How declining aerosols and rising greenhouse gases forced rapid warming in Europe since the 1980s



[1] Mainland Europe's temperature rise of about 1°C since the 1980s is considerably larger than expected from anthropogenic greenhouse warming. Here we analyse shortwave and longwave surface forcings measured in Switzerland and Northern Germany and relate them to humidity- and temperature increases through the radiation- and energy budget. Shortwave climate forcing from direct aerosol effects is found to be much larger than indirect aerosol cloud forcing, and the total shortwave forcing, that is related to the observed 60% aerosol decline, is two to three times larger than the longwave forcing from rising anthropogenic greenhouse gases. Almost tree quarters of all the shortwave and longwave forcing energy goes into the turbulent fluxes, which increases atmospheric humidity and hence the longwave forcing by water vapour feedback. With anthropogenic aerosols now reaching low and stable values in Europe, solar forcing will subside and future temperature will mainly rise due to anthropogenic greenhouse gas warming.

1. Introduction

[2] In the IPCC Forth Assessment Report (AR4), continental- and global-scale surface temperature is shown to decrease slightly from the 1950s to the 1970s, but drastically increase since the 1980s, with strongest temperature rises on the northern continents [Jones and Moberg, 2003; Intergovernmental Panel on Climate Change, 2007]. Solar radiation, measured at various regions around the globe, exhibit similar behaviour, with irradiances decreasing after the mid-1950s and showing increasing trends since the 1980s [Ohmura, 2006]. In Europe this dimming [Ohmura and Lang, 1989; Russak, 1990; Gilgen et al., 1998; Stanhill and Cohen, 2001; Liepert, 1997; Liepert and Tegen, 2002; Power, 2003] and brightening of surface solar radiation [Wild et al., 2005; Pinker et al., 2005; Norris and Wild, 2007] was related to cloud changes and changes of the aerosol load in the troposphere, which strongly increased after the Second World War, and drastically decreased due to enormous efforts to curb emissions and air pollution since the mid-1980s [Sliggers and Kakebeeke, 2004]. Aerosol optical depth (AOD) records from six remote locations in Europe confirmed the recent clearing of the air, particularly in the lower troposphere, showing a 60% decline of AOD and hence anthropogenic aerosol concentration since 1986 [Ruckstuhl et al., 2008]. The same report also confirmed increasing atmospheric transmission by surface solar irradiance measurements at a large number of stations in Northern Germany and Switzerland since 1981, showing significant global radiation increases, which strengthen and confirm the claim for solar brightening.

[3] Here we show the climate impact solar brightening had on a continental and a more maritime region in Europe, and contrast it to longwave radiative forcings in order to show the role declining aerosols and rising anthropogenic greenhouse gases had on the observed rapid temperature rise since the early 1980s. For this we analyze the radiation and energy budget using measured shortwave and empirically derived longwave radiative forcings at the surface, and relate them to observed temperature and humidity changes in Switzerland and Northern Germany. We use two completely independent sets of data and average over 25 MeteoSwiss radiation stations that cover the north plateau and southern parts of Switzerland below 1000 m a.s.l., and over eight stations from the German Weather Service that cover Northern Germany from south of Berlin to the Baltic Sea [Ruckstuhl et al., 2008].

2. Temperature, Humidity and Shortwave Radiation Changes

[4] Figure 1 shows the three climate parameters, temperature (T), absolute humidity (Uabs) and solar global irradiance or the total of direct and diffuse shortwave downward radiation (SDR), which are all measured at the screen level height (2 meter above ground). Annual means from 1981 to 2005 are averaged over the 25 Swiss (left graphs) and the eight German stations (right graphs). Linear regression lines through the 25 years (red) and trend values (red numbers) are given per decade with the 95 percent confidence interval in square brackets [±2 stdev] (trends shown in the text have the same notation). All parameters show large and statistically significant positive trends and solar brightening is evident in both regions. Variations in total solar irradiance measured from spacecraft since 1979, are too small to have appreciably contributed to this increase [Foukal et al., 2006].

Figure 1.

Temperature, humidity and radiation increase since 1981. Annual mean values of temperature (T), absolute humidity (Uabs) and shortwave downward radiation (SDR) averaged over (left) 25 stations in Switzerland and (right) eight stations in Northern Germany. Linear regression lines and decadal trends with 95% confidence interval in brackets (red) are shown for the period 1981 to 2005. Linear regressions and trends (blue) are also given for the same period but without the year 2003. The rapid temperature and humidity increase is influenced by a gradual solar radiation increase but not specifically by the extreme summer 2003.

