The Role of Soil in Forest Drought Response: Remote Sensing-Based Monitoring of Disturbance Hotspots in Central Europe

Abstract

Prolonged drought and increased susceptibility to biotic stressors have led to a far-reaching calamity in forests dominated by Norway spruce (Picea abies (L.) Karst.) across Central Europe. European beech (Fagus sylvatica L.) has suffered from crown defoliation and increased mortality. The drastic consequences for forestry and ecosystems urge for comprehensive insights to guide future forest management. The recent drought represents an experimental setting for applying remote sensing-based anomaly detection to understand the role of site conditions for drought response. As quantitative information on soils is scarce and usually available at coarse spatial resolution, knowledge on the role of soil properties is limited. To close this gap, our study pioneers a fine-scale assessment on the role of soil properties based on satellite remote sensing-derived forest disturbance. We applied an existing forest disturbance modeling framework, based on Sentinel-2 time series data on 340 km² in Central Germany, representing hotspots of forest disturbance. Our approach allowed for a reconstruction of spatio-temporal-dynamics of forest disturbance at 10 m spatial resolution over the initial period (2019 to 2021) of the recent drought. Forest disturbance information was intersected with fine-scale soil information (1:10,000) based on roughly 2,870 soil profiles in three study areas. We investigated on how disturbed area varied among sites with different soil type, texture, stoniness, effective rooting depth and available water capacity (AWC). Our approach enabled to retrace where initial disturbance took place and how disturbance developed over time. For Norway spruce, we found that stands were most affected on deep Cambisols with medium to high AWC (90 to 160 mm) and rather low stone content, i.e., on soils usually considered as suitable in silviculture. However, these stands seemed to be more prone to suffer from unexceptional water scarcity. In contrast, we could not find evidence for pronounced disturbance on shallow soils or soils with high stone content, even though stands on soils with high AWC (> 160 mm) were least affected. Compared to spruce, the drought response of beech seemed less clearly directed to soil properties, but based on our results, we support the general concern on drought vulnerability of this species. In this regard, only stands on Luvisols with very high AWC (> 180 mm) remained undisturbed. Although long-term post-drought effects are unknown, stands initially affected did not necessarily develop the highest proportions of disturbed area, indicating recovery or adaptive mechanisms. We conclude that the integration of remote sensing-based forest disturbance monitoring with fine-scale soil information allowed us insights into soil-related drought risks. In view of the currently still high level of spruce die-back due to bark beetle infestation, disturbance in the hotspot regions we investigated will hardly to be stopped. Nevertheless, other areas with a high proportion of spruce could benefit from our findings, by identifying vulnerable stands and target in situ monitoring at an early stage of drought. For long-term strategies related to forest conversion and recovery, we consider fine-scale and quantitative information on soils crucial for implementing precision forestry.