EcoHealth
Application of remote sensing in ecosystem health assessment in times of global change
PIs: Dr. André Große-Stoltenberg, Prof. Till Kleinebecker
Researcher: Mojdeh Safaei
Duration: 2020-2024
|
Publications:
|
Abstract
Climate change, Land Use/Land Cover Changes (LULCCs), and biological invasions are transforming ecosystems globally, posing significant challenges to human well-being. Understanding and monitoring terrestrial ecosystem health—a multifaceted concept reflecting an ecosystem’s structure, function, resilience, and recovery capacity—is essential for sustainable development. Combining ground-based methods with advanced remote sensing technologies makes assessing and monitoring ecosystem health over extensive spatial scales possible, providing critical insights for Sustainable Development Goals (SDGs). The study pursued three key objectives: (1) comparing ground-based and remote sensing methods for ecosystem health assessment, (2) employing the Dynamic Habitat Index (DHI) to monitor ecosystem health dynamics over time, and (3) analyzing the sensitivity of DHIs to environmental changes across diverse LULC types.
This study showed the potential of DHIs—derived from Normalized Difference Vegetation Index (NDVI) data—to evaluate the health of coniferous forests under extreme drought conditions. We also evaluated the sensitivity of the DHI to changing environmental conditions across various Land Use/Land Cover (LULC) types. The analysis highlighted the effectiveness of DHIs in capturing the impacts of drought on Central European coniferous forest ecosystems. DHIs successfully distinguished between damaged and nondamaged forest areas, showing promise as an early warning system for ecosystem degradation and functional changes. Integrating DHIs with meteorological and ancillary geodata enhanced their interpretive power, highlighting the dynamic interplay of pedo-climatic factors in shaping ecosystem health.
The findings illustrate the strengths and limitations of different approaches, emphasizing the importance of indicator selection related to regional contexts, historical background, and environmental conditions. The integrated methodologies developed in this research offer valuable tools for land managers and decision-makers, contributing to sustainable land use strategies and advancing SDG indicators related to land degradation.
Future research could focus on identifying ecological thresholds, tipping points, and spatial early warning signals of ecosystem transitions to alternative stable states. Advanced methods, such as deep learning and composite remote sensing indices (e.g., the DHI), can be computed at high spatial resolutions to provide continuous metrics for assessing forest damage. For example, these methods could be used to map decreasing annual productivity or detect stagnation at low levels following disturbances. Incorporating structural data and canopy height information from LiDAR, supported by automated field measurements for calibration, would be a crucial next step in enhancing the identification of tipping points and regime shifts to other types of (forest) ecosystems.
