Analysis Correlation of Meteorological Factors with PM_2.5 Concentrations in Forecasting Air Quality of The City of Jakarta

Data monitoring Air Quality Index (AQI) indicates that Jakarta ranks among the cities with the poorest air quality in the world in 2023. This measurement index is primarily due to the high concentration of 2.5 Micron Particles (PM2.5) in the ambient air. This decline in air quality is also evidenced by the brownish grays haze visible in Jakarta's sky over several weeks. Generally, two factors influence this phenomenon: emission sources and dispersion processes due to Jakarta's ambient meteorological conditions. Assuming a uniform and constant emission source factor, this study analyses the correlation of meteorological factors on the stability of PM2.5 concentration in the atmosphere. The meteorological data used is from the BMKG Kemayoran Jakarta from 2022 to 2023, including rainfall, temperature, humidity, and wind speed. Statistically, these meteorological data are tested for correlation with PM2.5 quality using Spearman correlation analysis, and trend data analysis is also conducted. The study results show an increase in PM2.5 and temperature in Jakarta and a decrease in rainfall and air humidity from 2022 to 2023. Furthermore, changes in PM2.5 concentration are meteorologically influenced by temperature, humidity, and rainfall factors, with correlation factors of 0.55, -0.55, and -0.43, respectively, according to Spearman's method analysis. This correlation relationship can subsequently be used to forecast Jakarta's ambient air PM2.5 quality with a regression equation. PM2.5 = -60.5 + 1.658 tropospheric ozone + 5.64 temperature - 0.495 humidity - 0.246 rainfall Where if some meteorological parameter or air quality parameter can be reduced, it can prevent the occurrence of photochemical smog phenomena. The model has a determination value of 0.99, which means that the model is quite good at making predictions, but the model still has a high RMSE value of 22.95, so the model still needs improvement by adding parameters or other prediction models.