Inconsistencies between monthly mean longwave cloud forcing and

Inconsistencies between monthly mean longwave cloud forcing and dynamical fields from reanalyses Richard Allan ESSC, Reading University Mark Ringer Hadley Centre, Met Office

INTRODUCTION • Clouds and Climate • Advances: – Clear-sky fluxes from satellite (e. g. Ramanathan et al. 1989, Science) “Cloud Forcing” LWCF=OLRc-OLR – Sub-sampling by “dynamical regime” (e. g. Bony et al. 1995, Climate Dynamics) • This Study: see Allan and Ringer (2003) GRL 30(9)

Had. AM 3 minus ERBS (1985 -89) Clear-sky OLR differences > LWCF differences in some regions

Illustration of clear-sky sampling bias using Had. AM 3 3 -hourly data

Clear-sky OLR Composite with OMEGA and SST OMEGA (mb/day) More POSITIVE=more stable=drier/cloud-free=higher OLRc +ve w (mb/day) More LW emission OLRc -ve Higher SST More moisture OMEGA (mb/day) More NEGATIVE=less stable=moister=lower OLRc

OLRc (Wm-2)

DOLRc (Wm-2)

Using ERA-40 Daily data to illustrate clear-sky sampling bias of CERES data

Using ERA 40 clearsky OLR to evaluate dynamical regimes ERA 40 -CERES similar ERA 40 < CERES ERA 40 minus CERES clear-sky OLR (January-August 1998)

d. LWCF/d. SWCF CERES only: -1. 20 CERES/ERA 40 -1. 33

CONCLUSIONS • Model Evaluation using satellite data – Sub-sampling of dynamical regimes – Clear-sky radiation • Clear-sky radiation measurements preferentially sample more stable, drier atmospheric profiles • Subsampling by vertical motion exagerates clearsky sampling bias (~15 Wm-2) • Use of reanalysis data (ERA 40) may improve interpretation of cloud radiative effect • Satellite LWCF may overestimate degree of cancellation between LWCF and SWCF