Search

Disturbances, such as wildfire, insect outbreak, wind throw and building construction, can significantly alter ecological processes and impact the supply of mountain ecosystem services (ESs). The comprehensive impacts of vegetation-related disturbances on ESs are poorly known, as most of existing studies have been focused on the impacts of single disturbance. Using the Eastern Helan Mountain as a case study, the vegetation-related disturbances during 1989–2020 were detected and mapped through the BEAST algorithm. Then two critical regulating ESs of carbon sequestration and sand fixation were quantified using CASA model and RWEQ model, respectively. Bayesian Network (BN) was further applied to explore the impacts of disturbances on ESs and identify the "win-win" solutions of ESs in terms of disturbance. The results suggested that BEAST could effectively identify vegetation-related disturbances at various spatio-temporal scales, which was difficult to be captured by normal land use/cover products. The established BN model could simulate the spatio-temporal pattern of sand fixation and carbon sequestration with an average classification error of 13.36% and 9.83%, respectively. Forest conservation and return pasture to forest policy can reduce disturbance frequency and intensity, and improve these two ESs. Both ES generally decreased with disturbance intensity, but firstly decreased and then increased with disturbance frequency. Thus, slight disturbance is beneficial for these ESs, especially in grassland at medium elevation. The "win-win" solutions of these two ESs can be achieved under frequent disturbances with the lowest intensity. Despite the low impacts of disturbances on ESs than other nodes in the BN, their impacts should also be taken into account into ecosystem management due to their inevitability. This BN-based approach turns out to be a promising decision support tool for mountain ecosystem managers to integrate small-scale disturbances into management and provide key information for post-disturbance recovery. • A Bayesian-based approach was designed for adaptive mountain management. • BEAST algorithm could effectively detect multiple types of vegetation related disturbances at high spatial resolution. • Thousands of scenarios were generated to explore the "what-if" boundary of ESs. • Forest conservation policy could reduce the frequency and intensity of disturbances. • Disturbance at high frequency and low intensity is beneficial for mountain ecosystem services. [ABSTRACT FROM AUTHOR]

Understanding the impacts of 2D/3D urban structure on PM 2.5 is critical for protecting resident health and sustainable development. However, accurate understanding is limited by the lack of high-resolution PM 2.5 simulation in cities with sparse monitoring stations and the simultaneously consideration of different landscape types both in 2D and 3D directions. Using Yinchuan City as a case study, an integrated simulation model combining satellite AOD, NDVI and meteorological data was developed using random forest (RF) algorithm to simulate city-scale PM 2.5 variation at 30 m resolution. Then, the most critical 2D/3D urban metrics in altering PM 2.5 and their impacts were explored through RF analysis. Finally, optimal urban structure was identified using Bayesian Network and multi-scenario analysis. The results indicated that 1) the established RF model could effectively simulate within-city variation of PM 2.5 at 30 m with an R2 of 0.75. Thus, it is feasible to use the abundant temporal information and high-resolution AOD to map city-scale PM 2.5 and overcome the limitation of spare monitoring stations. 2) The emission-related metric of distance to polluting enterprise and city center had the largest impacts on PM 2.5 , followed by NDVI, building height, impervious surface area and shape, and UGS green volume, area, shape and connectivity. Optimized 3D urban structure for reducing PM 2.5 was proposed from the three aspects of emission source location, UGS pattern and impervious surface. This study deepens the understanding of how 2D/3D urban structure impact PM 2.5 and provides scientific references for optimization possibility of urban air quality in both horizontal and vertical directions. [ABSTRACT FROM AUTHOR]