Article Meteorologische Zeitschrift, Vol. 18, No. 6, 585-596 (December 2009) c by Gebrüder Borntraeger 2009 Intercomparison of radiative forcing calculations of stratospheric water vapour and contrails G UNNAR M YHRE1,2 ∗ , M ARIA K VALEV ÅG1 , G ABY R ÄDEL3 , J OLENE C OOK3 , K EITH P. S HINE3 , H ANNAH C LARK4,5 , F ERNAND K ARCHER4 , K RZYSZTOF M ARKOWICZ6 , A LEKSANDRA K ARDAS6 , PAULINA WOLKENBERG6 , Y VES BALKANSKI7 , M ICHAEL P ONATER8 , P IERS F ORSTER9 , A LEXANDRU R AP9 and RUBEN RODRIGUEZ D E L EON10 1 Department of Geosciences, University of Oslo, Norway Research-Oslo (CICERO), Oslo, Norway Reading, UK 4 CNRM/GAME Météo France, Toulouse, France 5 current affiliation: Laboratoire d’Aérologie, Université de Toulouse, Toulouse, France 6 Institute of Geophysics, University of Warsaw, Poland 7 LSCE/IPSL, Laboratoire CEA-CNRS-UVSQ, Cedex, France 8 Deutsches Zentrum für Luft und Raumfahrt (DLR), Institut für Physik der Atmosphäre, Oberpfaffenhofen, Germany 9 School of Earth and Environment, University of Leeds, Leeds, UK 10 Manchester Metropolitan University, UK 2 Center for International Climate and Environmental 3 Department of Meteorology, University of Reading, (Manuscript received March 17, 2009; in revised form June 8, 2009; accepted July 27, 2009) Abstract Seven groups have participated in an intercomparison study of calculations of radiative forcing (RF) due to stratospheric water vapour (SWV) and contrails. A combination of detailed radiative transfer schemes and codes for global-scale calculations have been used, as well as a combination of idealized simulations and more realistic global-scale changes in stratospheric water vapour and contrails. Detailed line-by-line codes agree within about 15 % for longwave (LW) and shortwave (SW) RF, except in one case where the difference is 30 %. Since the LW and SW RF due to contrails and SWV changes are of opposite sign, the differences between the models seen in the individual LW and SW components can be either compensated or strengthened in the net RF, and thus in relative terms uncertainties are much larger for the net RF. Some of the models used for global-scale simulations of changes in SWV and contrails differ substantially in RF from the more detailed radiative transfer schemes. For the global-scale calculations we use a method of weighting the results to calculate a best estimate based on their performance compared to the more detailed radiative transfer schemes in the idealized simulations. Zusammenfassung Sieben Forschungsgruppen haben an einem Ergebnisvergleich hinsichtlich des Strahlungsantriebes von Kondensstreifen und stratosphärischem Wasserdampf teilgenommen. Dies umfasst einen Vergleich sowohl von aufwändigen Strahlungsübertragungsmodellen mit Parametrisierungsschemata aus globalen Klimamodellen als auch einen Vergleich von idealisierten und realistischen Mustern von Kondensstreifen und stratosphärischem Wasserdampf. Aufwändige Linienmodelle weichen für den langwelligen (LW) und kurzwelligen (SW) Strahlungsantrieb um etwa 15 % voneinander ab, in einem Fall beträgt die Abweichung 30 %. Da die LW und SW Komponenten sowohl für Kondensstreifen als auch für stratosphärischen Wasserdampf ein unterschiedliches Vorzeichen haben, können sich die Differenzen in den Komponenten bzgl. des Netto-Strahlungsantriebes kompensieren und verstärken. Daher sind relativ gesehen die Modellunsicherheiten beim Netto-Strahlungsantrieb am größten. Einige der Modelle, die in globalen Klimasimulationen zur Anwendung kommen, weichen für die Strahlungsantriebe von Kondensstreifen und Änderungen des stratosphärischen Wasserdampfes deutlich von den Ergebnissen aufwändigerer Strahlungsmodelle ab. Um eine beste Abschätzung für die Berechnungen des globalen Strahlungsantriebes geben zu können, wichten wir die Ergebnisse der einzelnen Modelle unter Berücksichtigung der Abweichungen, die in den idealisierten Vergleichsrechnungen im Vergleich zu den Ergebnisse der detailliertesten Modelle festgestellt wurden. 1 Introduction Human-induced climate change is caused by many radiative forcing mechanisms, due to either their influ∗ Corresponding author: Gunnar Myhre, Center for International Climate and Environmental Research-Oslo (CICERO), P.O. Box. 1129, Blindern, N-0318 Oslo, Norway, e-mail: gunnar.myhre@cicero.uio.no DOI 10.1127/0941-2948/2009/0411 ence on the thermal infrared radiation, on solar radiation, or on a combination of the two (F ORSTER et al., 2007). Radiative transfer codes are generally adopted for estimating the radiative forcing, which introduces an uncertainty. Many intercomparison exercises have been performed to identify and quantify the uncertainty in such calculations (B OUCHER et al., 1998; C OLLINS 0941-2948/2009/0411 $ 5.40 c Gebrüder Borntraeger, Berlin, Stuttgart 2009 586 G. Myhre et al.: Stratospheric water vapour and contrails et al., 2006; E LLINGSON et al., 1991; F ORSTER et al., 2005; F ORSTER et al., 2001; G OHAR et al., 2004; M EERK ÖTTER et al., 1999; S HINE et al., 1995). For changes in the atmospheric constituents that affect both thermal infrared radiation and solar radiation, their contributions may have the same sign (e.g. changes in ozone in the troposphere) but it is more challenging if their contribution gives rise to radiative forcings of opposite sign. Here we will investigate two radiative forcing mechanisms that have positive longwave (LW) radiative forcing (RF) and negative shortwave (SW) RF for an increase in their abundance in the atmosphere, namely water vapour in the stratosphere and contrails. Since the net radiative forcing is the residual of LW and SW radiative forcings, the uncertainties