Supplementary material to "Local spatial structure of forest biomass and its consequences for remote sensing of carbon stocks"

Résumé fourni par la source

Appendix S2: A second reliability study approachIn this reliability study, we assumed that additional subplots (i.e.replicates) of 0.04 ha (n=19), 0.1 ha (n=9), 0.25 ha (n=4), 0.5 ha (n=2), 1 ha (n=1) and 2 ha (n=1) were established randomly inside the 4-ha pixel.The numbers of additional subplots were chosen to be realistic and increasing these numbers did not affect the results (not shown).Two cases were simulated: additional subplots did not overlap each other or with the original subplot ("Replicate Rr") or additional subplots were allowed to overlap each other or the original subplot ("Replicate Rr Overlap").The original subplot and the replicates were then used to calculate the ICC considering the 4-ha pixel as factor in the one-way analysis of variance (see methods).Results showed that "Replicate Rr" tended to overestimate the correction needed for all subplot areas and "Replicate Rr Overlap" overestimated the correction needed for small subplot but underestimate the correction for larger subplot area (Fig. S3).In all cases, these corrections did not outperform the "within subplot Rr" bootstrap approach for a given surveyed area.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.