Skip to the main content

Original scientific paper

Multiple Linear Regression (MLR) model simulation of hourly PM10 concentrations during sea breeze events in the Split area

Tanja Trošić Lesar orcid id ; Meteorological and Hydrological Service Zagreb, Croatia
Anita Filipčić ; Department of Geography Faculty of Science University of Zagreb, Croatia

Full text: english pdf 3.556 Kb

page 77-85

downloads: 651



The main objective of this study is to simulate the daily change of the particulate matter (PM10) concentrations using a multiple linear regression (MLR) model for the selected sea breeze cases in Split and Kaštel-Sućurac, which are situated in the central part of the Eastern Adriatic coast, for the period from 2007 to 2009. A new predictor characteristic to the daily and nightly part of the coastal circulation is included in the MLR model. The MLR model results well match the PM10 concentration measurements during the selected sea breeze cases in Split (with the correlation coefficient R2=0.77 and the index of the agreement IA = 0.89) and Kaštel-Sućurac (R2=0.65, IA =0.81), and they also match the measurements during land breeze (for Split R2=0.66, IA=0.86 and for Kaštel-Sućurac R2=0.66, IA=0.70).


PM10 concentrations, sea breeze, MLR model, Split area

Hrčak ID:



Visits: 1.370 *