in the net forcing may be particularly large, especially if the LW and SW components are of a similar magnitude. Changes in stratospheric ozone and mineral dust are examples of other radiative forcing mechanisms that have opposite signs for the LW and SW forcings. The problem is also relevant to the direct aerosol component that has an opposing forcings for the absorbing and scattering SW components (S CHULZ et al., 2006). Water vapour emitted by aircraft perturbs background concentrations of water vapour in the atmosphere, and can lead to the formation of contrails (S AUSEN et al., 2005; S CHUMANN, 2005). It is in the stratosphere where increase in water vapour from aircraft traffic may be significant because the background concentration is low and removal is slow (G AUSS et al., 2003). It is likely that stratospheric water vapour (SWV) has increased over recent decades, but the magnitude of the trend and causes for the trend are uncertain (S CHERER et al., 2008). Oxidation of methane in the stratosphere and a climate change feedback are the most probable cause for the observed increase in SWV (H ANSEN et al., 2005). Model simulations indicate that the SWV abundance responds to changes in the upper tropospheric and lower stratospheric temperature, such as can be caused by aerosols from volcanic eruptions and by changes in stratospheric ozone (J OSHI and S HINE, 2003; S TUBER et al., 2001). The current contribution from aircraft to the increase in SWV is likely to be small, but could be significantly larger in the future, especially should there be fleets of high-flying supersonic aircraft (S ØVDE et al., 2007). There is much uncertainty in the radiative forcing of contrails and SWV from aircraft traffic (S AUSEN et al., 2005). This is due to many factors such as uncertainties in the amount and spatial distribution of the water vapour emitted and the amount, spatial distribution and properties of contrails formed, and to uncertainties in the radiative transfer calculations themselves. In this study we concentrate on quantifying the uncertainties introduced by the radiative transfer simulations. We perform idealized calculations for one vertical profile to identify differences just due to the radiation codes themselves. Thereafter, we perform global-scale calcu- Meteorol. Z., 18, 2009 lations to explore the impact of factors such as temperature, clouds, and surface albedo in addition to the radiation codes on the RF due to stratospheric water vapour and contrails. The purpose of the intercomparison is to assess the uncertainty in the RF calculations, in particular for the broad-band codes used in the climate model calculations. It is not the purpose of the paper to explain the differences, as this would be a formidable challenge, although in some cases we are able to identify which models are likely to be more reliable. 2 Models A short description of each model together with appropriate references is given in Table 1. The radiative transfer codes are divided into three groups: line-by-line (LBL) models, intermediate complexity models, and broad-band models used for global calculations either for stand-alone RF calculations or as a part of a general circulation model (GCM). Two LBL codes (with different versions of one of the LBL codes for stratospheric water vapour) are included in this study. There are several possible reasons why radiation codes may disagree. These include the number of spectral bands, the method of calculating multiple scattering, the origin of the data used to model gaseous absorption and the way that absorption data is used in the model. We note in particular that there have been many updates to the near-infrared spectral line data for water vapour in HITRAN in recent years (see e.g. C HAGAS et al., 2001; G IVER et al., 2000; ROTHMAN et al., 2005) which have led to a systematic increase in the calculated absorption. These factors mean that models with higher spectral resolution are not necessarily more reliable, and this especially applies to the differences between intermediate and broadband models, where the transmittance calculation has to be parameterized in some way. 3 Intercomparison setup Calculations with the mid-latitude summer single profile were performed for various solar zenith angles and surface albedos. These calculations acted as a reference. The mid-latitude summer profile with a solar zenith angle of 30 degrees and a surface albedo of 0.2 was selected for the rest of the single column calculations and perturbations to stratospheric water vapour, and contrails were then introduced. The concentration of stratospheric water vapour was fixed at 3.0 ppmv and then at 3.7 ppmv and the RF due to this change was calculated. For contrails, the optical depth was varied between 0.1, 0.3, and 0.52 with no wavelength dependence. We span this range because of uncertainties in the actual contrail optical depth. To ensure that differences in single-scattering properties of contrail particles did not influence the intercomparison, the asymmetry Meteorol. Z., 18, 2009 Table 1: Model descriptions. G. Myhre et al.: Stratospheric water vapour and contrails 587 588 G. Myhre et al.: Stratospheric water vapour and contrails Meteorol. Z., 18, 2009 Table 2: LW, SW, and net RF for global SWV changes (in Wm−2 ). LW and net RF shown for both instantaneous and stratospheric temperature adjusted RF. SWV 3.0 – 3.7 ppmv LW inst LW adj SW Net inst Net adj UiO_BBM 0.41 0.29 -0.058 0.35 0.23 UoR_NBM 0.40 0.27 -0.020 0.38 0.25 UoR_E-S 0.52 0.29 -0.030 0.49 0.26 ECHAM4/ATT 0.41 0.25 -0.051 0.35 0.20 ECHAM4/SLT 0.25 0.19 0.16 UW_FU 0.37 CNRM_ARPEGE 0.31 UoL_E-S 0.49 0.40 -0.034 0.22 -0.052 0.32 -0.055 0.26 -0.021 0.49 0.38 Aircraft 2050 LW inst LW adj SW Net inst Net adj UiO_BBM 0.058 0.062 -0.011 0.047 0.050 UoR_NBM 0.081 0.052 -0.005 0.076 0.047 ECHAM4/ATT 0.056 0.043 -0.008 0.047 0.035 ECHAM4/SLT 0.044 0.035 -0.007 0.037 0.028 UW_FU 0.059 -0.009 0.049 CNRM_ARPEGE 0.044 -0.009 0.035 factor was specified to be wavelength independent (0.8) and the single scattering albedo was 1.0 in the solar spectrum and 0.6 in the terrestrial spectrum. These are typical values for contrail optical properties (S TRAUSS et al., 1997). Simulations were performed for a 100 % contrail cover, with the contrails between 10 and 11 km height. Instantaneous LW, SW, and net RF are reported at the tropopause (179 hPa). Global calculations for the change in stratospheric water vapour from 3.0 to 3.7 ppmv and for change due to emissions by sub and supersonic aircrafts were performed. In these simulations SWV is fixed to these abundances and no feedback from the temperature change from SWV is taken into account. The aviation water vapour increase has been derived from water vapour emissions for 2050 generated by the SCENIC project (S ØVDE et al., 2007). Radiative forcing (instantaneous as well as with stratospheric adjustment) at the tropopause is reported. For the intercomparison of the effect of contrails on the global scale, a 1 % homogeneous cover with a contrail top at 11 km was used. The contrails had an optical depth of 0.3, and the other optical properties were similar to the single profile exercise. Radiative forcing at the top of the atmosphere is reported. Each modelling group used their own data for temperature, clouds, water vapour in the troposphere, surface albedo, and tropopause altitude for the globalscale calculations. 4 Results 4.1 Single profile cases 4.1.1 Stratospheric water vapour The results for the change in SWV from 3.0 to 3.7 ppmv are presented in Figure 1. For the net RF, the difference among the models is more than a factor of two, but the separate LW and SW RF also differ by this same factor. Differences between the net RF for the LBL models are small, but that is due to a compensation of larger differences for the LW and SW RF, particularly the latter. The method of integration to irradiances has little influence on the results for the two versions of the UoR LBL codes. The stronger short-wave forcing in the UoR LBL is likely to be due to the use of more recent spectral line data (HITRAN-2001) compared to the UiO LBL calculations (HITRAN-1992) (see section 2). To investigate the implementation of radiative transfer schemes, three types of codes are used at two different groups. The Streamer codes give identical results, and the two groups using the Fu-Liou code have identical LW results and with SW results being only slightly different. The E-S code has been implemented at UoR and UoL with identical LW RF, but with SW RF that differs by more than 50 %. This is due to differences in the absorption data in a different version of the E-S code. Three of the intermediate complex models give a weak SW RF (at least a factor of two weaker than the LBL codes) and thus a high net RF. The differences G. Myhre et al.: Stratospheric water vapour and contrails Meteorol. Z., 18, 2009 LONG WAVE for the differences. In general the deviations are larger for the broad-band models compared to the intermediate complexity models (in particular for LW) indicating a need for further investigation of the parameterizations for shortwave absorption and their dependence on the source of absorption data. UiO_LBL UoR_RFM* UoR_RFM UW_MOD UW_STR UoR_NBM IPSL_STR UiO_CCM UiO_BBM UoR_FU UoR_E−S UoL_E−S UW_FU CNRM_ARPEGE ECHAM4 0 0.1 0.2 0.3 589 0.4 0.5 0.6 0.7 4.1.2 Contrails Wm−2 SHORT WAVE UiO_LBL UoR_RFM* UoR_RFM UW_MOD UW_STR UoR_NBM IPSL_STR UiO_CCM UiO_BBM UoR_FU UoR_E−S UoL_E−S UW_FU CNRM_ARPEGE ECHAM4 −0.2 −0.18 −0.16 −0.14 −0.12 −0.1 −0.08 −0.06 −0.04 −0.02 0 Wm−2 NET UiO_LBL UoR_RFM* UoR_RFM UW_MOD UW_STR UoR_NBM IPSL_STR UiO_CCM UiO_BBM UoR_FU UoR_E−S UoL_E−S UW_FU CNRM_ARPEGE ECHAM4 0 0.1 0.2 0.3 0.4 0.5 0.6 Wm−2 Figure 1: Instantaneous long-wave (light-grey), shortwave (grey), and net (black) RF at the tropopause (179 hPa) for a change in stratospheric water vapour from 3 to 3.7 ppmv for single column conditions (see text). between the eight codes used for global-scale simulations are particularly large. The UoR-E-S, UoL-E-S, and CNRM ARPEGE deviate particularly from the rest. The UoR-E-S, UoL-E-S models show the highest RF in the LW which is not offset in the net, because of the small negative SW RF. These models thus have the largest net RF values. CNRM ARPEGE and ECHAM4 use a LW scheme with few broad bands and have been optimized for computationally-fast calculations in GCMs, which may lead to absorption band combinations which are unfavourable for representing SVW RF (F ORSTER et al., 2001). Both models produce a weak LW RF, but as the negative SW RF from ECHAM4 is also weak, the CNRM APEGE model comes out with the lowest value of all for the net RF. Except for the UoR E-S and UoL E-S models, the codes used for global calculations underestimate the net RF compared with the LBL models. The differences between the radiative transfer codes for SWV are generally larger than for other components. This indicates that absorption data for water vapour, particularly in the stratosphere, is probably the main cause The second part of the intercomparison concerned the addition of contrails to the single-profile cases. The properties of the contrails were described in section 3. Figure 2 shows the RF due to contrails for the different models and at three different