Evaluating Post-fire Vegetation Recovery in Canadian Mixed Prairie Using Remote Sensing Approaches

This study investigated a wildfire occurred in April 2013 at Grasslands National Park, aiming to quantify vegetation's post-fire recovery with both field and remote sensing approaches. Biophysical parameters and hyperspectral reflectances were collected through field surveys conducted one year prior to the fire as well as five continuous years post-fire at growing seasons. These data were processed into burned and unburned samples followed by significance test to reveal biophysical differences across samples. Results indicated an overall recovery of the grassland within 4-5 years, with different vegetation forms recovering at various post-fire growing seasons. Green grass was the most resilient component that fully recovered one year post-fire, followed by forbs at two years post-fire, with shrubs and soil organic crust taking longer than four years to recover compared to the adjacent unburned communities. Hyperspectral dataset was used to establish the utility of remote sensing approaches in grasslands fire-study. Results suggested the potential of satellite remote sensing data in such application. Furthermore, Landsat dataset were processed and significance test was repeated to further prove the sensitivity of Landsat product (especially NDVI) in distinguishing burned and unburned samples, as well as good agreement with conclusions established from field data analysis. Finally, major driving factors were analyzed with ANOVA and results indicated the significant role of meteorological variables and topography in vegetation's post-fire recovery. Findings from this research contribute to a better understanding of fire's effect on the under-studied Canadian northern mixed prairie. Also, the successful validation of RS based approaches can provide as the theoretical basis for potential future RS applications in modelling grassland post-fire recovery in the mixed prairie.