[5] Year to year variations of SDR values are quit different between Swiss and German records except for the year 2003 showing very high solar irradiance in both regions. In order to show that the observed temperature rise is not merely due to the extreme summer 2003 radiation increase, linear regression lines are added without 2003 to all parameters (dashed blue). Decadal trends (blue numbers) show marginally smaller temperature rises, while absolute humidity trends are slightly larger without 2003. SDR trends instead declined by about 1 W m−2 dec−1, but are still largely positive. Since the interest here is to study gradual “long-term” changes from 1981 to 2005, the strongly cloud affected year 2003 has been neglected in the following analysis.

3. Surface Radiation Budget Changes

[6] The surface radiation budget with the individual radiation components and respective decadal trends and 95 percent confidence intervals is shown in Figure 2 for annual means averaged over the Swiss (left) and the German stations (right). In the radiation budget solar radiation only appears as shortwave net radiation (SNR), with the reflected part due to the surface albedo (A) being subtracted by multiplying SDR by (1 − A). Since we don't have continuous albedo measurements over the full measuring period we cannot make assumptions on possible climate forcings due to changing albedo, and instead used annual average values of A = 0.28 measured at Payerne for all month and stations. This value may be high for Germany and therefore rather underestimates the climate forcing effects.

Figure 2.

Radiation budget and surface forcings. Annual mean values averaged over (left) 25 stations in Switzerland and (right) eight stations in Northern Germany are given for the individual components of the surface radiation budget. The graphs show shortwave net radiation (SNR), longwave downward radiation (LDR), total absorbed radiation (TAR), longwave upward radiation (LUR) and total net radiation (TNR) from 1981 to 2005 all in W m−2. Downward fluxes are positive upward fluxes negative. Decadal trends are given in W m−2 dec−1 with the 95% confidence interval in brackets. TNR is the balance between downward and upward fluxes at the surface and represents the energy available for the sensible and latent energy fluxes. TNR increases primarily increase water vapour in the atmosphere, but part of it is lost by circulation changes, ground fluxes and ocean heat uptake.

[7] Longwave downward radiation (LDR) is derived from monthly values of absolute humidity (Uabs) measured at all stations and an empirical relation determined with measured monthly values of LDR and Uabs at the two reference stations Locarno-Monti (388 masl.) and Payerne (498 masl.) in Switzerland from 2001 to 2005. The used relation is: LDR = 189.36 * (Uabs ^ 0.26) [Ruckstuhl et al., 2007]. However, this empirical relation is based on humidity and related temperature changes, and does not account for LDR increases by anthropogenic greenhouse gases (aGHG), nor does it account for LDR changes that are due to changing cloud amount and optical thickness. The radiative forcing at the tropopause of all the long-lived, well mixed greenhouse gases increased from 1.7 Wm−2 in 1979 to 2.65 Wm−2 in 2004 [Hofmann et al., 2006]. An estimated surface forcing due to aGHG increases of +0.35 [+0.30 to +0.40] W m−2 per decade is therefore added. Hence, LDR shown in Figure 2 represents annual averages of longwave downward radiation, that change due to changes of temperature and all greenhouse gases, but changing cloudiness is not included here.

[8] Total absorbed radiation (TAR), the sum of SNR and LDR, represents the total radiative energy available to maintain the Earth's surface temperature and to sustain the turbulent (sensible and latent) heat fluxes in the atmosphere. The observed increase of TAR is the sum of the increase of SNR and LDR, and this radiation increase forced the observed temperature rise and increased the water vapour in the atmosphere.

[9] The longwave upward radiation (LUR) is the thermal radiation emitted from the Earth's surface. Earth's surface emissivity ranges from 0.9 to 0.99 with an average of about 0.93. However, a large percentage of the emitted longwave upward radiation comes back as longwave downward radiation (apparent sky emissivity 0.82 in Europe) [Philipona et al., 2005]. Since absorption is equal to emission, LDR is partly reflected on the ground, adding to LUR and leading to an apparent emissivity of about 99 percent [Marty et al., 2002]. LUR is negative in the radiation budget and is calculated using the Stefan-Boltzmann law and monthly mean temperature values measured at the individual radiation stations. Trends are directly related to the temperature trends.