solar zenith angles (30, 60, and 75 degrees) upon which the RF is known to depend substantially (M EERK ÖTTER et al., 1999; M YHRE and S TORDAL, 2001; S TUBER et al., 2006). All results shown are for a 100 % contrail cover and for a contrail optical depth of 0.3. The differences in the LW and SW components between the models are much smaller for the contrail RF than they were for the RF due to SWV. The differences are a maximum of 35 % for both LW and SW RF if the CNRM ARPEGE code is excluded. The magnitude of the SW RF for contrails is, for some solar zenith angles, more comparable to the LW RF than is the case for SWV. Therefore, the relative difference in the net RF between models can be large, despite rather small differences in SW and LW RF. Differences for the LBL codes are about 15 % for LW (smaller for SW) and the difference for the net RF is enhanced, since the deviations for LW and SW RF go in the same direction for the two codes. The sign for the net RF is still the same for the two LBL codes for the three solar zenith angles. The two Streamer codes differ somewhat for LW and SW RF, but for the net RF the deviation is small. The net RF changes from positive for a solar zenith angle of 30 degrees to negative for 75 degrees, except for CNRM ARPEGE. By contrast, the sign of the net RF at a solar zenith angle of 60 degrees varies amongst the models. CNRM ARPEGE has a high LW RF, and the SW code caused a less distinct strengthening in the RF with an increasing solar zenith angle compared to the other models, leaving the net RF significantly positive also in the 75 degree case. 4.2 Global calculations 4.2.1 Stratospheric water vapour Figure 3 shows global and annual mean LW, SW, and net RF resulting from a change in SWV from 3.0 ppmv to 3.7 ppmv, and Figure 4 shows global and annual mean LW, SW, and net RF of estimated change in SWV from sub and supersonic aircraft in 2050 (S ØVDE et al., 2007). The results shown in the figures include stratospheric temperature adjustment. Table 2 summarizes RF for instantaneous and adjusted RF. The agreement in Figure 3 590 G. Myhre et al.: Stratospheric water vapour and contrails Meteorol. Z., 18, 2009 LONG WAVE UiO_LBL UoR_RFM UW_MOD UW_STR IPSL_STR UiO_BBM UoR_FU UW_FU UoL_E−S CNRM_ARPEGE 0 5 10 15 20 25 30 35 40 45 50 −2 Wm SHORT WAVE sza = 30° NET sza = 30° UiO_LBL UiO_LBL UoR_RFM UoR_RFM UW_MOD UW_MOD UW_STR UW_STR IPSL_STR IPSL_STR UiO_BBM UiO_BBM UoR_FU UoR_FU UW_FU UW_FU UoL_E−S UoL_E−S CNRM_ARPEGE CNRM_ARPEGE −20 −18 −16 −14 −12 −10 −8 −6 −4 −2 0 −10 −5 0 5 −2 Wm 10 15 20 25 30 5 10 15 20 5 10 15 20 Wm−2 sza = 60° sza = 60° UiO_LBL UiO_LBL UoR_RFM UoR_RFM UW_MOD UW_MOD UW_STR UW_STR IPSL_STR IPSL_STR UiO_BBM UiO_BBM UoR_FU UoR_FU UW_FU UW_FU UoL_E−S UoL_E−S CNRM_ARPEGE CNRM_ARPEGE −40 −35 −30 −25 −20 −15 −10 −5 0 −20 −15 −10 −5 −2 sza = 75° sza = 75° UiO_LBL UiO_LBL UoR_RFM UoR_RFM UW_MOD UW_MOD UW_STR UW_STR IPSL_STR IPSL_STR UiO_BBM UiO_BBM UoR_FU UoR_FU UW_FU UW_FU UoL_E−S UoL_E−S CNRM_ARPEGE −50 0 Wm−2 Wm CNRM_ARPEGE −45 −40 −35 −30 −25 Wm−2 −20 −15 −10 −5 0 −20 −15 −10 −5 0 Wm−2 Figure 2: Longwave (light-grey, left), shortwave (grey, lower left), and net (black, lower right) RF at the top of the atmosphere due to contrails with optical depth 0.3. Shortwave and net RF are shown for solar zenith angles of 30 degrees, 60 degrees, and 75 degrees (see text for further details). 591 G. Myhre et al.: Stratospheric water vapour and contrails Meteorol. Z., 18, 2009 LONG WAVE LONG WAVE UiO_BBM UiO_BBM UoR_NBM UoR_NBM UoR_E−S ECHAM4/ATT ECHAM4/ATT 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 Wm−2 UoL_E−S 0 0.1 0.2 0.3 0.4 0.5 SHORT WAVE 0.6 −2 Wm UiO_BBM SHORT WAVE UoR_NBM UiO_BBM UoR_NBM ECHAM4/ATT UoR_E−S −0.02 −0.015 −0.01 −0.005 0 −2 Wm ECHAM4/ATT NET UoL_E−S −0.07 UiO_BBM −0.06 −0.05 −0.04 −0.03 −0.02 −0.01 0 −2 Wm UoR_NBM NET ECHAM4/ATT UiO_BBM −0.1 −0.06 −0.04 −0.02 0 0.02 0.04 0.06 0.08 0.1 −2 Figure 4: Global and annual mean longwave (light-grey), shortwave (grey), and net (black) RF for a global change in stratospheric water vapour from sub and super sonic aircrafts estimated for 2050 in the SCENIC project (S ØVDE et al., 2007). Stratospheric temperature adjustment is included in the simulations. UoR_E−S ECHAM4/ATT UoL_E−S −0.1 −0.08 Wm UoR_NBM 0 0.1 0.2 0.3 0.4 0.5 0.6 Wm−2 Figure 3: Global and annual mean longwave (light-grey), shortwave (grey), and net (black) RF for a global change in stratospheric water vapour from 3.0 to 3.7 ppmv. Stratospheric temperature adjustment is included in the simulations. is better than in Figure 1. The difference is more than 50 % in the net RF, but this is less than in Figure 1, even though additional factors, such as spatial variability in meteorological fields of temperature, humidity, clouds as well as stratospheric temperature adjustment, have been introduced. In Table 2, the strong reduction in the RF (LW and net) as a result of stratospheric temperature adjustment for the homogeneous SWV change from 3.0 to 3.7 ppmv can be seen in all six models that calculate both instantaneous and adjusted RF. The strongest reduction occurs in the UoR E-S and the weakest in the ECHAM/SLT and UoL E-S. The strong effect of stratospheric temperature adjustment in the UoR E-S explains the much better agreement in Figure 3 than in Figure 1 for the E-S model. The UoL E-S has a weaker effect of the stratospheric temperature adjustment than the UoR E-S. The former model has a tropopause level at 100 hPa in the tropics and 250 hPa at mid and high latitudes. However, the stratospheric temperature adjustment is included at altitudes above 250 hPa at all latitudes and this could be a source for the