// what we've learnt about Austria

The first index to fuse cadastral parcels with 1 m LiDAR terrain, 0.2 m orthophoto, Sentinel-2/1 and Hansen forest change — every parcel and object segmented, classified, queryable. Every number below is computed live against the index.
● loading…

1 · forest-loss shock [hansen / temporal]

annual loss inside processed KGs — hectares

→ recent clearfells · earthwork / salvage · regenerating · deforestation hotspots

2 · land cover [composition]

→ top KGs by tree area · most built-up · high-biodiversity

3 · parcel auto-classification [15-class]

→ disturbed · farmsteads · alpine pastures · vineyards

4 · dominant type per parcel [frav vector]

single most-covering object type per parcel — indexed for sub-50 ms cross-Austria filtering

→ tree-dominant >70% · pure grass · agroforestry · tree+grass+crop mosaic

5 · giant-tree census [discovery]

→ tallest (conf>0.9) · highest-biomass KGs

6 · buildings [development]

→ apartment blocks 5+ · most built-up parcels · densest roof cover

7 · elevation gradient [ecology]

how landscape character changes with altitude — per-KG band means

→ high-alpine · flat lowlands · steep S-facing meadows

8 · aspect & terrain ruggedness [1 m DTM]

which way parcels face, and how forest cover tracks ruggedness

→ rugged forest · roughest parcels · flat cropland · solar potential

9 · earth on the move [temporal]

net DTM change between the 2022 and 2024 LiDAR campaigns, co-registered at 1 m

→ active earthworks · erosion / subsidence · construction sites

10 · where the classifiers disagree [quality]

→ all divergence pairs · high-divergence KGs · lowest-confidence KGs