Regresar

Nonparametric Geostatistical Prediction of Daily PM<inf>2.5</inf> Concentrations Based on Satellite Measurements of Aerosol Optical Depth

Abstract:

The prediction of particulate matter (PM2.5 ) has been a widely studied topic, and most of the techniques proposed are based on correlating the concentrations of this variable, obtained by sensors on the ground, with satellite measurements of Aerosol optical depth (AOD). In the present work, a non-parametric geostatistical method based on the local linear estimator is proposed in order to obtain flexible predictions in positions not observed by the monitoring network. This procedure was applied to real daily data, and the results obtained by this approximation were compared with those values derived from models based on linear regression and parametric kriging for a validation sample.