A precise air pollution forecast is the basis for targeted pollution control and sustained improvements in air quality. It is desirable and crucial to select the most suitable model for air pollution forecasting (APF). To achieve this goal, this paper provides a comprehensive evaluation of performances of different models in simulating the most common air pollutants (e.g., PM_2.5, NO₂, SO₂, and CO) in Guangzhou (23.13° N, 113.26° E), China. To simulate temporal variations of the above-mentioned air pollutant concentrations in Guangzhou in September and October 2020, we use a numerical forecasting model (i.e., the Weather Research and Forecasting model with Chemistry (WRF-Chem)) and two artificial intelligence models (i.e., the back propagation neural network (BPNN) model and the long short-term memory (LSTM) model). WRF-Chem is also used to simulate the meteorological elements (e.g., the 2 m temperature (T2), 2 m relative humidity (RH), and 10 m wind speed and direction (WS, WD)). In order to investigate the simulation accuracies of classical APF models, we simultaneously compare the simulations of the WRF-Chem, BPNN, and LSTM models to ground truth observations. Comparative assessment results show that WRF-Chem simulated air pollutant (i.e., PM_2.5, NO₂, SO₂, and CO) concentrations have the best correlations with ground measurements (i.e., Pearson correlation coefficient R = 0.88, 0.73, 0.61, and 0.61, respectively). Furthermore, to evaluate model performance in terms of accuracy and stability, the normalized mean bias (NMB, %) and mean fractional bias (MFB, %) are adopted as the standard performance metrics (SPMs) proposed by Boylan et al. The comparison results indicate that when simulating PM_2.5, WRF-Chem was more effective than the BPNN but less effective than the LSTM. While simulating concentrations of NO₂, SO₂, and CO, the WRF-Chem model performed better than the BPNN and LSTM models. With regards to WRF-Chem, the NMBs and MFBs for the PM_2.5 simulations are, respectively, 6.49% and 0.02%, –11.96% and –0.031% for NO₂, 7.93% and 0.019% for CO, and 5.04% and 0.012% for SO₂. Our results suggest that WRF-Chem has superior performance and better accuracy than the NN-based prediction models, making it a promising and useful tool to accurately predict and forecast regional air pollutant concentrations on a city scale.

This study presents a novel, satellite-based framework for quantifying the relative contribution of regional transport in biomass burning aerosol (BBA) formation over the South China Sea (SCS). We integrate the biomass burning emission (BBE) rates from potential source regions with a random forest regression model, which is driven by backward trajectory analysis. This approach isolates and evaluates the relative contribution from transported sources. The model demonstrates robust predictive skill for BBA concentrations (R² = 0.78 on an independent test set), using only transport-weighted BBE rates and meteorological data as inputs. Quantitative interpretation via SHAP (Shapley Additive exPlanations) analysis reveals nonlinear relationships and the distinct importance of transport-source features. A key finding is that BBE originating from northern Laos and Thailand contributes 21.23% to the predicted BBA concentrations over the SCS. Furthermore, there is a clear nonlinear positive correlation between the regional BBE rates and downwind BBA concentration, except for transport from Cambodia. Our results pinpoint that the impact of regional transport is paramount, governed by a combination of source emission intensity, transport duration, and trajectory pathway. This study establishes a satellite-driven methodology for attributing aerosol sources and clarifies the dominant controls on BBA concentration variability in a major maritime receptor region.

This study proposes a numerical model to simulate the thermal characteristics of stratospheric airships (SA). This model introduces a simple yet accurate method for the selection of the optical bands for target detection based on multispectral features. The key to this model is the selection of the spectral bands in which a significant radiation difference exists between the target and its surrounding background. Using the signal-to-noise ratio (SNR), detection probability, and false alarm rate as model evaluation metrics, and based on the analysis of the radiation differences between the target and atmospheric background, it is found that there are two narrow bands possible for SA detection: one with a central wavelength of 8.83 µm and a bandwidth of 0.35 µm and the other with a central wavelength of 11.51 µm and a bandwidth of 0.34 µm. Further numerical simulations and theoretical analyses under varying space environments demonstrate that considerable detection potential can be achieved; thus, the presented optical detection model is useful for the night detection of SA targets.