Assessing postfire recovery of chamise chaparral using multi-temporal spectral vegetation index trajectories derived from Landsat imagery

Land managers and ecologists working in Mediterranean-type ecosystems require information on postfire recovery of shrublands as a means of identifying long-term changes in these sensitive systems. This study evaluates the utility of postfire regrowth trajectories, derived from multi-temporal Landsat satellite surface reflectance imagery, as a basis for estimating postfire recovery of chamise chaparral in southern California. Postfire recovery metrics are applied to time series trajectories of spectral vegetation indices (SVIs) including NDVI, EVI, SAVI, MSAVI, NBR, and NBR2. Shrub fractional cover changes following postfire recovery are estimated at 74 plots using multi-date high spatial resolution orthoimagery, enabling stratification of the recovery outcomes. The sensitivity of Landsat-based regrowth trajectories to postfire shrub recovery is assessed by ANOVA statistical testing based on the stratified recovery outcomes. We evaluate several combinations of postfire recovery metrics and SVIs as postfire recovery indicators. We compare SVI trends extending 19 to 29years postfire with field-collected data on postfire recovery trends reported in previous studies, in order to assess the sensitivity of the SVIs to short- and long-term phases of regrowth. The utility and necessity of normalizing SVI time series based on unburned control plots to reduce moisture effects is also evaluated. A primary finding is that the Scaled Recovery Metric (SRM), a variant of the Regeneration Index which incorporates the pixel-specific SVI value immediately after fire, is a particularly useful indicator of postfire recovery based on time-sequential SVI trajectories and facilitates inter-site comparisons. Time series of several SVIs (especially NDVI and NBR2) provide statistically significant (p