Concentración Concentración de material particulado (PM2.5) en función de la humedad y reflectancia atmosférica usando imágenes landsat-8 en Lima Metropolitana, 2015 – 2016

• Autor: García Rivas Plata, Cecilia Edith
• Formato: tesis doctoral
• Fecha de Publicación: 2018
• Institución: Universidad Nacional de Trujillo
• Repositorio: UNITRU-Tesis
• Lenguaje: español
• OAI Identifier: oai:dspace.unitru.edu.pe:20.500.14414/16953
• Enlace del recurso: https://hdl.handle.net/20.500.14414/16953
• Nivel de acceso: acceso abierto
• Materia: Material particulado fino; PM2.5; Contaminación del aire; NDMI; Imagen satelital; Teledetección

The work consists of finding the mathematical model that estimates the concentration of PM2.5 particulate matter as a function of humidity with the calculation of the Normalized Index of Humidity Difference (NDMI) and the atmospheric reflectance of Landsat 8 satellite imagery in Metropolitan Lima. the years 2015 and 2016. We used 19 Landsat 8 OLI satellite images downloaded from United State Geological Survey (USGS) servers in Path 7 and Row 68 for the years 2015 and 2016. Those with the lowest possible cloud cover were considered. Observations of daily PM2.5 concentrations were obtained from seven fixed monitoring stations located in Metropolitan Lima administered by the National Meteorology and Hydrology Service of Peru (SENAMHI). The digital levels of the Landsat 8 OLI satellite images were converted to TOA Reflectance (on the roof of the atmosphere) with angular correction. Applying multiple linear regression it was found that the concentration of daily PM2.5 has a high correlation with atmospheric reflectances for bands 1-4 and NDMI. The least squares method obtained the best model that correlates the five predictor variables RB1, RB2, RB3, RB4 and NDMI with the concentration of PM2.5 in Metropolitan Lima for the years 2015 and 2016. When applying the significance test F It was found that the proposed model as a whole explains the variance of the variable PM2.5 better than expected (p-value = 4.716x10-5) whose predictors significantly favor the correlation. The significance test was also xiv performed by individual predictor by means of F-test, checking that all predictors contribute to the model, so there is a correlation of the variables atmospheric reflectances in the visible bands and NDMI with the concentration of PM2.5, obtaining a correlation R = 0.847 and coefficient of determination R2 = 0.717409.