A new approach for modeling delayed fire‐induced tree mortality

Ecosphere

Matteo Garbarino¹  Davide Ascoli¹  Andrew Hacket‐Pain²  Janet Maringer³  Marco Conedera³ 

[1] Department of Agriculture, Forest and Food Sciences University of Turin Largo Paolo Braccini 2 Grugliasco10095Italy;Department of Geography and Planning School of Environmental Science University of Liverpool LiverpoolL69 7ZTUK;Insubric Ecosystems Swiss Federal Institute for Forest, Snow and Landscape Research WSL A Ramél 18 CadenazzoCH‐6593Switzerland;

关键词: Cox proportional hazards model;    Fagus sylvatica;    fire ecology;    fungi infestation;    Kaplan–Meier estimator;    novel disturbance;   

DOI  :  10.1002/ecs2.3458

来源: DOAJ

【 摘 要 】

Abstract Global change is expanding the ecological niche of mixed‐severity fire regimes into ecosystems that have not usually been associated with wildfires, such as temperate forests and rainforests. In contrast to stand‐replacing fires, mixed‐severity fires may result in delayed tree mortality driven by secondary factors such as post‐fire environmental conditions. Because these effects vary as a function of time post‐fire, their study using commonly applied logistic regression models is challenging. Here, we propose overcoming this challenge through the application of time‐explicit survival models such as the Kaplan‐Meier (KM‐) estimator and the Cox proportional‐hazards (PH‐) model. We use data on tree mortality after mixed‐severity fires in beech forests to (1) illustrate temporal trends in the survival probabilities and the mortality hazard of beech, (2) estimate annual survival probabilities for different burn severities, and (3) consider driving factors with possible time‐dependent effects. Based on our results, we argue that the combination of KM‐estimator and Cox‐PH models have the potential of substantially improve the analysis of delayed post‐disturbance tree mortality by answering when and why tree mortality occurs. The results provide more specific information for implementing post‐fire management measures.

【 授权许可】

Unknown