Parameterization and comparative evaluation of the CCN number concentration on Mt. Huang, China

Abstract

Quantifying regional CCN concentration is important for reliable estimations of aerosol indirect effects. Based on observational data of the number concentrations of total aerosol (NCN) and cloud condensation nuclei (NCCN), particle number size distribution (PNSD) and, size-resolved activation ratio (SRAR) obtained on Mt. Huang in southeast China from September 19 to October 11, 2012, seven parameterization schemes are used to calculate NCCN employing CCN spectra, bulk activation ratio, cut-off diameter and SRAR. The calculations and the observations are compared and analyzed at four supersaturations (S) from 0.109% to 0.67%. Results show that (1) the parameterization using the average cut-off diameter Dm, which is derived from the various measured PNSD and NCCN, provides the best estimate of NCCN, with coefficient of determination, R2=0.70–0.90 and NCCN,cal/NCCN,obs=0.92–1.11, followed by the method of combining an average size-resolved activation curve with the PNSD, with R2=0.71–0.91 and NCCN,cal/NCCN,obs=0.71–0.91; average D50 together with the PNSD also provides a rational scheme for NCCN prediction, with NCCN,cal/NCCN,obs=0.86–0.94 and R2=0.70–0.89; (2) the method of parameterizing CCN spectra, though straightforward, has limits under polluted conditions. Reasonable NCCN estimate could only be obtained at high S (R2≥0.85 at S=0.39% and 0.67%). (3) For the method employing the bulk activation ratio ARB(S), NCCN are substantially overestimated by using total mode-based ARB(S) (NCCN,cal/NCCN,obs=0.94–1.39, R2=0.17–0.67), while applying ammonium sulfate-based ARB(S) yields improved CCN predictions (NCCN,cal/NCCN,obs=0.91–1.11, R2=0.70–0.91). In southern China, when determining the parameterization schemes in climate models, it is first recommended to use the method of average cut-off diameter or SRAR, with the various measured PNSD to predict NCCN. Besides, the method using ammonium sulfate-based ARB(S) and parameterizing CCN spectra also provides a viable way to calculate NCCN.