Fig. 3. Spatio-temporal classification and attribution of subsidence hot spots across Java Island. (A) Spatial distribution of subsidence hotspot derived from k-means clustering of InSAR-derived VLM time series and spatial coherence refinement using DBSCAN. Each color represents a unique cluster (cluster ID). Detailed figure of each cluster is shown in fig. S9. (B) Conceptual model showing the dominant land use or geologic mechanism associated with each group of clusters, determined using multivariate DTW of their temporal signatures. Clusters are grouped into six classes: coastal alluvial compaction (clusters 1 to 4, 17, and 21), deltas and river channels (clusters 5, 11, and 23), industrial areas (clusters 7, 10, 12 to 14, and 24), residential areas (clusters 6, 8, 9, 15, 16, and 25 to 27), agricultural areas (clusters 18 to 20, 22, and 28), and the Lusi mud volcano (cluster 29). The underlying basemap in (A) is provided by Esri, TomTom, Garmin, FAO, NOAA, and USGS.
[8/8] Ainsi la vulnérabilité côtière dépend désormais autant des pratiques locales que du changement climatique global. Adapter ces territoires suppose donc de réguler les usages de l’eau et de repenser l’aménagement littoral.
#geography #subsidence #risk #sealevelrise #Indonesia
