Fig. 1: Postfire surface radiometric warming in summer and its amplification by fire size.
From: Forest fire size amplifies postfire land surface warming
a, Annual burnt area over 2003–2016. b, Surface radiometric temperature change (ΔΤ) in summer (June–August) one year after fire. c,d, Slope βΔΤ and R2 derived by fitting a linear regression model (ΔΤ = α + βΔΤ × log10[fire size]) in the grid cells with a minimum of 10 fires. The solid dots indicate pixels with locally significant regressions (P < 0.05, the two-tailed t-test), all of which have passed a more rigorous field significance test corrected for the false-discovery rate (FDR; αFDR = 0.10; Methods). e, Spatial distribution of forest type composition. The colour opacity indicates the proportion of forest area to land area. f, Regional mean postfire summer ΔT (in violet, significantly greater than zero for all years at α = 0.05, one-tailed t-test, with negligible s.e.m.), and the domain-wide βΔT value derived by fitting a single linear regression model (ΔΤ = α + βΔΤ × log10[fire size]) across the whole study domain (in orange; P < 0.05 for all years, Student’s t-test; the shaded areas show standard errors), both for up to 14 years after fire. All the maps have a 2° spatial resolution. The light-grey background indicates northern temperate and boreal forests (40° N–70° N) with a >10% ground coverage. Figure developed using Python open-source tools.
