标题：Fine scale modeling of wintertime aerosol mass, number, and size distributions in central California
作者：Zhang, Yang; Liu, Ping; Liu, Xiao-Huan; Pun, Betty; Seigneur, Christian; Jacobson, Mark Z.; Wang, Wen-Xing
作者机构：[Zhang, Yang; Liu, Ping; Liu, Xiao-Huan] N Carolina State Univ, Dept Marine Earth & Atmospher Sci, Raleigh, NC 27695 USA.; [Pun, Betty; Seigneur, Ch 更多
通讯作者地址：[Zhang, Y]N Carolina State Univ, Dept Marine Earth & Atmospher Sci, 1125 Jordan Hall,Campus Box 8208,2800 Faucette Dr, Raleigh, NC 27695 USA.
来源：JOURNAL OF GEOPHYSICAL RESEARCH-ATMOSPHERES
摘要：In light of nonattainment of PM2.5 in central California, the CMAQ-MADRID 1 model is applied to simulate PM2.5 mass, number, and size distributions observed during the California Regional PM10/PM2.5 Air Quality Study (CRPAQS) winter episode of 25-31 December 2000. The simulations with 12 and 24 size sections at a horizontal grid resolution of 4 km reproduce well the 24 h average mass concentrations of PM2.5 (with normalized mean biases (NMBs) of -6.2% to 0.5%), but with larger biases for organic matter, nitrate, and elemental carbon (with NMBs of -67% to 40.2%) and a weaker capability of replicating temporal variation of PM2.5 and its components. The coagulation process leads to a 40%-91% reduction in simulated PM2.5 number concentrations. The 24 section simulation with coagulation shows the best agreement with the observed PM number and size distributions (with an NMB of -13.9%), indicating the importance of coagulation for predicting particle number and the merits of using a fine particle size resolution. Accurately simulating PM2.5 number and size distributions continue to be a major challenge, due to inaccuracies in model inputs (e. g., meteorological fields, precursor emissions, and the initial size distribution of PM emissions and concentrations), uncertainties in model formulations (e. g., heterogeneous chemistry and aerosol formation, growth, and removal processes), as well as inconsistencies and uncertainties in observations obtained with different methods.