GIS-BASED REGRESSION ANALYSIS OF AIR QUALITY AND LAND SURFACE TEMPERATURE IN METRO MANILA AIRSHEDS

• Published in: International archives of the photogrammetry, remote sensing and spatial information sciences. remote sensing and spatial information sciences., 2024-04, Vol.XLVIII-4/W8-2023, p.77-83
• Main Authors: Broqueza, M. J., Cruz, L. M., Mana-ay, M., Jamboy, A. M., Ramos, R. V.
• Format: Article
• Language: English
• Subjects: Air; Air quality; Climate adaptation; Climate change adaptation; Clustering; Correlation; Geographical information systems; Ground stations; Land surface temperature; Mathematical models; Modelling; Nitrogen dioxide; Parameters; Regression analysis; Statistical analysis; Sulfur dioxide; Surface temperature; Urban areas
• ISSN: 2194-9034; 1682-1750; 2194-9034
• DOI: 10.5194/isprs-archives-XLVIII-4-W8-2023-77-2024

This study examines the spatial relationship of land surface temperature (LST) derived from MODIS, elevation, and air quality parameters (CO, NO2, and SO2) derived from Sentinel 5P in the Metro Manila airsheds from January 2019 to March 2023. Using Ordinary Least Squares (OLS) and Generalized Linear Regression (GLR), mean LST and heat index from the Philippine Atmospheric, Geophysical, and Astronomical Services Administration (PAGASA) ground stations exhibit a strong positive correlation, allowing the use of LST for further analysis. Across different combinations between LST, elevation, and air quality parameters, a weak to low negative correlation was seen between DEM to LST, CO, and NO2. In addition, weak to low positive correlation was seen between LST to CO and NO2. Almost no correlation was found between DEM and SO2, and LST and SO2. These results may be unreliable due to overfitting, non-stationarity, bias, or misspecification as implied by their statistical parameters. To enhance the reliability, it is recommended to investigate additional air quality parameters such as Normalized Difference Built-Up Index (NDBI) as high LST, CO, and NO2 have shown clustering in urban areas of Metro Manila. Moreover, it suggests exploring other regression modeling and methodologies, such as training and test sets, to identify the best-fit model. In conclusion, this study provides an exploratory foundation for future research and comparative assessment on using different methods for modeling these variables. This comprehensive approach enhances understanding of the complex interplay between temperature, elevation, and air quality, aiding the development of informed urban climate adaptation strategies.