Evaluating deep learning time series models for PM2.5 forecasting across diverse horizons

The study area is located at Chengdu, Sichuan Province , encompassing five air quality monitoring stations and two meteorological monitoring stations that provide air quality and meteorological data, respectively. The five air quality monitoring stations consist of Shilidian station (1432A), Shahepu station (1434A), Renmin Park station (1437A), Dashi West station (2880A), Sanwayao station (1433A). Datasets for NO2, CO, SO2, O3, and PM2.5 are collected from the five air quality stations mentioned above, ranging from November 1st 2022 to October 31th 2024. The two meteorological stations, Wenjiang station and Shuangliu station, provided datasets of temperature (T), pressure (P), relative humidity (RH), precipitation (Prec) and wind speed (WS) are collected over the same period. Average values from five air quality stations and the two metrological stations were used to represent the overall spatial quality and meteorological data for the study area, respectively.