODS) (mainly CFCs) and cosmic ray (CR) intensity in the stratosphere as only two variables. This is also true for the extratropic/extrapolar regions (not shown in Paper I). This means that the global lower stratospheric temperature (GLST) has been controlled primarily by halogenated ODS (CFCs) and CRs. Despite its striking contrast to the expectation from current climate models, this conclusion is solidly supported by substantial observed datasets, as will be presented in this paper. It is generally agreed that the change in GLST should mirror the change in GMST. Indeed, a CFC-warming theory of GMST has been proposed by the author in previous publications 1, 3,6,9 . After removal of the CR effect, the Antarctic ozone hole has shown a clear recovery since the mid-1990s, closely following the change trend of anthropogenic ODSs 6,9,11 . A similar conclusion was also reached by others 12 . The total ODS level has turned to a declining trend measured in the troposphere since around 1994 due to the Montreal Protocol 13 . It is reasonable to expect a similar trend for the Arctic ozone hole, though the latter was not significant for every spring Arctic. In contrast, the recovery of the tropical ozone hole or ozone at mid-latitudes in the lower stratosphere was delayed by around a decade from the declining trend of ODSs 6,[9][10][11] .
Given the observations outlined above, it is important to understand why there was a warming stopping in 2000-2015 and why the GMST has risen again in recent few years. An associate question is: which greenhouse gas (GHG) is dominantly governing global climate change, CFCs or CO2? This paper aims to answer this question using substantial observation datasets including GLST, land surface air temperature (LSAT), sea ice extent (SIE), snow cover extent (SCE), GMST, as well as straightforward quantum-physical model calculations of GMST.
Here, we make a statement for the studies including Paper I 10 and this paper: we choose to use observed data as direct as possible when available instead of 'adjusted' or 'processed' data based on understanding in current climate models. The latter has led to changes in historical observed datasets. For example, the IPCC AR5 assessed estimate for historical warming between 1850-1900 and 1986-2005 was 0.61 [0.55 to 0.67] °C, while it has been changed to 0.69 [0.54 to 0.79] °C in AR6 for this same warming period due to 'changes in observational understanding' 14 . Similar 'adjustments' since 2013 can be found in most data sources used in the Report. This could create errors, given that current climate models have still significant discrepancies from observations.

Results and Discussion
(I) Global lower stratospheric temperatures (GLSTs). Furthering the presentation of timeseries annul mean LST anomaly datasets at the lower stratospheres (100-30 hPa) over the polar regions and the tropics since the 1960s in Paper I 10 , we plot the LSTs of southern hemisphere (SH) extratropic (30S-90S), northern hemisphere (NH) extratropic (30N-90N) and the global (90S-90N) from the same multiple data sources, as shown in Figure 1. Note again that the LST anomalies in original datasets were normalized to various reference temperatures and had therefore to be offset to compare with each other, but this offsetting has no effects on their long-term trends.
Here, the original LST anomalies are offset to the same level in the 2000s-2010s, during which all the datasets were available and they should have smaller uncertainties than earlier measurements.
There are large discrepancies between ground-based and satellite-based measured data prior to 1995. Interestingly, however, the LSTs from all measured datasets consistently show clear drops between the late 1970s and 1995 and have become constant since the mid-1990s with no significant change over the last two decades.
Given transport lags of 1 and 10 years for ODSs transported from the troposphere/surface to the lower stratosphere over the polar regions and the tropics respectively and an estimated lag of 5 years for mid-latitudes (30N-60N and 30S-60S), we get transport lags of approximately 4 and 7 years of ODSs respectively for the NH/SH extratropic and global mean lower stratospheres to model the LSTs using the CRE equation. To keep the simplicity, we use the same constants kAnt and kTro to calculate the LSTs of the extratropic and the global as those used for calculating LSTs over the polar regions and the tropics in Paper I respectively, except using respective different transport lags of 4 and 7 years for ODSs. That is, no optimizations to get the best fitting results are performed. This is expected to cause some discrepancy between observed and calculated data for extratropic LSTs, but it should be a good approximation for the GLST, considering that the tropical area is nearly double the total area of the South and North polar regions and the tropical ozone/temperature hole is all-season in contrast to the seasonal polar holes. The thus calculated results of LSTs are also shown in Figure 1, which match the observed data surprisingly well, especially for GLSTs (as expected). Overall, the calculated curves show excellent agreements with observed data since 1995. It is worth noting the complexity that CFCs themselves are highly effective GHGs, which tend to cause stratospheric cooling too; the destruction of CFCs in the lower stratospheres over the polar regions and the tropics (shown in Paper I) should reduce this cooling effect, compared with that over mid-latitudes. However, the stratospheric cooling effect of ozone depletion should be dominant. In spite of the obvious simplifications and approximations in above CRE-model calculations, both observed and calculated results clearly demonstrate that both extratropical and global LSTs have well been controlled only by the level of halogenated ODSs and the CR intensity and that stratospheric cooling has stopped since the mid-1990s with no significant change over the past 20 years. This is consistent with the measured trend in ODSs and observed recovery in ozone depletion shown in Paper I. Then, an obvious question arises: why has the GMST risen again in recent few years (after 2015)?  to observed temperature anomalies (symbols).  