Development of an empirical model to predict nitrogen dioxide concentrations from weather variables for sites across the UK

The prediction of nitrogen dioxide (NO2) concentrations from simple and available data such as meteorological records accurately over time steps shorter than 1 yr would be of great benefit. Predictions of monthly: or more frequent, concentrations of NO2 would be especially valuable when calculating nitrogen deposition to vegetation. as concentrations of NO2 can fluctuate through the year. A model that predicts concentrations of NO2 (mug m(-3)) on a fortnightly time step at specific sites using mean fortnightly ambient air temperatures and a local annual concentration of NO2 from the nearest representative monitoring site has been developed. The model was tested at a variety of sites across the UK and showed a reasonable pattern of agreement with measured values (R-2 from 0.45 to 0.61) for the majority of sites. The model could be used to predict deposition of NO2-N from commonly recorded data sets. within the limits that have been tested. (C) 2004 Elsevier Ltd. All rights reserved.