Patterns and controls of ecosystem service values under different land-use change scenarios in a mining-dominated basin of northern China

Exploring future land use/cover change (LULC) and ecosystem service values (ESV) dynamics is crucial for sustainable development from regional to global scales. However, ESV dynamics is significantly affected by economic development and policy changes, and future projections of ESV are generally characterized by large uncertainties due to complexity in LULC dynamics. Mining activities are supposed to exert significant effects on ESV dynamics, while the patterns and controls of ESV in mining-dominated basins are still less understood, especially under different land-use scenarios. Here, based on multi-source remote sensing data and land-use data in the Fenhe River Basin, we integrated the SD model, Markov model and PLUS model to simulate LULC in the Fenhe River Basin under four scenarios, i.e., the economic development priority (EDP), natural development priority (ND), cultivated land protection priority (CLP), and ecological protection priority (EPP). Based on ESV equivalent coefficient method and geographical detector, we further determined the patterns and controls of ESV. Results showed that construction land and cultivated land were mainly concentrated in the central parts of the basin, and grassland and woodland were concentrated in the edge area of the basin. Over the past two decades, construction land and woodland have experienced significant expansion, while the areas of grasslands and cultivated lands have declined due to mining activities. The high-value and sub-high-value areas of ESV were concentrated in woodland, grassland, reservoir and concentrated water area of mainstream and tributary. The low-value area was mainly located in the gathering area of construction land. The total ESV showed a downward trend with time, and the highest decline in the total ESV under the four land use scenarios was the EDP, while the lowest drop was the EPP. The highest contribution rate of ESV was woodland, which was equivalent to the contribution rate of cultivated land, grassland and water area, totally accounting for 58.54% of the whole ESV. The geographical detector method indicated that distance to the residential area, soil erodibility, distance to the river, soil type and annual average temperature were the main driving factors affecting the spatial patterns of ESV. Our results highlighted the fundamental role of mining activity in affecting LULC characterized by grassland and cultivated land shrinkage, and suggested that mining areas rehabilitation should be given priority in terms of further improving ESV in mining-dominated basins.