difference in the effect of stratospheric temperature adjustment. It can also be seen in Table 2 that UW FU and CNRM have a weaker instantaneous RF than most of the other models. ECHAM4/SLT and ECHAM4/ATT both indicate derivatives of the well-established ECHAM4 climate model, with increased vertical resolution in the upper troposphere/lower stratosphere region (L AND et al., 2002). In ECHAM4/ATT the operational semiLagrangian transport (SLT) scheme for water vapour and cloud water has been replaced by the Lagrangian advection scheme ATTILA (R EITHMEIER and S AUSEN, 2002). Using Lagrangian transport significantly reduces a distinct wet and cold bias in the extratropical lowermost stratosphere (S TENKE et al., 2008) with a number of beneficial side effects such as a lowering of the extratropical tropopause (which is located at a too high altitude in ECHAM/SLT). The impact of the difference between ECHAM4/SLT and ECHAM4/ATT is shown 592 G. Myhre et al.: Stratospheric water vapour and contrails in Table 2 to be significant and almost as large as the range between the other models, even though the radiation scheme is identical in both model versions. In the rest of the analysis we use the ECHAM/ATT model. The deviations in net RF for SWV change estimated for 2050 in Figure 4 are significant, but still smaller than shown for the single profile cases. This is partly due to a compensation of the LW and SW RF, as well as the different effects of the stratospheric temperature adjustment. For a homogeneous SWV increase, cooling will occur in the entire stratosphere; however, for water vapour changes that reach their maximum at altitudes higher than the tropopause, heating may occur just above the tropopause (M YHRE et al., 2007). Whereas UoR NBM and ECHAM4/ATT have a strong reduction in the RF due to the effect of the stratospheric temperature adjustment, the UiO BBM has a weak increase. The much weaker SW RF for UoR NBM is also consistent with the results shown in Figure 1 and 3. The SW RF results for SWV show larger differences than the LW RF. This is partly because stratospheric temperature adjustment reduces the difference in the LW calculations but also that some of the models with the weakest LW RF have not calculated the full RF. However, it seems that factors such as background humidity, temperature and clouds do not introduce an additional large uncertainty in the calculation of RF due to SWV. 4.2.2 Contrails Global and annual mean RF for a 1 % homogeneous contrail cover is shown in Figure 5 for clear and all-sky conditions. For UiO BBM, UoR FU, and UW FU the difference between the clear and all-sky RF is small, as shown in earlier studies (M YHRE and S TORDAL, 2001; R ÄDEL and S HINE, 2008; S TUBER and F ORSTER, 2007). The agreement between the five models is slightly better for the all-sky conditions than for the clear-sky conditions, because UoL E-S has a large impact of clouds on the LW RF and that UoL E-S and CNRM ARPEGE have a weak impact of clouds on the SW RF. Differences between some of the models are of similar magnitude for the clear sky condition as in Figure 2, but slightly amplified for the global simulations (e.g. UW FU versus UiO BBM). CNRM ARPEGE and UoL E-S have substantially stronger LW RF for the clear-sky conditions, and for the former model this is consistent with the findings in Figure 2. The relatively large clear sky LW RF from UoL E-S compared to results in Figure 2 could be due to several factors among them temperatures and water vapour. Figure 6 shows the geographical distribution of the annual mean all-sky RF for a 1 % homogeneous contrail cover (global mean shown in lower panel of Figure 5). Despite differences in the magnitude, the spatial pattern from the five models has many similarities with maximum values in sub-tropical regions and particularly over the Sahara. The latter is due to a maximum LW RF and Meteorol. Z., 18, 2009 weak SW RF because of the high reflectivity over the Saharan region. It should be mentioned here that contrail formation in some of the regions with maximum RF is not so likely at 10–11 km due to the dry conditions. Weak RF is simulated near the equator due to large cloud cover, in particular for UiO BBM with even negative values in a narrow region. There is also a consistent pattern between the models with rather low values at high latitudes (even negative in CNRM ARPEGE). To convert the idealized experiment with a homogeneous contrail cover over to a more realistic case, the contrail cover as estimated in R ÄDEL and S HINE (2008) is adopted. The UiO BBM, UoR FU, UW FU, and CNRM ARPEGE net RFs for a realistic contrail cover (and contrail optical depth of 0.3) are 9.3, 10, and 12, and 15 mWm−2 , respectively. This is a slightly smaller difference than shown in Figure 5 because the estimated contrail cover is small where the differences in Figure 5 are the largest (i.e. in the tropics). 