[10] Total net radiation (TNR) finally is the sum of the surface absorbed and the surface emitted radiation fluxes, TAR plus LUR. TNR is positive and in the physical sense represents the excess of radiation energy that has not been used to heat the Earth's surface, and instead is available for ground fluxes (usually very small) and primarily to sustain the sensible and latent heat fluxes, thereby closing the surface energy budget.

[11] An increasing TNR over the investigated period demonstrates, that part of the increasing TAR was used to rise the sensible and latent fluxes and hence increased water vapour in the atmosphere, but part of it may also have been lost by ground fluxes, changing atmospheric circulation, or ocean heat uptake in Northern Germany. In Switzerland the +1.29 [+0.19 to +2.41] W m−2 dec−1TNR increase represents a 3% dec−1 increase with respect to absolute TNR averaged over the period (see Figure 2). Absolute humidity Uabs on the other hand, which for monthly means is proportional to atmospheric column integrated water vapour (IWV) [Ruckstuhl et al., 2007] increased by 2.8% dec−1 (see Figure 1). Hence, atmospheric water vapour increased more or less in proportion to increasing TNR. In Northern Germany however, TNR increased by 4.8% dec−1, whereas Uabs only by 2.2% dec−1, which suggests that here part of the rising TNR energy was apparently absorbed by ocean heat uptake or lost by changing circulation.

4. Shortwave and Longwave Surface Forcings

[12] In Table 1 we show the individual forcings of the SNR and LDR components that are due to changing aerosols, cloudiness and greenhouse gases, as well as flux changes that are due to changing temperature and humidity. In our recent analysis [Ruckstuhl et al., 2008], it was shown that the strong AOD decline in Europe had a larger impact on SNR through the direct aerosol effect (SNR AODdir) than through cloud mediated indirect aerosol effects (SNR AODind). With cloud mediated effects we mean changes in total cloudiness or cloud optical depth due to changing aerosols or due to changing circulation (our measurements do not allow to separate the two effects). It was further shown, that a large part of the SNR AODind surface forcing, which increased due to decreasing cloudiness, was compensated by decreasing longwave downward radiation. In Switzerland cloud forcing analyses showed that about three quarters of any SNR increase is compensated by LDR decrease [Marty et al., 2002]. Hence, to determine the effective shortwave climate impact of direct and indirect aerosol effects combined (SNR AODdir-ind), the longwave compensated part (three quarters of SNR AODind) needs to be subtracted as SNR Cloud-comp (see Table 1). The effective shortwave climate impact SNR AODdir-ind therefore becomes +0.76 [+0.03 to +1.50] W m−2 dec−1 in Switzerland, and +1.25 [+0.17 to +2.32] W m−2 dec−1 in Northern Germany.

Table 1. Shortwave and Longwave Surface Forcings Due to Aerosols, Clouds, a GHG, Water Vapor, and Temperature Changes
 25 Swiss Stations (W m−2 dec−1)8 German Stations (W m−2 dec−1)
SNR AODdir+0.62+1.07
SNR AODind+0.50+0.62
SNR Cloud-comp−0.36−0.44
SNR AODdir-ind+0.76+1.25
LDR Uabs+T+2.37+2.05
LDR aGHG+0.35+0.35
LDR Cloud-comp+0.36+0.44
LDR T−2.09−1.89
LDR aGHG+WV+0.99+0.95
LDR WV+0.64+0.60

[13] On the longwave side, LDR increased due to rising temperature and humidity (LDR Uabs+T) and due to increasing anthropogenic greenhouse gases (LDR aGHG). However, in order to assess the individual effects that temperature, humidity and anthropogenic greenhouse gas increases had on LDR, the longwave cloud compensation part (LDR Cloud-comp), which was subtracted as SNR Cloud-comp on the shortwave side needs to be added on the longwave side. From the sum of all longwave downward forcings LDR Uabs+T, LDR aGHG, and LDR Cloud-comp, we now subtract the effect that is due to the rising temperature at the surface and in the atmosphere (LDR T). LDR T is calculated by multiplying the LUR increase by the apparent longwave sky emissivity 0.82 for Europe [Marty et al., 2002]. The longwave forcing due to all greenhouse gases (LDR aGHG+WV) therefore becomes +0.99 [+0.40 to +1.58] W m−2 dec−1 in Switzerland and +0.95 [+0.26 to +1.64] W m−2 dec−1 in Germany, and by subtracting LDR aGHG we find the forcing that is due to the atmospheric water vapour feedback (LDR WV), which is +0.64 [+0.26 to +1.02] W m−2 dec−1 in Switzerland and +0.60 [+0.16 to +1.04] W m−2 dec−1 in Northern Germany.

