EO4Bay

EO4Bay - Developing an Earth observation-based forest monitoring system for Bavaria

Funding and duration:

Bayerisches Staatsministerium für Ernährung, Landwirtschaft, Forsten und Tourismus | 2024-2027

Summary:

Approximately one-third of Bavaria’s land cover is forested and is under increasing pressure from climate-related extremes. Forest owners are facing major challenges from disturbances like bark beetle calamities in spruce and drought stress in beech. Addressing these issues needs spatially explicit, up-to-date and detailed information about the location, timing and intensity of these disturbances as well as their impact on wood supply and forest recovery processes.

Current forest monitoring methods, such as aerial photography, take several weeks or even months from data acquisition, analysis and making the data available, delaying timely decision- making. Thus, it is necessary to build an earth observation- based monitoring system capable of providing semi-automated, near real-time information on the changes in forest canopy and it’s impact on timber stocks.

EO4Bay is dedicated to developing and piloting such a semi-automated earth observation-based monitoring system specifically designed for Bavarian forests at a 10m spatial resolution.

EO4Bay aims to enhance the resilience and sustainability of Bavarian forests by providing critical data to inform management practices and policy decisions. The project is funded by the Bavarian StMELF (Bayerisches Staatsministerium für Ernährung, Landwirtschaft, Forsten und Tourismus) with the objectives:

Development of Bavarian Earth Observation Data Cube: This involves the automatic processing of all available data from Copernicus Sentinel-2 mission for Bavaria
Creation of an AI-based Approach for Vegetation Height Mapping: Deep Learning models (e.g. UNet) are trained on Sentinel 2 images with spaceborne laser scanning data from GEDI (Global Ecosystems Dynamics Investigations) system. It is validated from high- resolution airborne laser scanning data.
Transformation of Vegetation Height Maps into Timber Stock Maps: By applying allometric functions derived from the federal forest inventory and validating them against the state’s own inventory areas.
Temporal Analysis of Vegetation Height Maps: Analyzing vegetation height maps over time to detect negative changes caused by disturbance and crown thinning as well as positive changes caused by height increase and associated changes in wood stock.