Near-real-time mapping for causal agents of forest disturbances in China using harmonized Landsat and Sentinel-2 dataset

2026-05-06

PRE

Ronghua Liao

Chengcheng Guo

Lingkun Chen

Yulin Jiang

Yuchen Tao

Jiani Liao

Rui Lu

Huaguo Huang

Zhou Shi

Su Ye *

DOI：10.1016/j.rse.2026.115461

Original

Original request for help

Save

Abstract

En 中文

• Developed a near-real-time framework for mapping forest disturbance agents in China. • Proposed a transferring-guided strategy to produce local training data efficiently. • Constructed stage-based random forest models using HLS spectral trajectory features. • Achieved first-alert and level-off lags of 11.6 and 15.5 days with 77% and 84% accuracy, respectively. • Demonstrated reduced false alarms compared to DIST-ALERT with longer first-alert lag.

Keywords:

forest disturbance

near-real-time mapping

Landsat and Sentinel-2

random forest model

training data transfer

Near-real-time mapping for causal agents of forest disturbances in China using harmonized Landsat and Sentinel-2 dataset

Abstract

Journal

Organization

Cited Papers

Citing Papers