BEGIN:VCALENDAR VERSION:2.0 PRODID:-//pretalx//pretalx.earthmonitor.org//global-workshop-2026//MXM8GC 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-2026-7DU3RA@pretalx.earthmonitor.org DTSTART;TZID=Europe/Amsterdam:20261007T151500 DTEND;TZID=Europe/Amsterdam:20261007T153000 DESCRIPTION:Semi-natural grasslands are critical ecosystems that provide a range of essential services\, with their role as habitats for diverse spec ies being among the most significant. However\, over the past century\, se mi-natural grasslands that once covered vast areas across Europe have larg ely been transformed into intensively managed agricultural lands\, abandon ed\, or converted into forests. These large-scale land-use changes have le d to considerable biodiversity loss\, making the conservation and restorat ion of semi-natural grasslands an important component of sustainable lands cape management. \n\nWe utilized 97 in-situ herbaceous biomass samples col lected during the summer of 2019 from alvar grasslands in Western Estonia\ , all restored between 2015 and 2019. Samples were collected from 20 × 20 cm plots nested within 2 × 2 m botanical plots. Sentinel-1 and Sentinel- 2 imagery from the same period was used\, with median band values and deri ved indices (e.g.\, NDVI\, BSI\, SAVI\, VH/VV) included as predictors. \n\ nRandom Forest models were developed using Sentinel-1 and Sentinel-2 spect ral bands and derived indices as predictors. Model robustness was evaluate d using 5-fold cross-validation. Two approaches for linking field and sate llite data were tested: point sampling and a 3 × 3 kernel mean\, with poi nt sampling performing slightly better. \n\nThe model achieved an RMSE of 98 ± 54 g/m²\, an MAE of 71 ± 30 g/m²\, and an R² of 0.32 ± 0.08\, r eflecting the high spatial variability of semi-natural grasslands. SHAP an alysis identified SAVI\, NDVI\, and the vegetation red edge band B8A as th e most important predictors\, while Sentinel-1 variables contributed less to model performance. \n\nThese results highlight the dominant role of opt ical data in herbaceous biomass estimation and demonstrate that simple poi nt-based sampling can outperform spatial averaging approaches. The propose d methodology provides a practical and scalable solution for monitoring gr assland restoration. DTSTAMP:20260624T065156Z LOCATION:Aula Magna SUMMARY:Monitoring herbaceous biomass and restoration of semi-natural grass lands using machine learning on Sentinel-1 and Sentinel-2 imagery - Iris L uik URL:https://pretalx.earthmonitor.org/global-workshop-2026/talk/7DU3RA/ END:VEVENT END:VCALENDAR