Most cities in China are experiencing severe air pollution due to rapid economic development and accelerated urbanization. Long-term air pollution data with high temporal and spatial resolutions are needed to support research into physical and chemical processes that affect air quality, and the corresponding health risks. For the first time, data on PM₁₀, PM_2.5, SO₂, NO₂, O₃ and CO concentrations in 23 ambient air quality automatic monitoring stations and routine meteorological were collected between January 2014 and December 2016 to determine the spatial and temporal variation in these pollutants and influencing factors in Chengdu. The annual mean concentrations of PM_2.5 and PM₁₀ exceeded the standard of Chinese Ambient Air Quality and World Health Organization guidelines standards at all of the stations. The concentrations of PM₁₀, PM_2.5, SO₂ and CO decreased from 2014 to 2016, and the NO₂ level was stable, whereas the O₃ level increased markedly during this period. The air pollution characteristics in Chengdu showed simultaneously high PM concentrations and O₃. High PM concentrations were mainly observed in the middle region of Chengdu and may have been due to the joint effects of industrial and vehicle emissions. Ozone pollution was mainly due to vehicle emissions in the downtown area, and industry had a more important effect on O₃ in the northern area with fewer vehicles. The concentrations of PM₁₀, PM_2.5, NO₂ and CO were highest in winter and lowest in summer; the highest SO₂ concentration was also observed in winter and was lowest in autumn, whereas the O₃ concentration peaked in summer. Haze pollution can easily form under the weather conditions of static wind, low temperature and relative humidity, and high surface pressure inside Chengdu. In contrast, severe ozone pollution is often associated with high temperature.

Predicting the susceptibility of a specific part of a landslide (SSPL) involves predicting the likelihood that the part of the landslide (e.g., the entire landslide, the source area, or the scarp) will form in a given area. When predicting SSPL, the landslide samples are far less than the non-landslide samples. This class imbalance makes it difficult to predict the SSPL. This paper proposes an advanced artificial intelligence (AI) model based on the dice-cross entropy (DCE) loss function and XGBoost (XGB_DCE) or Light Gradient Boosting Machine (LGB_DCE) to ameliorate the class imbalance in the SSPL prediction. We select the earthquake-induced landslides from the 2018 Hokkaido earthquake as a case study to evaluate our proposed method. First, six different datasets with 24 landslide influencing factors and 10,422 samples of a specific part of the landslides are established using remote sensing and geographic information system technologies. Then, based on each of the six datasets, four landslide susceptibility algorithms (XGB, LGB, random-forest (RF) and linear discriminant analysis (LDA)) and four class balancing methods (non-balance (NB), equal-quantity sampling (EQS), inverse landslide-frequency weighting (ILW), and DCE loss) are applied to predict the SSPL. The results show that the non-balanced method underestimates landslide susceptibility, and the ILW or EQS methods overestimate the landslide susceptibility, while the DCE loss method produces more balanced results. The prediction performance of the XGB_DCE (average area under the receiver operating characteristic curve (0.970) surpasses that of RF (0.956), LGB (0.962), and LDA (0.921). Our proposed methods produce more unbiased and precise results than the existing models, and have a great potential to produce accurate general (e.g., predicting the entire landslide) and detailed (e.g., combining the prediction of the landslide source area with the landslide run-out modeling) landslide susceptibility assessments, which can be further applied to landslide hazard and risk assessments.