Modeling above-ground carbon storage: a remote sensing approach to derive individual tree species information in urban settings

Vegetation has gained importance in respective debates about climate change mitigation and adaptation in cities. Although recently developed remote sensing techniques provide necessary city-wide information, a sufficient and consistent city-wide information of relevant urban ecosystem services, such as carbon emissions offset, does not exist. This study uses city-wide, high-resolution, and remotely sensed data to derive individual tree species information and to estimate the above-ground carbon storage of urban forests in Berlin, Germany. The variance of tree biomass was estimated using allometric equations that contained different levels of detail regarding the tree species found in this study of 700 km2, which had a tree canopy of 213 km2. The average tree density was 65 trees/ha per unit of tree cover and a range from 10 to 40 trees/ha for densely urban land cover. City-wide estimates of the above-ground carbon storage ranged between 6.34 and 7.69 tC/ha per unit of land cover, depending on the level of tree species information used. Equations that did not use individually localized tree species information undervalued the total amount of urban forest carbon storage by up to 15 %. Equations using a generalized estimate of dominant tree species information provided rather precise city-wide carbon estimates. Concerning differences within a densely built area per unit of land cover approaches using individually localized tree species information prevented underestimation of mid-range carbon density areas (10–20 tC/ha), which were actually up to 8.4 % higher, and prevented overestimation of very low carbon density areas (0–5 tC/ha), which were actually up to 11.4 % lower. Park-like areas showed 10 to 30 tC/ha, whereas land cover of very high carbon density (40–80 tC/ha) mostly consisted of mixed peri-urban forest stands. Thus, this approach, which uses widely accessible and remotely sensed data, can help to improve the consistency of forest carbon estimates in cities.