Spatio-temporal urban ecosystem service analysis with Sentinel-2A MSI data
(English)Manuscript (preprint) (Other academic)
Continuous urbanization changes the surface of our globe raising questions of sustainability, ecological functionality and living quality in metropolitan regions. Remote sensing enables us to obtain timely and reliable information on the state of urban areas and their changing patterns. The objectives of this study are to evaluate the contribution of Sentinal-2A data for urban ecosystem service mapping and to evaluate spatio-temporal characteristics of ecosystem service provisional patches through landscape metrics as an extension of the ecosystem service concept. Changes in service patterns over a 10-year time frame are mapped in the megacity of Beijing, China using Landsat TM data from 2005 and Sentinel-2A data from 2015. Landscape metrics are generated based on the classification results to evaluate the changes of urban ecosystem service provision bundles. The images are segmented using KTH-SEG, an edge-aware region growing and merging algorithm. The segments are then classified using a SVM classifier according to a classification strategy that is designed to distinguish between four natural and managed urban classes based on underlying ecosystem function and three artificial urban structures, i.e. buildings and roads that negatively influence ecosystem service provision to varying degrees and in different ways. These negative impacts are quantified through seven spatial attributes of green and blue patches and their configuration, namely area (CA), connectivity (COHESION), core area (TCA), diversity (SHDI), edge effects (CWED), percentage of land cover (PLAND) and a proximity measure. The 2015 classification accuracy of 90.2% was higher than the 2005 classification accuracy with 84.7%. Beijing’s urban development is characterized by a decrease in agricultural areas in the urban fringe in favour of new high and low density built-up areas, urban green space and golf courses. In total, high density and low-density urban areas have increased ca. 21%. Furthermore, the deconstruction of former high density low-rise suburban agglomerations into urban green space can be observed. The planar increase in urban areas is partly counteracted by the creation of managed urban green spaces. Ecosystem service bundles based on underlying land cover classes and similar spatial factors that influence service quality were derived for 2005 and 2015. Changes in landscape composition and configuration resulted in decreases of more than 30% in the bundles that represent food supply, noise reduction, waste treatment, global climate regulation. Temperature regulation/moderation of climate extremes, recreation/place values and social cohesion, aesthetic benefits/cognitive development and least affected by the observed land cover changes. The extension of the ecosystem service concept through spatio-temporal characteristics of ecosystem service provisional patches by landscape metrics is believed to give a more realistic appraisal of ecosystem services in urban areas.
Ecosystem Services, KTH-SEG, Landscape Metrics, MSI, Sentinel-2, Urban land cover
Research subject Geodesy and Geoinformatics
IdentifiersURN: urn:nbn:se:kth:diva-181866OAI: oai:DiVA.org:kth-181866DiVA: diva2:900894
QS 20162016-02-052016-02-052016-02-09Bibliographically approved