Assessing the impact of tree felling on the loss of urban greenness using remote sensing data

Kurzfassung

Urban trees are an important component of urban landscapes, but their removal directly reduces vegetation cover and urban greenness. This study combined a Before–After–Control–Impact (BACI) framework with remote sensing data to quantify the relative NDVI loss caused by 31 documented tree felling events in the city of Ingolstadt, Germany. Each impact site was matched with three control sites. Relative NDVI loss was calculated at the impact site and within seven cumulative circular buffer zones with radii ranging from 30 to 90 m. The calculations used NDVI derived from Sentinel-2, Landsat 8/9 and color-infrared imagery (CIR). For Sentinel-2 and Landsat 8/9 imagery, two date selection approaches were applied. After the result sets produced by the different datasets and approaches were compared in terms of their internal robustness, one result set was selected for the subsequent analysis. The CIR imagery produced substantially less stable results than the satellite datasets. Among the five result sets, the results from Sentinel-2 imagery with average method showed the highest overall robustness in the internal comparison and was therefore selected for the subsequent analyses. Overall, relative NDVI loss generally decreased as the analyzed spatial extent increased, although its magnitude and spatial pattern varied considerably among the 31 tree felling cases. Post-felling underlying surface type and felling area showed comparatively clear associations with relative NDVI loss. Impervious surfaces were associated with higher relative NDVI loss at the impact sites, while larger felling areas were associated with higher predicted loss across the buffer zones. No stable independent relationship was identified between greenness proportion and relative NDVI loss. Overall, the study demonstrates the potential of combining archived remote sensing imagery with a BACI framework to retrospectively quantify urban greenness loss caused by tree felling events. It also provides initial empirical evidence for detecting and quantifying urban greenness loss associated with individual felling events using archived remote sensing data. However, the findings should be interpreted as exploratory because of the limited number of cases, incomplete information on some felling events and the absence of an independent reference dataset.