5. Discussion

[14] According to this analysis the initial climate forcings that let to the observed rapid temperature rise since 1981 are of anthropogenic nature. Interestingly, in Switzerland, the combined direct and indirect aerosol shortwave forcing SNR AODdir-ind is twice as large as the longwave anthropogenic greenhouse forcing LDR aGHG, and in Northern Germany even three times as large. Models estimate that the aGHG forcing at the surface may be smaller than at the tropopause [Allan, 2006], which would make the ratio between aerosol- and greenhouse forcing effects even larger. The anthropogenic forcings are further amplified by the longwave water vapour feedback LDR WV that is also larger than the LDR aGHG forcing. The analysis of the radiation- and the energy budget further shows that the different radiative forcings observed and determined at the surface are well balanced by measured temperature and absolute humidity increases, at least in continental Switzerland. The smaller temperature rise under larger radiative forcing in Germany can be explained by larger energy losses through ocean heat uptake in the more maritime Baltic Sea region.

[15] The balanced energy budget in Switzerland is very instructive. If we sum all shortwave and longwave radiative forcings (SNR AODdir-ind + LDR aGHG + LDR WV) we find a total climate forcing of +1.75 [+0.59 to +2.91] W m−2 dec−1. Of this total forcing, TNR or the radiation balance of +1.29 [+0.19 to +2.41] W m−2 dec−1 was used to increase the sensible and latent fluxes and hence atmospheric water vapour. The residual or +0.46 [+0.16 to +0.76] W m−2 dec−1 of the forcing was responsible for the measured temperature rise of +0.49 [+0.21 to +0.77]°C dec−1 in Switzerland, and this results in a heating rate of about +1°C/W m−2, which is a reasonable value for direct radiation heat conversion over land. It is though interesting to observe, that almost three quarter of the energy of all radiative forcings was used to increase the turbulent fluxes and hence water vapour in the atmosphere.

6. Conclusions

[16] Recent studies reported decreases of anthropogenic aerosols in the industrialized world and over the oceans, and suggested connections between solar brightening and aerosol effects [Streets et al., 2006; Mishchenko and Geogdzhayev, 2007; Wild et al., 2007]. Given the climate impact of aerosols and greenhouse gases shown in this study, rising air pollutants from the 1950s to the 1980s, and subsequent decreases due to drastic emission reductions in Europe and other industrialized nations [Sliggers and Kakebeeke, 2004; Stern, 2006], likely explain the observed temperature evolution over mainland Europe since 1950 (see Figure 3). Slightly decreasing temperatures since the 1950s (green) are in line with rising aerosols and a solar dimming, which offset anthropogenic greenhouse warming [Charlson et al., 1992]. Of the rapid temperature rise since the 1980s (blue) however, about two thirds are shown by our analysis to be likely forced by aerosol decline and related solar brightening that strongly reinforced anthropogenic greenhouse forcing. With anthropogenic aerosols now reaching low and stable values over Europe [Ruckstuhl et al., 2008; Sliggers and Kakebeeke, 2004], solar brightening will subside and temperature trends for the next decades will likely reduce to about +0.2°C dec−1, a level similar to the overall long-term increase since 1950 (red), which in the future will mainly be driven by anthropogenic greenhouse gas warming.

Figure 3.

Temperature rise over mainland Europe since 1950. Annual temperature differences for the period 1950 to 2005 with respect to the 1961–1990 climatological mean for mainland Europe (45°–55°N; 5°–15°E). Linear regression lines and decadal trends with 95% confidence interval show the temperature decline from 1950 to 1980 (green), and the temperature rise to be twice as large for the period 1981 to 2005 (blue) than for the whole period 1950 to 2005 (red). Crutem2 data from CRU/UEA.


[17] This study was completed as part of the National Center of Competence in Research on Climate (NCCR Climate), an initiative funded by the Swiss National Science Foundation (NSF).