Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Master → Etats Unis → 2019 → Multi-temporal Remote Sensing of Vegetation Regrowth After a Large Wildfire

University of California Santa Barbara (2019)

Multi-temporal Remote Sensing of Vegetation Regrowth After a Large Wildfire

Kibler, Christopher Linscott

Titre : Multi-temporal Remote Sensing of Vegetation Regrowth After a Large Wildfire

Auteur : Kibler, Christopher Linscott

Université de soutenance : University of California Santa Barbara

Grade : Master of Arts (MA) in Geography 2019

Résumé partiel
Large wildfires occur regularly in southern California and disturb a variety of vegetation communities that exist in different environmental conditions across the landscape. Different plant species have different functional adaptations to fire that affect their recovery. Chaparral shrublands are well adapted to high intensity crown fires and will recover rapidly in the years following a fire. Other vegetation types, such as coniferous forests, are less adapted to severe fires and will take longer to regenerate in areas of canopy mortality. As large wildfires become more frequent across the western United States, it is important to monitor how these fires affect different vegetation communities. Changes in the species composition of individual patches affect carbon and nutrient cycling, local hydrology, and other aspects of ecosystem function. In some cases, patches of vegetation may not return to their pre-fire conditions.

This study examines how patches of chaparral shrubland and mixed conifer forest recovered from the 2007 Zaca wildfire in Santa Barbara County, California. It combines multi-temporal remote sensing imagery and field survey data to identify characteristic recovery signals for different vegetation types. Landsat imagery from 2000-2018 was used to compute the relative differenced normalized burn ratio, green vegetation fractions, and shade fractions for the entire burn scar. Land cover data for the chaparral shrublands were collected from 82 field survey transects in the summer of 2018. Land cover data for the mixed conifer forests were created by manually classifying high resolution aerial imagery. The resulting remote sensing trajectories were used to compare recovery behavior across different vegetation types. The conifer land cover data was also used to develop a statistical model that identified the environmental predictors of conifer mortality during the fire. Finally, I quantified the fractional cover of standing dead wood in Quercus chrysolepis crowns to determine if standing dead wood affects remote sensing estimates of green vegetation regrowth.

Présentation et version intégrale

Page publiée le 31 décembre 2022