Evaluation of cloud properties in atmospheric models using cloud-scale observations

Long-term continuous cloud radar measurements and satellite retrievals of cloud properties were used to evaluate cloud resolving models (CRMs) and Single-Column models (SCMs) in new ways that provide more reliable conclusions.; Cirrus cloud properties from a cloud radar-IR radiometer retrieval were used to evaluate the cirrus cloud properties in a 29-day simulation of continental convective cloud systems performed by a 2D CRM. The CRM reproduced most of the cirrus properties revealed by the observations, except that the CRM's cirrus clouds are physically thicker.; The cloud radar-IR radiometer retrievals and results from the CRM simulation were used to evaluate the cirrus properties simulated by a SCM based on the National Centers for Environmental Prediction (NCEP) global model. Synthetic SCM subgrid-scale cloud fields were generated to obtain cloud quantities on spatial scales comparable to those of cloud radar observations. The ice water path (IWP) and ice water content (IWC) in the SCM thin cirrus clouds are too large, and the SCM's IWC decreases with cloud physical thickness, which is opposite to the observations. These differences are related to the detrainment and microphysical processes in the SCM.; Pixel-level satellite cloud products were used to evaluate the occurrence frequencies and cloud radiative forcings (CRFs) of various cloud types in the CRM simulation. The CRM cloud types were determined using the ISCCP (International Satellite Cloud Climatology Project) simulator program. The CRM reproduced the 3 observed predominance and relative distribution of high cloud types but it underestimated high cloud amounts. The CRM thick high cloud amount temporally correlated better with the satellite observations than did the thin high cloud amounts. The CRM's SW and LW CRFs are −31 and 31 W/m 2, respectively, compared to the observed −43 and 38 W/m 2.; Compared to cloud radar observations, the satellite retrieved cloud-top heights were too low for daytime optically thin clouds while the CRM simulated a similar cloud-top height frequency distribution. The satellite-retrieved IWP in overcast single-layer thin cirrus was the same as the cloud radar-IR radiometer retrieval at the 90% confidence level. The CRM-simulated daytime cloud water and ice path was between those from two different satellite retrievals.