Impact of touristification and landscape pattern on habitat quality in the Longji Rice Terrace Ecosystem, southern China, based on geographically weighted regression models

Abstract

Although the roles of tourism development and landscape value in the conservation of rice terraces have become increasingly important in recent years, rice terraces still face abandonment and desertification. Habitat quality (HQ) assessments evaluate the biodiversity conservation status of rice terraces; however, the spatial and temporal evolution of rice terrace HQ and influencing factors remains poorly understood. Here, a case study was performed of the Longji Terraces, Longsheng, Guangxi, China. Land use and land cover (LULC) remote sensing data for 1985–2020 (every 5 years) were obtained, and the InVEST-HQ model was used to evaluate and map the spatial and temporal evolution patterns and characteristics of HQ. A landscape pattern index and tourism indicators were combined and used to construct geographically weighted regression (GWR) models to explore the impact of tourism and landscape pattern on HQ. Our findings indicated that: (1) forests and terraces are the dominant landscape patches in Longji because they are better protected. Forest area and HQ changes exhibited temporal consistency, meaning that as forest area increased, HQ improved. (2) Longji Terraces exhibited optimum HQ overall, with a multi-year average of 0.977 and insignificant inter-annual variations. Spatially, the geographical HQ distribution was characterised by ‘low local levels, high global level.’ However, the percentage of HQ degraded areas in the final period reached 68 %, which is significantly greater than that of the improved areas (28 %). Moreover, improved areas had more spatial overlap with administrative villages that present superior tourism development. (3) HQ was significantly negatively correlated with night-time light (NTL), showed periodic changes (first negative, then positive) with Shannon’s diversity index (SHDI), was positively correlated with the contagion index (CONTAG), and showed periodic transitions (initially positive, then negative) with the mean of patch area (AREA_MN), population (POP), and patch density (PD). (4) Over time, the study area HQ showed weaker clustering and increased dispersion. In addition, the influence of tourism and landscape patterns increased from 2000 to 2020, with R2 increasing from 0.42 to 0.95. The results establish a scientific foundation for the conservation of rice terraces and sustainable use of landscape resources.

Keywords

• Rice terrace
• Ecosystem service
• Habitat quality
• GWR
• InVEST