Anthropogenic greenhouse forcing and strong water vapor feedback increase temperature in Europe



[1] Europe's temperature increases considerably faster than the northern hemisphere average. Detailed month-by-month analyses show temperature and humidity changes for individual months that are similar for all Europe, indicating large-scale weather patterns uniformly influencing temperature. However, superimposed to these changes a strong west-east gradient is observed for all months. The gradual temperature and humidity increases from west to east are not related to circulation but must be due to non-uniform water vapour feedback. Surface radiation measurements in central Europe manifest anthropogenic greenhouse forcing and strong water vapor feedback, enhancing the forcing and temperature rise by about a factor of three. Solar radiation decreases and changing cloud amounts show small net radiative effects. However, high correlation of increasing cloud-free longwave downward radiation with temperature (r = 0.99) and absolute humidity (r = 0.89), and high correlation between ERA-40 integrated water vapor and CRU surface temperature changes (r = 0.84), demonstrates greenhouse forcing with strong water vapor feedback.

1. Introduction

[2] Global annual mean temperature at Earth's surface increased over the past 150 years [Jones et al., 2001; Parker et al., 1995]. On the Northern Hemisphere (NH) sea surface and land-surface air temperature measurements show similar increases up until the 1980s, but from that time on greater warming is measured over land [Houghton et al., 2001]. In Central Europe (CE) temperature rises three times faster than the northern hemisphere average [Philipona and Dürr, 2004]. Temperature increases are known to be larger during winter [Jones and Moberg, 2003], however, no consensus exists on why temperature increases.

[3] In his article [Tyndall, 1861] John Tyndall argued “now if the chief influence be exercised by the aqueous vapor, every variation of this constituent must produce a change of climate”. There is yet no doubt that atmospheric water vapor is the most abundant and effective greenhouse gas [Held and Soden, 2000]. In response to human emissions of greenhouse gases, like carbon dioxide, the Earth warms; more water evaporates from the surface, and the amount of water vapor in the atmosphere increases, leading to a further increase in the surface temperature. This effect is known as ‘positive water vapor feedback’. Its existence and size however, have been contentiously argued for several years [Cubasch and Cess, 1990; Cess et al., 1990; Lindzen, 1990; Schneider et al., 1999; Hall and Manabe, 1999; Forster and Collins, 2004].

[4] Unexpected large temperature and humidity rises over the past two decades were recently observed in central Europe [Philipona and Dürr, 2004]. While increasing temperature and absolute humidity are consistent with measured surface radiative forcings (changes of radiative fluxes over time), they are two to three times larger than predicted by general circulation models (GCM) [Philipona et al., 2004]. The main question therefore is, whether the rapid increases are in fact heightened by positive water vapor feedback.

[5] Here we give observational evidence, that in Europe large-scale weather patterns uniformly but weakly influence annual mean temperature by advection on the one hand, whereas radiative forcings in conjunction with non-uniform strong atmospheric water vapor build-up dominate temperature changes on the other hand. Surface temperature is analyzed using the monthly average records of the University of East Anglia's Climate Research Unit (CRU) [Jones and Moberg, 2003]. The strong relation between surface temperature and atmospheric integrated water vapor is evidenced with ERA-40 reanalysis data from the European Center for Medium-Range Weather Forecasts (ECMWF) [Simmons and Gibson, 2000; Bengtsson et al., 2004]. To evidence the physical coherences, temperature and humidity are contrasted and correlated to monthly changes of surface radiative forcings measured in the Alpine Surface Radiation Budget (ASRB) network [Philipona et al., 1996; Marty et al., 2002] in Central Europe. We used measurements from the six ASRB stations Locarno-Monti (370 m), Payerne (490 m), Davos (1610 m), Cimetta (1670 m), Weissfluhjoch (2690 m) and Jungfraujoch (3580 m), which are distributed over an area of about 200 by 200 km square in the Alps.

2. Non-Uniform Temperature Changes in Europe

[6] Surface temperature in Europe is analyzed from 10°W to 20°E and 35°N to 60°N, using the 5° latitude × 5° longitude grid cells monthly averages of measured land air temperature anomalies of CRU (crutem2). An overall temperature increase is observed over the period 1980 to 2002 (Figure 1, green labels) with increases from +0.2 in western to +0.7°C decade−1 in central and eastern Europe. Considerably larger temperature changes are measured from 1995 to 2002 (Figure 1, orange labels) with rather decreasing temperatures in western regions but very large increases of more than +2°C decade−1 in eastern zones. These 1995 to 2002 temperature changes are not necessarily characteristic and would be lower if the warm year 1994 were included. However, the non-uniform large changes during this period are well suited to investigate cause and effects of temperature changes, if contrasted to radiative forcings measured at the ASRB network (only available since 1995), and ERA-40 water vapor changes in the atmosphere.

