Forest Change Detection and Monitoring
Using Passive and Active Remote Sensing Data

Background

Forest ecosystems under pressure

The RS4FOR project investigated forest cover dynamics in the Polish Carpathians using a multi-source Earth Observation framework spanning over four decades. By combining dense Landsat time series (1978–2020) with Sentinel-2 multispectral imagery, Sentinel-1 SAR backscatter, and airborne LiDAR point clouds, the project addressed fundamental challenges in discriminating gradual canopy degradation from abrupt stand-replacing disturbances. Machine learning methods were applied to map forest stand composition at fine spatial resolution, whilst LiDAR-derived structural metrics enabled characterisation of old-growth attributes that are otherwise inaccessible through optical remote sensing alone.

Objectives

Research objectives

• 01 Forest cover dynamics — detection and classification of conversions and modifications across a 40-year Landsat archive
• 02 Species composition mapping — machine learning classification of forest stand types using Sentinel-2 time series and environmental predictors
• 03 Forest health assessment — early detection of decline and disturbance through multi-source optical and radar data
• 04 Old-growth characterisation — structural mapping of old-growth stands using LiDAR metrics for biodiversity and conservation assessments

Publications

Selected outputs