To solve the above mystery, it is important to obtain more detailed information from the map of GLST, which is made from available MSU-UAH satellite datasets 15 . The map for the annul mean GLST difference of the 2010s minus the 2000s is shown in Figure 3. It is now clearly revealed that stratospheric cooling is indeed being reversed in majority of global areas, with exceptions in some local high-latitude Arctic and Antarctic areas. The enhanced stratospheric cooling is most marked in the north and northeastern Russia, extending to the Far-East region and Alaska; it also occurs at some areas of Antarctica, especially at the region of 60S-90S, 50-100 W. This regional/local phenomenon cannot arise from ODSs and associated ozone depletion. It must originate from a different mechanism, which will be revealed by observed data to be presented below.  Figure 4(a) clearly shows a significant and highly inhomogeneous global warming in the late half of last century. In contrast, Figure 4(b) shows that, consistent with the above observations in GLST (Figure 3), surface warming stopping or reversal has occurred in most areas of the globe and only regional warming has continued in north and northeastern coasts of Russia and in Alaska of USA since the 2000s. The continued warming over some areas at Antarctica is not visible in Figure 4(b), which is probably due to few measurement stations at Antarctica.  To reveal the mechanism for these regional warmings, it is worth noting the cause of the socalled 'Arctic amplification (AA)' phenomenon in the observed much faster warming in the Arctic than the rest of the world over the last half century. Although the precise mechanism for AA is still under debate, there is an interesting mechanism proposed by Dai et al. 16 that AA is closely related to the surface albedo feedback associated with sea-ice loss, leading to increased outgoing longwave radiation and heat fluxes from newly opened waters. This mechanism was shown to be effective only from October to April and only over areas with significant sea-ice loss and to largely disappear when the sea ice is fixed or melts away. This AA mechanism is obviously consistent with the observed GLST results shown in Figures 2-4. Furthermore, it also agrees with the observed seasonal NH LSAT difference of 2010-2020 minus 2000-2010 as shown in Figure 5, which shows that continued regional warming at the Arctic coasts only occurs in the seasons of DJF (December, January and February) and MAM (March, April and May) but not JJA and SON.
Moreover, NH and SH sea-ice extent (SIE) data shown in Figure 6 indeed confirm the continued sea ice loss over the Arctic and the sudden melting at Antarctica around 2015. Over the past two decades, the GMST has been at its highest records since 1850 with a rise of 0.6-0. 8  shown in Figures 1-6, we can make a solid conclusion that regional warmings at the Arctic coasts of Russia and Alaska (USA) and some areas of Antarctica is due to continued/new sea ice loss.  Note that for these regional LSAT changes, the natural El Niño and volcanic effects were not removed as it was hard to do so for local regions or individual countries. A similar result can be found from the UK Met Office's central England temperature (CET) dataset 17 , which is the longest instrumental record of temperature in the world started in 1659, as shown in Figure S1. Moreover, time-series snow cover extent (SCE) data over NH and North America since 1965 are shown in Figure 9, both of which exhibit stabilized since ~1995. All these observed data point to a fact that global warming has stopped since around 2005.   In striking contrast to the excellent agreement between observed data in Figures 1-9 and the warming mechanism of halogenated GHGs, these robust observations do not agree with current climate models which give GMST rises caused dominantly by radiative forces of increased CO2, CH4 and N2O atmospheric concentrations. Note that there is no current controversy about the fact that the absorption of terrestrial radiation emitted from the earth surface by these major nonhalogen GHGs at the centers of their IR bands (e.g., CO2 at 667 cm 1 ) has been entirely saturated.
What is under the current debate is whether the absorption at the edge wings of their IR bands would continue to increase with rising gas concentrations and contribute considerably to the observed global surface temperature change. In current climate models, it is assumed that the greenhouse effect of these major GHGs is not yet saturated and non-linear (e.g., logarithmic for CO2) relationships between the radiative force ΔF and gas concentration for CO2, CH4 and N2O are applied, in contrast to a linear relationship for halogenated GHGs. As the author pointed out previously 3,6,9 , there are at least two major problems with this assumption. Niño southern oscillation (ENSO) is the largest source of year-to-year variability. We simply adopt the empirical model developed by Lean and Rind 22,23 , which played an important role in proving that the warming in the late half of last century was due to anthropogenic influences. In this model, lags are 4 months for ENSO, 6 months for volcanic aerosols, and 120 months for anthropogenic force. The latter (the 10-year lag), which was chosen to maximize the explained variance in Lean- Rind model, turns out to agree excellently with our observed ozone recovery trend at tropical and mid-latitudes delayed from the tropospheric halogenated ODS peak 1,6,9 . The natural contributions to the GMSTs were 0.2 C warming during major ENSO events in 1997-98, and about 0.3 C cooling in 1992 following the large Pinatubo volcanic eruption 22,23 . These values are used to remove ENSO and volcanic effects from observed GMST data without further optimization, with details given in SI and Figure S2.
Since GMST is still around the peak, ice melting at the Arctic is most likely to continue and at Antarctica may be increasing with the recovery of the Antarctic ozone hole due to the increased HFCs and PFCs by international Agreements including the highly successfully and extremely important Montreal Protocol and its Amendments, however, it is most likely to see a gradual global reversal in GMST in coming decades. Nevertheless, this reversal will come true only with continued international efforts in phasing out all halogenated ODSs and GHGs. Therefore, the tremendous efforts from international governments and community are required if humans desire to reverse the climate change caused by anthropogenic halogenated ODSs and GHGs.