DEVELOPMENT OF LAND-USE REGRESSION MODELS FOR PARTICULATE MATTER ESTIMATION IN NATIONAL CAPITAL REGION, PHILIPPINES

Abstract

Regression models are commonly used to estimate unknown variables, such as environmental parameters. Multiple Linear Regression (MLR) is one of the techniques used to model air quality and measure air pollutant concentrations. Specifically, a technique called Land-Use Regression (LUR) enables the user to generate air pollutant models using geographical layers as input parameters. The study aims to generate models for fine and coarse particulate matter (PM2.5 and PM10, respectively) using LUR for the National Capital Region in 2019. Independent variables considered in this study are road network, traffic count, Normalized Difference Vegetation Index (NDVI), population density, and elevation. The final model results showed significant estimates based on the model parameters. For PM2.5, the model resulted in high values of R2 and adjusted R2 and an RMSE of 0.77 μg/m3. For PM10, model parameters showed that the generated final model for PM10 was significant with a 55% R2 value. Maps were then generated using the final LUR models of PM2.5 and PM10. The models can be improved by adding more types of input variables and longer observation periods.