BEGIN:VCALENDAR VERSION:2.0 PRODID:-//pretalx//talks.osgeo.org//foss4g-2026//talk//XW3XTM BEGIN:VTIMEZONE TZID:JST BEGIN:STANDARD DTSTART:20000101T000000 RRULE:FREQ=YEARLY;BYMONTH=1 TZNAME:JST TZOFFSETFROM:+0900 TZOFFSETTO:+0900 END:STANDARD END:VTIMEZONE BEGIN:VEVENT UID:pretalx-foss4g-2026-XW3XTM@talks.osgeo.org DTSTART;TZID=JST:20260903T130000 DTEND;TZID=JST:20260903T133000 DESCRIPTION:Introduction   \nRemote sensing-based spectral diversity ha s emerged as a scalable proxy for field-measured biodiversity [1-4]\, gr ounded in the Spectral Variation Hypothesis\, which posits that greater sp ectral heterogeneity in remotely sensed imagery reflects greater ecologica l diversity on the ground [5]. While studies have examined methodological choices known to influence spectral diversity mapping performance [6\,7]\, the effect of input feature type\, specifically the choice between spectr al bands\, Principal Component Analysis (PCA) transformed spectral bands\, and spectral indices (SI)\, has not been empirically tested. The biod ivMapR package [8] implements the spectral species concept for biodivers ity mapping and acknowledges both PCA-transformed bands and SI as valid input options\, but provides no firm recommendation\, explicitly calling f or users to test and compare approaches against ground observations [9]. W ithin the FOSS4G ecosystem\, reproducible biodiversity monitoring workflow s based entirely on open data and open-source software remain under-evalua ted\, particularly with respect to methodological parameterisation cho ices. To address this gap\, we compare three spectral diversity comput ation approaches across three ecologically distinct hemi-boreal forest sit es in  Estonia and validate them against field-measured tree speci es diversity. We aim to address two research questions: (1) Does PCA-bas ed dimensionality reduction produce stronger correlations with field-measu red tree species diversity than non-PCA spectral bands or SI used  as biodivMapR inputs? (2) Are results consistent across ecologicall y distinct hemi-boreal forest sites?  \nMethods   \nWe selected  three study sites representing characteristic hemi-boreal landscape t ypes in Estonia: Otepää\, a glacial moraine upland characterised b y mixed nemoral forests dominated by spruce (n = 3\,758 plots)\; Soomaa\ , a peatland and floodplain forest complex dominated by pine\, birch\, and alder with extensive bog-forest mosaics (n = 10\,259)\; and Rapla\, a l owland managed forest mosaic representative of intensively managed hemi-bo real production forests (n = 4\,322). We filtered forest inventory valid ation plots using a morphological erosion criterion\, retaining only p lots with a minimum residual area of 900 m² after applying a 30-metre inw ard buffer\, to exclude geometrically irregular or insufficiently large pl ots where moving-window diversity estimates would be strongly influenced b y non-forest edge pixels. We pre-processed the Sentinel-2 imagery from the Copernicus Open Access Hub using open-source geospatial libraries in Pyth on. We masked forests prior to spectral diversity computation\, ensuring a nalysis was restricted to forested areas across all sites.   \nWe comp ared three approaches: (1) PCA applied to ten Sentinel-2 bands\, with the first three principal components used as input features for Shannon spectr al diversity mapping\; (2) bands without dimensionality reduction (non-PCA )\; and (3) a set of three ecologically relevant SI - Normalised Differenc e Vegetation Index (NDVI)\, Normalised Difference Water Index (NDWI)\, and Leaf Area Index proxy derived from Soil-Adjusted Vegetation Index (LAI-SA VI)\, derived from the same Sentinel-2 bands. Shannon alpha-diversity was computed using a moving window of three pixels (30 m × 30 m) across tiled processing grids\, with mosaicked outputs validated against field-mea sured tree species Shannon diversity using Spearman rank correlation. The effect of inter-quartile range (IQR) outlier filtering\, an internal b iodivMapR preprocessing step that masks pixels whose spectral values fal l outside Q1 - 4 × IQR or Q3 + 4 × IQR\, was additionally assessed by co mparing results with and without this filtering applied. The entire workfl ow was implemented using open Sentinel-2 imagery and the open-source R pac kage biodivMapR\, ensuring full methodological transparency and reproduc ibility within the FOSS4G ecosystem.  \nResults   \nPCA-based spectr al diversity consistently achieved the highest correlation with field-meas ured tree species diversity across all three sites\, outperforming both th e SI and spectral band approaches. The performance advantage of PCA was modest but consistent across sites. SI outperformed spectral bands at tw o of three sites. IQR filtering produced no consistent improvement across sites or methods. Pixel-level analysis confirmed that IQR masked fewer t han 1.4% of pixels in the pre-masked forest-only imagery\, regardles s of site or method.  \nDiscussion  \nResults indicate a moderate relationship between spectral and field-measured diversity across all meth ods and sites. While remote sensing-derived spectral diversity cannot serv e as a standalone biodiversity indicator at this spatial resolution [1\,4] \, it provides a reproducible\, scalable\, and open-source means of moni toring relative diversity patterns across landscapes.  \nThe modest but consistent advantage of PCA over spectral indices reflects its ability to condense spectral variance across correlated bands into uncorrelated compo nents\, reducing noise while retaining biodiversity-relevant signal. T his comes at the cost of additional processing overhead and user-guide d component selection\, which limits full workflow automation. SI\, by contrast\, are theoretically grounded\, computationally efficient\, and d irectly interpretable\, making them a practical alternative for large-ar ea applications where automation and transferability are priorities. Spect ral bands alone are not recommended due to band collinearity\, increased s pectral noise. The negligible effect of IQR filtering suggests that\, in p re-masked forest-only imagery\, this step may not be necessary. The ecolog ical significance of the pixels removed warrants further attention.   \nConclusions   \nWe evaluated three input feature approaches for spec tral diversity mapping in hemi-boreal forests using Sentinel-2 and biodi vMapR\, validated against field-measured tree species diversity across three ecologically distinct sites. PCA consistently outperformed spectral bands and SI\, though the advantage was modest. SI represent a practica l open-source alternative for large-area applications where computational efficiency is a priority. We identify integration of complementary env ironmental predictors is identified as a priority for future work\, as we expect it to strengthen this relationship. Because the workflow relies e xclusively on open data and tools\, it transfers directly to other forest ecosystems without licensing constraints. All scripts and parameter config urations will be made openly available to support reproducible biodiversit y monitoring.  \nReferences  \n[1] Fassnacht\, F.E. et al.\, 2022. Abo ut the link between biodiversity and spectral variation. Appl. Veg. Sci. 2 5\, e12643.\n[2] Rocchini\, D. et al.\, 2021. rasterdiv: An Information Th eory tailored R package for measuring ecosystem heterogeneity from space. Methods Ecol. Evol. 12\, 1093–1102.\n[3] Torresani\, M. et al.\, 2024. R eviewing the Spectral Variation Hypothesis: Twenty years in the tumultuous sea of biodiversity estimation by remote sensing. Ecol. Inform. 82\, 1027 02.\n[4] Wang\, R.\, Gamon\, J.A.\, 2019. Remote sensing of terrestrial pl ant biodiversity. Remote Sens. Environ. 231\, 111218.\n[5] Palmer\, M.W. e t al.\, 2002. Quantitative tools for perfecting species lists. Environmetr ics 13\, 121–137.\n[6] Robertson\, K.M. et al.\, 2023. Effects of Spatia l Resolution\, Mapping Window Size\, and Spectral Species Clustering on Re mote Sensing of Plant Beta Diversity. J. Geophys. Res. Biogeosciences 128\ , e2022JG007350. \n[7] Schmidtlein\, S.\, Fassnacht\, F.E.\, 2017. The spe ctral variability hypothesis does not hold across landscapes. Remote Sens. Environ. 192\, 114–125. \n[8] Féret\, J.\, De Boissieu\, F.\, 2020. bi odivMapR: An r package for α‐ and β‐diversity mapping using remotely sensed images. Methods Ecol. Evol. 11\, 64–70. \n[9] Féret\, J.\, De B oissieu\, F. (n.d.). biodivMapR: Produce diversity maps from optical image s. https://jbferet.github.io/biodivMapR/articles/biodivMapR_02.html DTSTAMP:20260717T225735Z LOCATION:Cosmos1 SUMMARY:Evaluating spectral diversity approaches for tree species diversity mapping in hemi-boreal forests using Sentinel-2 and biodivMapR - Arathi Biju URL:https://talks.osgeo.org/foss4g-2026/talk/XW3XTM/ END:VEVENT END:VCALENDAR