Chao Ma

Massive landslides over large regions can be triggered by heavy rainfalls or major seismic events. Mapping regional landslides quickly is important for disaster mitigation. In recent years, deep learning methods have been successfully applied in many fields, including landslide automatic identification. In this work, we proposed a deep learning approach, the ResU-Net, to map regional landslides automatically. This method and a baseline model (U-Net) were collectively tested in Tianshui city, Gansu province, where a heavy rainfall triggered more than 10,000 landslides in July 2013. All models were performed on a 3-band (near infrared, red, and green) GeoEye-1 image with a spatial resolution of 0.5 m. At such a fine spatial resolution, the study area is spatially heterogeneous. The tested study area is 128 km2, 80% of which was used to train models and the remaining 20% was used to validate accuracy of the models. This proposed ResU-Net achieved higher accuracy than the baseline U-Net model in this mountain region, where F1 improved by 0.09. Compared with the U-Net model, this proposed model (ResU-Net) performs better in discriminating landslides from bare floodplains along river valleys and unplanted terraces. By incorporating environmental information, this ResU-Net may also be applied to other landslide mapping, such as landslide susceptibility and hazard assessment.

As PM2.5 affect human health, it is important to target tree planting in the role of reducing air pollution concentrations. PM2.5 capture capability of greening trees is associated with leaf morphology, while quantitative research is scanty. In this paper, the PM2.5 capture capability of 25 species in Beijing and Chongqing were examined by a chamber device. Groove proportion, leaf hair, stomatal density, and stomata size were selected as indexes of leaf morphology. Results showed that groove proportion and stomata size significantly relate to PM2.5 capture quantity, while no significantly positive correlations were found for leaf hairs and stomatal density. Broadleaf species are conducive to PM2.5 capture for their rich leaf morphology and have an edge over coniferous in PM2.5 capture per leaf area. However, coniferous had a larger PM2.5 capture capability per tree due to the advantage of a large leaf area. Significant difference existed between the species in Beijing and Chongqing due to the different leaf morphology. Urban greening trees are diverse and the structures are complicated. Complex ecological environment may lead to different morphology characteristics. Climate and pollution conditions should be considered when greening.