Enhancing Forest‐Steppe Ecotone Mapping Accuracy through Synthetic ApertureRadar‐Optical Remote Sensing Data Fusion and Object-based Analysis

Abstract

In ecologically vulnerable regions with intricate land use dynamics, such as ecotones, frequent and intense land use transitions unfold. Therefore, the precise and timely mapping of land use becomes imperative. With that goal, by using principal component analysis, we integrated Sentinel-1 and Sentinel-2 data, using an object-oriented methodology to craft a 10-meter-resolution land use map for the forest‐grassland ecological zone of the Greater Khingan Mountains spanning the years 2019 to 2021. Our research reveals a substantial enhancement in classification accuracy achieved through the integration of synthetic aperture radar‐optical remote sensing data. Notably, our products outperformed other land use/land cover data sets, excelling particularly in delineating intricate riverine wetlands. The 10-meter land use product stands as a pivotal guide, offering indispensable support for sustainable development, ecological assessment, and conservation endeavors in the Greater Khingan Mountains region.