5 Interpretation of results The single profile calculation can be used in a ranking of the results and for an evaluation of a best estimate for the global calculations. For the single profile cases we have the benchmark LBL calculations available; despite the fact that there remain uncertainties in the LBL codes, they still must be regarded as the most reliable. The LBL model formulation and their complexity are much closer to first principles than schemes used for global calculations. We treat therefore LBL as more accurate, but acknowledge that there is an uncertainty associated with the LBL as well. However, because of their heavy computational requirements, it is impractical to use them for global calculations. To generate a “best estimate” for the global calculations from the models contributing to the intercomparisons, a number of options are available. These include (a) simply choosing the global results from the intermediate or broad-band model that are closest to the LBL code for the single profile cases, (b) taking a simple arithmetic average of all the model results or (c) finding an appropriate weighting to combine the model results. We regard option (a) as too restrictive, as just because a model agrees with the LBL results for one profile should not be taken to imply it does so for all cases. However option (b) neglects all information from the single profile comparisons, and the degree to which the model agrees with the reference results. Therefore we adopt option (c), but also compare it with the results from option (b) For option (c) we use a weighting adopted by (M URPHY et al., 2004) and the following formula: 2 Vi = e(−0.5∗∆RFi /σ 2) where ∆RFi is the absolute difference in RF between model i and the reference (mean of LBL results), and σ LONG WAVE LONG WAVE UiO_BBM UiO_BBM UoR_FU UoR_FU UW_FU UW_FU UoL_E−S UoL_E−S CNRM_ARPEGE CNRM_ARPEGE 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0 0.05 0.1 0.15 −2 UiO_BBM UoR_FU UoR_FU UW_FU UW_FU UoL_E−S UoL_E−S CNRM_ARPEGE CNRM_ARPEGE −0.08 −0.06 −0.04 −0.02 0 −0.2 −0.18 −0.16 −0.14 −0.12 −2 0.35 0.4 −0.1 −0.08 −0.06 −0.04 −0.02 0 −2 Wm Wm NET NET UiO_BBM UiO_BBM UoR_FU UoR_FU UW_FU UW_FU UoL_E−S UoL_E−S CNRM_ARPEGE −0.1 0.3 SHORT WAVE UiO_BBM −0.1 0.25 Wm SHORT WAVE −0.18 −0.16 −0.14 −0.12 0.2 −2 Wm −0.2 593 G. Myhre et al.: Stratospheric water vapour and contrails Meteorol. Z., 18, 2009 CNRM_ARPEGE −0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 −2 Wm −0.1 −0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 −2 Wm Figure 5: Global and annual mean net radiative forcing due to 1 % homogeneous contrails for clear sky (left) and all sky (right). Separate results for longwave (light-grey), shortwave (grey), and net (black) RF are shown. is the standard deviation of the results of the single profile calculations. Results from identical models have not been double counted in the calculation of the standard deviation. The weighting Vi for each model is calculated separately for SW and LW and thereafter applied for calculation of a best estimate for SW and LW by simply multiplying the percentage weighting of each model with the RF. This method has been applied to the global SWV changes with weighting factors 36.7 %, 41.9 %, 5.4 %, 10.5 %, and 5.4 % for LW and 41.5 %, 4.4 %, 11.6 %, 3.6 %, and 38.9 % for SW for UiO BBM, UoR NBM, UoR E-S, ECHAM4, and UoL E-S, respectively. The resulting best estimate is 0.284, –0.039, and 0.245 Wm−2 for LW, SW, and net RF of SWV change from 3.0 to 3.7 ppmv. A simple average of the results (option b above) gives 0.300, -0.038, and 0.262 Wm−2 for LW, SW, and net RF. Thus the weighting has a small impact on the SW RF, but gives a weaker LW RF since the models with very high LW RF have a large deviation from the LBL models. The SW single profile calculations of contrails revealed strong solar zenith angle dependence, as did the deviations from the LBL results. To calculate the SW weighting we have averaged the results for the three solar zenith angle results. The resulting weighting factors for LW are 12.6 %, 20.8 %, 20.8 %, 28.4 %, and 17.4 %, respectively, for UiO BBM, UoR FU, UW FU, CNRM ARPEGE, and UoL E S. The SW weighting factors for the same models are 24.9 %, 26.1 %, 24.7 %, 2.7 %, and 21.6 %. Including these factors in the calculations of 1 % contrail cover results in 0.260, –0.097, and 0.163 Wm−2 , respectively for LW, SW, and net RF. The corresponding numbers for a simple average is 0.250, –0.107, and 0.143 Wm−2 . A similar deviation for LW and SW RF for the two methods adds up to a twice as 594 G. Myhre et al.: Stratospheric water vapour and contrails a) b) c) d) Meteorol. Z., 18, 2009 e) Figure 6: Geographical distribution of the annual mean all sky net radiative forcing at the top of the atmosphere for a homogeneous 1 % contrail cover, UIO BBM a), UoR FU b), UW FU c), CNRM d), UoL E-S e). large deviation for the net RF of 0.020 Wm−2 . The uncertainties in terms of standard deviations among the models considered in the above analysis show similar ranges in global calculation as in the single profile cases for SWV changes, indicating that the uncertainties associated with input parameters do not exceed the uncertainties in the use of the radiation codes. For contrails, the standard deviations among the models are slightly higher for global calculations than the single profile calculations, but smaller than for SWV changes. • Changes in SWV lead to larger differences in RF than contrails. • Results from several of the models for global calculations of changes in SWV differ substantially from detailed radiative transfer schemes both for LW and SW radiation. 