Evaluation of ecosystem service capacity using the integrated ecosystem services index at optimal scale in Central Yunnan, China

Understanding and quantifying the dynamic features of local ecosystem services (ESs) and integrating various ecosystem assessment results are crucial foundations for regional ES management. However, existing methods for integrating and objectively evaluating multiple ESs remain limited. Therefore, this research evaluates four key services based on the InVEST and RUSLE models in the Central Yunnan Province (CYP) —during 2000 to 2020: water yield (WY), carbon storage (CS), habitat quality (HQ), and soil conservation (SC). It then constructs an Integrated Ecosystem Service Index (IESI) using Principal Component Analysis (PCA). Additionally, this study explores the factors driving the spatial divergence of ESs by employing the optimal parameters-based geographical detector model (OPGD) at the optimal spatial scale. This study offers a more scientific and effective approach to evaluating regional integrated ecosystem service capacity. It provides a comprehensive analysis tool for weighing la..., , , # Evaluation of ecosystem service capacity using the integrated ecosystem services index at optimal scale in Central Yunnan, China

[https://doi.org/10.5061/dryad.6t1g1jx93](https://doi.org/10.5061/dryad.6t1g1jx93)

## Description of the data and file structure

This study focuses on three main aspects. First, we selected WY, SC, CS, and HQ as key indicators related to human well-being to assess ESs using the InVEST and RUSLE models from 2000 to 2020 under the contexts of water scarcity, severe soil erosion, and habitat degradation in CYP. Second, we proposed an integrated method based on principal component analysis (PCA) to construct IESI. Finally, we applied the optimal parameters-based geographical detector (OPGD) model to identify the main driving factors in CYP.

### Files and variables

This study obtained land use data, remote sensing data, meteorological data, and other relevant datasets for the years 2000, 2005, 2010, 2015, and 2020. The data were clipped using the boundary of...,