BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.earthmonitor.org//KZU8DF
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-KZU8DF@pretalx.earthmonitor.org
DTSTART;TZID=Europe/Amsterdam:20261007T172000
DTEND;TZID=Europe/Amsterdam:20261007T172500
DESCRIPTION:Savanna ecosystems cover approximately one fifth of Earth's lan
 d surface and provide critical ecosystem services to over one billion peop
 le\, yet their dynamic vegetation layer remains difficult to monitor consi
 stently at scale. Spaceborne lidar from the Global Ecosystem Dynamics Inve
 stigation (GEDI) mission provides vegetation structure measurements\, such
  as canopy height and cover\, but its spatially sparse sampling necessitat
 es extrapolation using satellite remote sensing data. Temporal consistency
  of these wall-to-wall mapping products remains a key challenge\, particul
 arly in heterogeneous savanna systems characterized by pronounced seasonal
 ity and complex disturbance dynamics.\nThis study compares two approaches 
 for mapping GEDI-derived canopy height and cover across Kruger National Pa
 rk\, South Africa. The first uses hand-crafted Sentinel-1/2 features deriv
 ed from phenology-informed time series analysis. The second uses TESSERA f
 oundation model embeddings (pixel-wise representations of annual Sentinel-
 1/2 time series) as open\, analysis-ready features with lightweight regres
 sion heads. Both approaches use phenology-aligned GEDI samples anchored to
  leaf-on conditions as training data\, and are evaluated using temporal cr
 oss-validation and independent airborne lidar data acquired across multipl
 e sites in the study area\, with particular focus on temporal transferabil
 ity and label efficiency.\nThe comparison addresses a question of growing 
 practical relevance: does explicit phenological knowledge embedded in task
 -specific feature engineering outperform the implicit temporal representat
 ions learned by large-scale foundation models\, and under what conditions?
  The results will inform scalable\, open\, and reproducible approaches to 
 vegetation structure monitoring in African savannas\, with direct relevanc
 e for biodiversity conservation and carbon stock assessment.
DTSTAMP:20260624T082141Z
LOCATION:Aula Magna
SUMMARY:Comparing Foundation Model Embeddings and Phenology-Informed Featur
 e Engineering for Temporally Consistent Mapping of Savanna Vegetation Stru
 cture - Marco Wolsza
URL:https://pretalx.earthmonitor.org/global-workshop-2026/talk/KZU8DF/
END:VEVENT
END:VCALENDAR