6 Summary • Simulations of global changes in SWV differ by about a factor of 2 between the models, but cancelations of differences may indicate that uncertainties are larger. With varying complexity in radiative transfer schemes, we have explored the uncertainties in the RF of changes in stratospheric water vapour and contrails. Seven groups have been involved in this intercomparison study as part of the EU project QUANTIFY. Our main findings are: • Although the SW RF has a magnitude much more comparable to the LW RF for contrails than it does for SWV changes, the sign of the net RF due to contrails for various solar zenith angles is similar among most of the models Meteorol. Z., 18, 2009 G. Myhre et al.: Stratospheric water vapour and contrails • The net RF due to global changes in contrails agree within a factor of two for a homogeneous contrail cover; the agreement is even better for a realistic contrail cover. • The pattern of geographical distribution of the RF for a homogeneous contrail cover is consistent between five models, but the magnitude is significantly different. • Differences in the global distributions of input fields of temperature, humidity, clouds, and surface albedo do not introduce a substantial uncertainty compared to the uncertainty in the radiative transfer schemes in the RF for contrails and change in stratospheric water vapour. In the case of stratospheric water vapour, the choice of external parameters (like tropopause height) may require more attention, if they have a direct influence on the estimated perturbation itself. • Weighting the results based on their performance against detailed radiative transfer schemes does not lead to an average that differs substantially from a simple average, but is to be preferred since the best estimate is less influenced by outliers compared to detailed schemes. • The results, and in particular the deviations between the benchmark line-by-line models and the simpler models, indicate aspects of the simpler models that appear to be in need of attention, if their radiative forcing calculations are to be improved. This is particularly the case for absorption data for water vapour. In the solar spectrum the absorption data have been significantly updated in recent years and the absorption data in the models should be accordingly updated. In the LW spectrum the parameterization of absorption by water vapour in the stratosphere should be investigated further. Acknowledgments This study is funded by the EU FP6 project QUANTIFY. UoR would like to thank Piero C AU for valuable contributions in the early stages of the project. We thank two referees for their many useful comments. References B ERK , A., L.S. B ERNSTEIN , G.P. A NDERSON , P.K. ACHARYA , D.C. ROBERTSON , J.H. C HETWYND , S.M. A DLER -G OLDEN , 1998: MODTRAN cloud and multiple scattering upgrades with application to AVIRIS. – Remote Sens. Environ. 65, 367–375. 595 B OUCHER , O., S.E. S CHWARTZ , T.P. ACKERMAN , T.L. A NDERSON , B. B ERGSTROM , B. B ONNEL , P. C HYLEK , A. DAHLBACK , Y. F OUQUART, Q. F U , R.N. H ALTHORE , J.M. H AYWOOD , T. I VERSEN , S. K ATO , S. K INNE , A. K IRKEVAG , K.R. K NAPP, A. L ACIS , I. L ASZLO , M.I. M ISHCHENKO , S. N EMESURE , V. R AMASWAMY, D.L. ROBERTS , P. RUSSELL , M.E. S CHLESINGER , G.L. S TEPHENS , R. WAGENER , M. WANG , J. W ONG , F. YANG , 1998: Intercomparison of models representing direct shortwave radiative forcing by sulfate aerosols. – J. Geophys. Res. 103, 16979–16998. C HAGAS , J.C.S., A.A. N EWNHAM , K.M. S MITH , K.P. S HINE , 2001: Effects of improvements in near-infrared water vapour line intensities on short-wave atmospheric absorption. – Geophys. Res. Lett. 28, 2401–2404. C OLLINS , W.D., V. R AMASWAMY, M.D. S CHWARZKOPF, Y. S UN , R.W. P ORTMANN , Q. F U , S.E.B. C ASANOVA , J.L. D UFRESNE , D.W. F ILLMORE , P.M. F ORSTER , V.Y. G ALIN , L.K. G OHAR , W.J. I NGRAM , D.P. K RATZ , M.P. L EFEBVRE , J. L I , P. M ARQUET, V. O INAS , Y. T SUSHIMA , T. U CHIYAMA , W.Y. Z HONG , 2006: Radiative forcing by well-mixed greenhouse gases: Estimates from climate models in the Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report (AR4). – J. Geophys. Res. 111, D14317. D UDHIA , A., 1997. RFM v3 software user’s manual, Atmos., Oceanic, and Planet. Phys. – Clarendon Lab., Oxford, U.K. E DWARDS , J.M., A. S LINGO , 1996: Studies with a flexible new radiation code .1. Choosing a configuration for a largescale model. – Quart. J. Roy. Meteorol. Soc. 122, 689–719. E LLINGSON , R.G., J. E LLIS , S. F ELS , 1991: The intercomparison of radiation codes used in climate models – longwave results. - J. Geophys. Res. 96, 8929–8953. F ORSTER , P.M.D., K.P. S HINE , 1997: Radiative forcing and temperature trends from stratospheric ozone changes. – J. Geophys. Res. 102, 10841–10855. F ORSTER , P.M.D., M. P ONATER , W.Y. Z HONG , 2001: Testing broadband radiation schemes for their ability to calculate the radiative forcing and temperature response to stratospheric water vapour and ozone changes. – Meteorol. Z. 10, 387–393. F ORSTER , P.M.D., J.B. B URKHOLDER , C. C LERBAUX , P.F. C OHEUR , M. D UTTA , L.K. G OHAR , M.D. H UR LEY, G. M YHRE , R.W. P ORTMANN , K.P. S HINE , T.J. WALLINGTON , D. W UEBBLES , 2005: Resolution of the uncertainties in the radiative forcing of HFC-134a. – J. Quant. Spectrosc. Radiat. Transfer 93, 447–460. F ORSTER , P., V. R AMASWAMY, P. A RTAXO , T. B ERNTSEN , R. B ETTS , D. FAHEY, J. H AYWOOD , J. L EAN , D.C. L OWE , G. M YHRE , J. N GANGA , R. P RINN , G. R AGA , M. S CHULZ , R. VAN D ORLAND , 2007. Changes in atmospheric constituents and in radiative forcing. – In: S OLOMON , S., D. Q IN , M. M ANNING , Z. C HEN , M. M ARQUIS et al. (Eds.), Climate change 2007: The physical science basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change Cambridge University Press, United Kingdom and New York, NY, USA. F U , Q., K.N. L IOU , 1992: On the correlated k-distribution method for radiative-transfer in nonhomogeneous atmospheres. - J. Atmos. Sci. 49, 2139–2156. —, —, 1993: Parameterization of the radiative properties of cirrus clouds. – J. Atmos. Sci. 50, 2008–2025. G AUSS , M., I.S.A. I SAKSEN , S. W ONG , W.