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

Accueil du site → Doctorat → États-Unis → 2019 → Uncoupling impacts of drought and short-interval fire on chaparral in southern California using time-sequential Landsat imagery

San Diego State University (2019)

Uncoupling impacts of drought and short-interval fire on chaparral in southern California using time-sequential Landsat imagery

Storey, Emanuel Arnal.

Titre : Uncoupling impacts of drought and short-interval fire on chaparral in southern California using time-sequential Landsat imagery

Auteur : Storey, Emanuel Arnal.

Université de soutenance  : San Diego State University,

Grade : Doctor of Philosophy (PhD) in Geography 2019

Résumé partiel
Terrestrial vegetation growth, climate variation, and fire activity have been strongly coupled since the advent of vascular plants approximately 420 million years before present. Rises in anthropogenic land cover transformation, invasive species propagation, carbon emission, and artificial sources of fire ignition in recent centuries have greatly affected terrestrial plant communities. The populous, semi-arid region of southern California is marked by heightened fire activity and extreme drought episodes in recent decades. Contemporary regimes (spatial-temporal patterns) of fire and drought may exceed historical ranges of variability and thus threaten native plant communities, which may have adapted to quite different disturbance regimes. Postfire recovery of southern California shrub species (chaparral) may be impeded by fires that recur at short intervals or by severe drought. Evaluating the response of chaparral to compound fire-drought effects is key in conservation planning under projected climate, land use, and fire regime scenarios. Time-sequential remote sensing based on Landsat satellite imagery is a useful way to characterize shrub cover change over regional extents and several decades, and may provide a broader perspective than much of the ecological and biogeographic literature, which is mostly based on field studies. Landscape-scale variations of shrub type, soil reflectance, and terrain may introduce spatial bias into signals of change in fractional shrub cover (FSC) based on spectral vegetation indices (SVIs), including Normalized Difference Vegetation Index (NDVI). Chapter 2 evaluates these potential error sources in the application of postfire recovery assessment based on Landsat image series. Chapter 2 is based on three chaparral landscapes having internally-varied postfire recovery trajectories, and involves a cross-stratification variable-control approach. Results show that differences in shrub type linked to biomass variations can significantly influence linear NDVI-FSC relations, and such effects are accentuated in postfire recovery metrics based on NDVI difference. Minor effects due to soil spectral reflectance differences were observed at one site, but no significant effect of terrain illumination variability was found. Insights from Chapter 2 guided the methodologies of Chapters 3 and 4, which evaluate regional-scale influences of repeated fire and of drought on chaparral recovery, respectively. A regional map of change in FSC between the periods 1984-1989 and 2014-2018 was derived by transforming Landsat NDVI trajectories into calibrated estimates of FSC within 49 manually-delineated portions of the regional chaparral community, based on detailed shrub cover maps derived from high spatial resolution aerial imagery. Postfire recovery patterns were evaluated in 246 areas which burned in the period 1985-2008 and represent varied numbers of burns (one to three) and fire return intervals (1 to 23 years). Fire-return interval was not significantly related to recovery, except at sites that had burned three times within 25 years. Mean precipitation, soil hydrologic properties, and chaparral community type were significant predictors of recovery in the region. Limited postfire recovery was most acute in transmontane ecotonal chamise sites bounding the Colorado Desert

Présentation (ProQuest)

Page publiée le 26 mai 2021