Comparison of GCE-CRM and SCM-CAM5 with ARM data

Poster PDF

Description

The sensitivity of the changes of the cloud liquid water path (LWP) in response to the changes in aerosol concentrations is studied in a cloud resolving model (GCE-CRM) and a single column model of CAM5 (SCAM) using the forcing data derived from the Midlatitude Continental Convective Clouds Experiment (MC3E). We picked two days with low-level stratus clouds during the campaign period: 05/13/2011 with the background surface condensation nuclei (CN) varying from ~400 to ~800 #/cm3 and 05/27/2011 with the background surface CN varying from ~4000 to ~8000 #/cm3. On the low CN number day (05/13/2011), both models captured the evolution of the low-level clouds during the day with the CRM underestimating LWP while SCAM overestimates the LWP as compared to the retrieved LWP. Both models show increased LWP with increased CN with SCAM having a higher sensitivity (dlnLWP/dlnNa=0.20) than the CRM (dlnLWP/dlnNa=0.10). However, the CRM shows increased precipitation rate with increased CN (dlnPrec/dlnNa=0.24) while SCAM shows decreased precipitation rate with increased CN (dlnPrec/dlnNa=-1.0). On the high CN number day (05/27/2011), both models showed a delayed onset of the clouds as compared to the observation. With increased CN, the CRM shows decreased LWP and slightly increased precipitation rate. However, the SCAM shows increased LWP and precipitation is completely suppressed. The opposite responses of LWP to the increased CN in a high aerosol loading background in the two models may be explained by the different complexities of the two models with the SCAM lacking the entrainment-evaporating mechanism in its sub-grid parameterization. The opposite responses of the precipitation rate in the two models require further analysis.