Figure 1.

Temperature changes over Europe from 1980 to 2002 (green boxes) and from 1995 to 2002 (orange boxes) in °C decade−1 averaged over 5° latitude × 5° longitude grid boxes (crutem2 data from Climate Research Unit).

3. Similar Changes of Temperature and Integrated Water Vapor

[7] More insight is gained from monthly analyses over the period 1995 to 2002. Figure 2a shows similar monthly temperature changes in the Alps (6 ASRB stations), Central Europe (CE, red solid line box, Figure 1) and Main Europe (ME, red broken line box, Figure 1), with large changes for some months and almost no changes for others. The same monthly characteristics are in fact observed over all Europe, as shown for the six longitudinal zone averages (35°N–60°N) in Figure 2b. However, the magnitude of the changes for all months is strongly increasing from west to east. While the monthly ups and downs, observed over all zones, are likely related to changing weather patterns (circulations) driving more or less warm humid air into Europe, the west east gradient of the temperature changes cannot be explained by the same cause. Also, how could westerly humid air flows only increase humidity and temperature in eastern parts of Europe?

Figure 2.

Monthly temperature- (above) and integrated water vapor (IWV) changes (below) over Europe for different grid box- and longitudinal zone averages in °C, respectively percent of the IWV average, from 1995 to 2002. (Left) Alps, central Europe (red solid line box, see Figure 1) and main Europe (red dotted line box, see Figure 1). (Right) longitudinal average temperature respectively IWV changes from 10W to 15E (with 10W covering the area from 10W to 5W and from 35N to 60N).

[8] ERA-40 reanalysis data of atmospheric integrated water vapor (IWV), in percent to the average from 1995 to 2002 in the Alps, central and main Europe (Figure 2c) show very similar characteristics as changes of the CRU temperature (Figure 2a), at least for the first seven month (ERA-40 reanalysis data are only available up to August 2002 - for the last four months of 2002 we used the monthly averages of the three preceding years). Changes of IWV per month for the averages over longitudinal zones (Figure 2d) also show the same pattern as the temperature (Figure 2b) and the general increase for all months going from western towards eastern zones is likewise observed in the IWV as in the temperature. A correlation coefficient of r = 0.84 is found between CRU temperature and ERA-40 IWV for annual mean changes on all 5° × 5° grid cells analyzed over Europe.

4. Radiative Forcings Correlate With Temperature and Humidity Changes

[9] For central Europe (Alps) temperature changes from 1995 to 2002 are contrasted to absolute humidity changes and radiative flux changes (radiative forcings) measured over the same period at six ASRB stations. Figure 3a shows the radiative forcing [W m−2] for each month, that is due to changes of the longwave cloud effect (LCE) [Philipona and Dürr, 2004] contrasted to temperature changes (T) [°C], whereas radiative forcings from shortwave net radiation (SNR) and longwave downward radiation (LDR) are shown in Figure 3b. Positive LCE forcing, which means increasing cloudiness, like in February, March or April, involves a negative SNR forcing (shortwave cooling) but a positive LDR forcing (longwave warming). Negative LCE forcings, like in June, have inverse effects. Forcings over annual means given in the graphs with the 1σ standard deviation in parenthesis show a significant increase of LDR and a decrease of SNR [Philipona and Dürr, 2004]. The sum of SNR and LDR is shown in Figure 3c and is called total absorbed radiation (TAR). The forcing of TAR represents the total incoming radiative energy available to change surface temperature and latent- and sensible heat fluxes. Good correlation is observed between T and TAR, with a correlation coefficient r = 0.80 on the annual mean. However, much higher correlation of r = 0.99 is found between T and the longwave downward radiation for cloud-free situations (LDRcf) (Figure 3d). LDRcf is the difference between LDR and LCE and its similarity to TAR (difference of annual means 0.3 Wm−2) demonstrates, that cloud effects of SNR and LDR more or less cancel at mid latitudes [Marty et al., 2002] and that the increasing LDRcf is by far the dominant forcing to increase surface temperature.

Figure 3.

Monthly radiative forcings, temperature and absolute humidity changes over the time period 1995 to 2002 in the Alps. a) Monthly changes of longwave cloud effect (LCE) in Wm−2 contrasted to the temperature changes (T) in °C over the time period. b) shortwave net radiation (SNR) and longwave downward radiation (LDR). c) total absorbed radiation (TAR). d) longwave downward radiation for cloud-free situations (LDRcf). e) longwave downward radiation for cloud-free situations and temperature subtracted (LDRcf,Ts), and humidity subtracted (LDRcf,Ts,Us) versus absolute humidity for cloud-free situations (Ucf). Numbers on the right (i.e. T yr: +0.8(0.4) °C) represent changes on the annual means over the period, and 1σ standard deviation in parenthesis.

