Evaluating Land Use and Ecological Patterns in Xiong’an New Area of China with Machine Learning Methodology

Abstract

The Xiong’an New Area, following the precedent of the Shenzhen Special Economic Zone and Shanghai Pudong New Area, marks a significant development. This study introduces a method to optimize the feature variable selection for Sentinel-2 images from 2016 to 2022, aiming for precise land-use classification in Xiong’an using machine learning. The classification reveals substantial growth in the infrastructure and aquatic areas in Rongcheng and Xiongxian counties, outpacing Anxin from 2016 to 2022. The Remote Sensing-Based Ecological Index (RSEI) indicates a generally stable yet improving ecological landscape, especially in denser areas like Xiongxian and Rongcheng, aligning regional development with ecological enhancement. EOF analysis shows a spatial ecological division, with positive RSEI values in the western regions and negative values in the east, along with temporal fluctuations indicating a decline in the west and an increase in the east since 2017. Additionally, the RSEI’s short-cycle fluctuations emphasize the dynamic ecological state of the area, influenced by both long-term trends and transient factors.