engleski

Simulations of the concentrations of pollutant particles PM10 and PM2.5 for the sea breeze

The statistical prediction models are rarely focusing on pollutant concentrations modelling in hourly values, as well as their application in local winds. In this case a multiple linear regression (MLR) model is applied for the hourly concentrations simulation of pollution with PM2.5 and PM10 particles for sea breeze cases in Split, Croatia. A novel predictor characteristic for the daily and nightly part of the coastal circulation is found to be significant for both PM10 and PM2.5 particles for the sea breeze. The hourly simulations show very good agreement with an agreement index of IA=0.9 during the sea breeze and IA=0.8 for the land breeze. Numerical simulation with the Hybrid Single-Particle Lagrangian Integrated Trajectory model (HYSPLIT) from four point sources showed the main source of pollution in Split for the extreme case with the concentration of PM10 particles that reached the limit value for human health.

Multiple linear regression model, sea breeze, simulation, PM particles

nije evidentirano