Climate, landscape, and human influences on fire in southern Patagonia: A basin-scale approach

Abstract

Human activities and how they interact with climate and landscape to influence fire regimes is complex and still remains a major research challenge in most Patagonian forests. Recent studies on the fire history in Patagonia covered large areas, making it difficult to model fire occurrence and burned patch size. The main objective of this work was to evaluate the influence of climatic variability, landscape variations and human activities on the occurrence and size of fire at the basin scale. We aimed to update previous fire history of the Río de las Vueltas – Río Túnel basin, in southern Patagonian Andes, with the specific aims to (1) reconstruct in detail the spatio-temporal patterns of fire, (2) model fire variability as a function of climate variability, landscape and human activities, and (3) determine the influence of regional climate on fire occurrence at the basin scale. We used dendrochronological techniques to update the fire history of the study basin. At the basin scale, we used (1) maximum entropy models to estimate the influence of climate, landscape, and humans on fire probability, (2) general linear models to assess how these factors modulate fire size, and (3) superposed epoch analysis to assess the influence of regional water deficit on fire occurrence. We found that fire occurrence increased sharply with the arrival of settlers in the early 20th century and decreased in the early 21st century due to the shift in recognition of the ecosystem services provided by forests. Additionally, the mean fire interval, the number of fires during drier years, and fire size, increased along the northwest-southeast gradient of decreasing precipitation. Our results support that fire regimes are largely human driven, and conditioned by climate and landscape characteristics. The analysis at the basin scale facilitated the understanding of how these factors regulate fire regimes, and consequently, our comprehension at larger spatial scales.