M. Webb á C. Senior á S. Bony á J.-J. Morcrette Combining ERBE and ISCCP data to assess clouds in the Hadley Centre, ECMWF and LMD atmospheric climate models Received: 23 January 2000 / Accepted: 24 January 2001 Abstract This study compares radiative ¯uxes and cloudiness ®elds from three general circulation models the HadAM4 version of the Hadley Centre Uni®ed model, cycle 16r2 of the ECMWF model and version LMDZ 2.0 of the LMD GCM), using a combination of satellite observations from the Earth Radiation Budget Experiment ERBE) and the International Satellite Cloud Climatology Project ISCCP). To facilitate a meaningful comparison with the ISCCP C1 data, values of column cloud optical thickness and cloud top pres- sure are diagnosed from the models in a manner con- sistent with the satellite view from space. Decomposing the cloud radiative eect into contributions from low- medium- and high-level clouds reveals a tendency for the models' low-level clouds to compensate for underesti- mates in the shortwave cloud radiative eect caused by a lack of high-level or mid-level clouds. The low clouds fail to compensate for the associated errors in the longwave. Consequently, disproportionate errors in the longwave and shortwave cloud radiative eect in models may be taken as an indication that compensating errors are likely to be present. Mid-level cloud errors in the mid-latitudes appear to depend as much on the choice of the convection scheme as on the cloud scheme. Con- vective and boundary layer mixing schemes require as much consideration as cloud and precipitation schemes when it comes to assessing the simulation of clouds by models. Two distinct types of cloud feedback are dis- cussed. While there is reason to doubt that current models are able to simulate potential `cloud regime' type feedbacks with skill, there is hope that a model capable of simulating potential `cloud amount' type feedbacks will be achievable once the reasons for the remaining dierences between the models are understood. 1 Introduction Clouds play an important role in climate, not only be- cause they signi®cantly modify the distribution of the shortwave and longwave radiation absorbed and emitted by the Earth, in turn aecting temperature, humidity and the general circulation, but also because they are themselves sensitive to these changes. Consequently, clouds have the potential to modify the response of the climate system to anthropogenic forcing through cloud radiative feedbacks. Dierences between cloud feed- backs in climate models are a major contributing factor to the uncertainty in the models' responses to climate forcing e.g. Roeckner et al. 1987; Le Treut and Li 1991; Senior and Mitchell 1993). Senior 1998) points out that while many studies e.g. Cess et al. 1990) of cloud feedbacks in models focus on changes in cloud radiative forcing Harrison et al. 1990), the dierences between the responses of dierent models can only be understood when changes in cloud proper- ties such as the 3-D distribution of cloud amount, in- cloud albedo, in-cloud emissivity) are also examined and related to dierent physical assumptions in the models. If we are to have con®dence in predictions from climate models, a necessary although not sucient) requirement is that they should be able to reproduce the observed present-day distribution of clouds and their associated radiative ¯uxes. Again, an unambiguous link to the physical assumptions in models can only be made if the validation goes beyond that of the models' simu- lation of radiative ¯uxes at the top-of-the-atmosphere, to assess the various aspects of the cloud simulations that control these ¯uxes, those at the surface and also Climate Dynamics 2001) 17: 905±922 Ó Springer-Verlag 2001 M. Webb &) á C. Senior Hadley Centre for Climate Prediction and Research, The Met Oce, London Road, Bracknell, Berkshire RG12 2SY, UK E-mail: mjwebb@meto.gov.uk S. Bony MIT & LMD/CNRS, Department of Earth, Atmospheric and Planetary Sciences, Room 54-1721, 77 Massachusetts Avenue, Cambridge, MA 02139, USA J.-J. Morcrette European Centre for Medium-range Weather Forecasts, Reading, Berkshire RG2 9AX, UK