Impacts of landscape pattern evolution on typical ecosystem services in Ganjiang River Basin, China

Abstract

Understanding the response mechanism of ecosystem services (ES) to landscape patterns is of great signi cance for regional landscape planning and sustainable development. In this study, the landscape index and InVEST model were used to quantitatively analyze the spatio-temporal evolution of landscape patterns and ES in the Ganjiang River Basin of China from 1990 to 2020. Furthermore, the bivariate Moran's I method and spatial error model (SEM) were used to test the spatial correlation between landscape index and ES. The results showed that (1) cropland decreased and construction land increased, and the overall landscape tended to be fragmented, the patch shape tended to be complicated, and landscape diversity increased from 1990 to 2020. Water conservation (WC) and soil conservation (SC) capacity increased by 10.56 mm and 16.24 t hm -2 a -1 , respectively, whereas carbon storage (CS) decreased by 1.22 t hm -2 a -1 . (2) The responses of different typical ES to landscape patterns were different in landscape index and response degree. Typical ES negatively responded to Shannon's diversity index and patch density. WC was sensitive to the Splitting Index, whereas SC and CS were more responsive to the average patch area. (3) The overall purpose of territorial spatial planning within a basin should be to reduce the fragmentation and heterogeneity of the landscape. According to four local aggregation patterns of landscape index and ES, corresponding measures can be taken according to local conditions in different regions. The results can provide a quantitative basis for landscape management and ecological construction in the Ganjiang River basin and scienti c guidance for the Yangtze River conservation strategy. HighlightsGan River Basin landscape pattern and ES were evaluated Bivariate Moran's I methods and spatial regression were used for the analysis Impact of landscape patterns on ES were shown at the grid scale Landscape pattern optimization measures were proposed for four clustering methods considerably reduced carbon reserves are mainly located in urban areas, including Nankang, Zhanggong, Jizhou, and Yuanzhou. The continuous improvement of urbanization had impacted the structure and process of ES leading to a sharp decrease in carbon reserves within the region. Responses of ES to landscape patterns Spatial correlation analysisThe correlation analysis with landscape pattern index and ES in Ganjiang River Basin from 1990 to 2020 was used by GeoDa software. It revealed a conspicuous spatial correlation in the landscape pattern index and ES using global bivariate Moran's I (Fig. 7). Moran's I values of the CS, SC, and WC were below 0 for SHDI, PD, and SPLIT and greater than 0 for AREA_MN, and all passed the signi cance test. These results indicated that these three ES showed a negative correlation with SHDI, PD, and SPLIT, and a positive correlation with AREA_MN. Therefore, more heterogeneous patches and fragmented landscapes are detrimental to the maintenance of the functions of ES. The correlation between SHAPE_MN...