Corrigendum to “Substantial increases in burned area in circumboreal forests from 1983 to 2020 captured by the AVHRR record and a new autoregressive burned area detection algorithm” [Remote Sensing of Environment 325(2025) 114789]

Abstract

The authors regret that in our original paper, a data summary error resulted in a slight overestimate of our burned area estimates. This error occurred during our conversion of the 0.05-degree pixel burn fraction maps to total annual burned area. During this process, we multiplied subpixel burn fraction (0-1) by pixel size (hectares) to estimate subpixel burned area (hectares per pixel). However, we specified a constant pixel size corresponding to the area of a 0.05-degree cell at the equator and did not adjust to the latitude of the 0.05-degree cell. This resulted in the overestimation of pixel-level burned area and a slight overestimate of the percentage of circumboreal forests experiencing increasing trends in burned area. The error does not alter the overall conclusions of our article that decadal average annual burned area has increased substantially (prior estimate: 53 %, revised estimate: 45 %) and that annual burned area has increased in about a fifth of the area circumboreal forests (prior estimate: 19.6 %, revised estimate: 17.5 %). However, the error affected our results based on direct estimates of burned area. Tables and figures created using these direct estimates of burned area are revised to reflect revised burned area totals as described below.