BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.earthmonitor.org//VMEVPF
BEGIN:VTIMEZONE
TZID:Europe/Amsterdam
BEGIN:STANDARD
DTSTART:20001029T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T020000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-global-workshop-2026-VMEVPF@pretalx.earthmonitor.org
DTSTART;TZID=Europe/Amsterdam:20261007T175500
DTEND;TZID=Europe/Amsterdam:20261007T180000
DESCRIPTION:Forest structural diversity - the spatial heterogeneity of cano
 py architecture across vertical and horizontal dimensions - is a fundament
 al component of ecosystem functioning. Yet its continuous global mapping r
 emains constrained by the sparse orbital sampling of spaceborne LiDAR miss
 ions such as the Global Ecosystem Dynamics Investigation (GEDI). \nHere\, 
 we integrated globally distributed GEDI-derived structural diversity metri
 cs  with dense-vector representations from a geospatial vision foundation 
 model pretrained on multi-source satellite imagery. Specifically\, we used
  the Google Satellite Embedding dataset\, derived from the AlphaEarth Foun
 dations model\, which provides globally consistent 64-dimensional embeddin
 gs at 10 m resolution from multi-source satellite imagery. Our analysis sp
 ans GEDI's full tropical-to-temperate sampling domain (52°N–52°S)\, en
 compassing 14 major biomes from temperate conifer to tropical moist broadl
 eaf forests.\nRandom forest regression models were fitted within a spatial
 ly blocked cross-validation framework stratified by biogeographic region. 
 Cross-validated R² was consistently high across structural diversity dime
 nsions\, with low inter-fold variance indicating robust transferability ac
 ross held-out biogeographic regions. Predicted structural diversity reveal
 ed strong but metric-dependent spatial gradients\, reflecting the distinct
  axes of canopy architecture — from height and complexity to vertical pr
 ofile shape - captured across the global sampling domain.\nOur results dem
 onstrate that geospatial foundation-model embeddings capture information a
 cross both vertical and horizontal dimensions of forest canopy architectur
 e\, thus providing a scalable pathway for wall-to-wall inference of forest
  structural diversity from existing spaceborne observations
DTSTAMP:20260624T081745Z
LOCATION:Aula Magna
SUMMARY:Foundation-model embeddings predict global variation in forest stru
 ctural diversity - Marco Girardello
URL:https://pretalx.earthmonitor.org/global-workshop-2026/talk/VMEVPF/
END:VEVENT
END:VCALENDAR