[10] The annual mean LDRcf increase of +3.9(2.0) W m−2 is due to increasing greenhouse gases and increasing temperature. The forcing that is due to increasing greenhouse gases can be isolated by subtracting the effect of surface temperature rises from LDRcf. Using the first derivative of the Stefan-Boltzmann law we subtract for each month separately, the longwave radiation due to variations of temperature in each year (residuals to the linear regression) and due to temperature trends over the time period. An increase of +1.18(0.7) W m−2 remains on the annual mean for the cloud-free and temperature subtracted longwave downward radiation (LDRcf,Ts) (Figure 3e). The good correlation between surface absolute humidity [g m−3] under cloud-free situations (Ucf) and LDRcf,Ts of r = 0.89 manifests, that the dominant part of the LDRcf,Ts forcing is due to water vapor increase. Sensitivity values [Philipona et al., 2004] of 0.56 and 1.73 W m−2 at 500 respectively 3000 meters a.s.l., for a 0.1 g m−3 change of absolute humidity at the surface (gradual decrease assumed in the first 4 km), allow to subtract the water vapor forcing from LDRcf,Ts, and hence isolate the part that is just due to anthropogenic greenhouse gases. A remaining annual mean anthropogenic forcing of +0.35(0.4) Wm−2 for cloud-free, temperature subtracted and humidity subtracted longwave downward radiation (LDRcf,Ts,Us) is shown in Figure 3e.

5. Discussion and Conclusions

[11] Stand-alone MODTRAN radiative transfer model calculations show a +0.26 Wm−2 annual mean longwave downward forcing for the 12 ppm CO2 and other greenhouse gas increases in Europe from 1995 to 2002, apart from water vapor. The experimentally determined LDRcf,Ts,Us is quite close to this expected forcing at least for eight of the months and the annual mean. Hence, even though the change of the annual mean of LDRcf,Ts,Us barely reaches the 1σ confidence level, anthropogenic greenhouse forcing can be experimentally observed. Higher statistical significance is reached for the measured forcing of all greenhouse gases LDRcf,Ts, with +1.18(0.7) Wm−2, where the dominant water vapor forcing accounts for 70% or 2.4 times the anthropogenic forcing.

[12] Regarding the nature of the water vapor changes we already showed (Figure 2), that strong month-to-month changes, which are uniform over all Europe, are likely driven by air advection of changing large-scale circulations. However, the strong increase from west to east, which is observed for all months but dominant in the first part of the year, must be related to the abundance of water available at the surface for evapotranspiration. Radiative forcings over water rich surfaces apparently produce strong positive water vapor feedback, as observed in the ERA-40 data in central and northeastern Europe with integrated water vapor and temperature simultaneously increasing (Figure 4). Little changes are observed over the Atlantic Ocean, where radiative forcings are more efficiently absorbed. Decreasing temperatures over the Iberian Peninsula indicate lack of water and a drying out primarily in the second part of the year, resulting in a decreased water vapor greenhouse effect.

Figure 4.

Temperature- (left) and integrated water vapor (right) changes in °C, respectively percent, from 1995 to 2002. Decreasing changes are shown in blue, increasing changes in red (ERA-40 reanalysis data from ECMWF).

[13] While aerosols show small changes in higher elevations they are rather decreasing over the investigated time period in lowlands of central and north-eastern Europe [Weller et al., 2000]. Hence, solar radiation decrease measured at the surface is due to increasing water vapor absorption and the increasing cloud amount. Cloud forcings however, more or less cancel between shortwave and longwave radiative effects in mid latitudes. The strong increase of longwave downward radiation is shown to be due to increasing cloudiness, rising temperature, rising water vapor and due to the driving cause: the long-lived anthropogenic greenhouse gases. Greenhouse gas increases are accurately monitored and expected forcings are known with a high level of scientific understanding [Houghton et al., 2001]. With ASRB measurements we show for the first time that the different forcings can be separated and that anthropogenic greenhouse forcing is measurable, even though for the short time period it is not statistically significant. Hence, radiation measurements in central Europe and ERA-40 reanalysis data for all Europe demonstrate greenhouse warming and manifest a two to three times larger positive water vapor feedback than predicted by GCM calculations, in zones where sufficient surface water is available for evapotranspiration.


[14] This work was supported by the framework of the National Center of Competence in Research on Climate (NCCR Climate), an initiative funded by the Swiss National Science Foundation (NSF). We thank the Swiss Federal Office for Meteorology and Climatology (MeteoSwiss) for providing temperature and humidity data and valuable help at the ASRB stations. The authors also wish to thank Chris Hoyle from the Observatory at Davos and co-workers from MeteoSwiss for valuable help with the extraction of ERA-40 data.