-C. WANG, 2003: Impact of H2 O emissions from cryoplanes and 596 G. Myhre et al.: Stratospheric water vapour and contrails kerosene aircraft on the atmosphere. – J. Geophys. Res. 108, 4304. G IVER , L. P., C. C HACKERIAN , P. VARANASI , 2000: Visible and near-infrared (H2 O)-O-16 line intensity corrections for HITRAN-96. – J. Quant. Spectrosc. Radiat. Transfer 66, 101–105. G OHAR , L. K., G. M YHRE , K.P. S HINE , 2004: Updated radiative forcing estimates of four halocarbons. – J. Geophys. Res. 109, D01109. H ANSEN , J., M. S ATO , R. RUEDY,L. NAZARENKO , A. L ACIS , G.A. S CHMIDT, G. RUSSELL , I. A LEINOV, M. BAUER , S. BAUER , N. B ELL , B. C AIRNS , V. C ANUTO , M. C HANDLER , Y. C HENG , A. D EL G E NIO , G. FALUVEGI , E. F LEMING , A. F RIEND , T. H ALL , C. JACKMAN , M. K ELLEY, N. K IANG , D. KOCH , J. L EAN , J. L ERNER , K. L O , S. M ENON , R. M ILLER , P. M INNIS , T. N OVAKOV, V. O INAS , J. P ERLWITZ , J. P ERLWITZ , D. R IND , A. ROMANOU , D. S HINDELL , P. S TONE , S. S UN , N. TAUSNEV, D. T HRESHER , B. W IELICKI , T. W ONG , M. YAO , S. Z HANG , 2005: Efficacy of climate forcings. – J. Geophys. Res. 110, D18104, DOI:10.1029/2005JD005776. J OSHI , M.M., K.P. S HINE , 2003: A GCM study of volcanic eruptions as a cause of increased stratospheric water vapor. – J. Climate 16, 3525–3534. L AND , C., J. F EICHTER , R. S AUSEN , 2002: Impact of vertical resolution on the transport of passive tracers in the ECHAM4 model. – Tellus 54, 344–360. M EERK ÖTTER , R., U. S CHUMANN , D.R. D OELLING , P. M INNIS , T. NAKAJIMA , Y. T SUSHIMA , 1999: Radiative forcing by contrails. – Ann. Geophys.-Atmos. Hydro. Space Sci. 17, 1080–1094. M ORCRETTE , J.J., 1990: Technical report 64. – European Centre for Medium-range Weather Forecasts. M URPHY, J.M., D.M.H. S EXTON , D.N. BARNETT, G.S. J ONES , M.J. W EBB , M. C OLLINS , 2004: Quantification of modelling uncertainties in a large ensemble of climate change simulations. – Nature 430, 768–772. M YHRE , G., F. S TORDAL , 2001: On the tradeoff of the solar and thermal infrared radiative impact of contrails. – Geophys. Res. Lett. 28, 3119–3122. M YHRE , G., J.S. N ILSEN , L. G ULSTAD , K.P. S HINE , B. ROGNERUD , I.S.A. I SAKSEN , 2007: Radiative forcing due to stratospheric water vapour from CH4 oxidation. – Geophys. Res. Lett. 34, L01807. R ÄDEL , G., K.P. S HINE , 2008: Radiative forcing by persistent contrails and its dependence on cruise altitudes. – J. Geophys. Res. 113, D07105. R EITHMEIER , C., R. S AUSEN , 2002: ATTILA: atmospheric tracer transport in a Lagrangian model. – Tellus 54, 278– 299. ROTHMAN , L.S., D. JACQUEMART, A. BARBE , C.C. B EN NER , M. B IRK , L.R. B ROWN , M.R., C ARLEER , C. C HACKERIAN , K. C HANCE , L. C OUDERT, H. DANA , V.M. D EVI , J.M. F LAUD , R.R. G AMACHE , A. G OLD MAN , J.M. H ARTMANN , K.W. J UCKS , A.G. M AKI , J.Y. M ANDIN , S.T. M ASSIE , J. O RPHAL , A. P ER RIN , C.P. R INSLAND , M.A.H. S MITH , J. T ENNYSON , Meteorol. Z., 18, 2009 R. T OLCHENOV, R.A. T OTH , J. VANDER AUWERA , P. VARANASI , G. WAGNER , 2005: The HITRAN 2004 molecular spectroscopic database. – J. Quant. Spectrosc. Radiat. Trans. 96, 139–204. S AUSEN , R., I.S.A. I SAKSEN , V. G REWE , D. H AUGLUS TAINE , D.S. L EE , G. M YHRE , M.O. K OHLER , G. P ITARI , U. S CHUMANN , F. S TORDAL , C. Z EREFOS , 2005: Aviation radiative forcing in 2000: An update on IPCC (1999). – Meteorol. Z. 14, 555–561. S CHERER , M., H. VOMEL , S. F UEGLISTALER , S.J. O LTMANS , J. S TAEHELIN , 2008: Trends and variability of midlatitude stratospheric water vapour deduced from the re-evaluated Boulder balloon series and HALOE. – Atmos. Chem. Phys. 8, 1391–1402. S CHULZ , M., C. T EXTOR , S. K INNE , Y. BALKANSKI , S. BAUER , T. B ERNTSEN , T. B ERGLEN , O. B OUCHER , F. D ENTENER , S. G UIBERT, I.S.A. I SAKSEN , T. I VERSEN , D. KOCH , A. K IRKEVAG , X. L IU , V. M ONTANARO , G. M YHRE , J.E. P ENNER , G. P ITARI , S. R EDDY, O. S E LAND , P. S TIER , T. TAKEMURA , 2006: Radiative forcing by aerosols as derived from the AeroCom present-day and pre-industrial simulations. – Atmos. Chem. Phys. 6, 5225– 5246. S CHUMANN , U., 2005: Formation, properties and climatic effects of contrails. – Comptes Rendus Physique 6, 549– 565. S HINE , K.P., B.P. B RIEGLEB , A.S. G ROSSMAN , D. H AUGLUSTAINE , M. H UITING , V. R AMASWAMY, M.D. S CHWARZKOPF, R. VAN D ORLAND , W.-C. WANG , 1995. Radiative forcing due to changes in ozone: A comparison of different codes. – In: WANG , W.-C., I.S.A. I SAK SEN (Eds.), Atmospheric ozone as a climate gas, SpringerVerlag Berlin Heidelberg, New York, 373–396. S TAMNES , K., S.C. T SAY, W. W ISCOMBE , K. JAYAWEERA , 1988: Numerically stable algorithm for discrete-ordinatemethod radiative-transfer in multiple-scattering and emitting layered media. – Applied Optics 27, 2502–2509. S TENKE , A., V. G REWE , M. P ONATER , 2008: Lagrangian transport of water vapor and cloud water in the ECHAM4 GCM and its impact on the cold bias. – Climate Dynam. 31, 491–506. S TRAUSS , B., R. M EERKOETTER , B. W ISSINGER , P. W ENDLING , M. H ESS , 1997: On the regional climatic impact of contrails: microphysical and radiative properties of contrails and natural cirrus clouds. – Ann. Geophys.Atmos. Hydro. Space Sci. 15, 1457–1467. S TUBER , N., P. F ORSTER , 2007: The impact of diurnal variations of air traffic on contrail radiative forcing. – Atmos. Chem. Phys. 7, 3153–3162. S TUBER , N., M. P ONATER , R. S AUSEN , 2001: Is the climate sensitivity to ozone perturbations enhanced by stratospheric water vapor feedback? – Geophys. Res. Lett. 28, 2887–2890. S TUBER , N., P. F ORSTER , G. R ÄDEL , K. S HINE , 2006: The importance of the diurnal and annual cycle of air traffic for contrail radiative forcing. – Nature 441, 864–867. S ØVDE , O. A., M. G AUSS , I.S.A. I SAKSEN , G. P ITARI , C. M ARIZY, 2007: Aircraft pollution – a futuristic view. – Atmos. Chem. Phys. 7, 3621–3632.