ABoVE: Spatial Estimates of Carbon Combustion from Wildfires across SK, Canada, 2015 : Arctic-Boreal Vulnerability Experiment (ABoVE)

This dataset provides spatial estimates of carbon combustion from all 2015 wildfire burned areas across Saskatchewan, Canada, on a 30-m grid. Carbon combustion (kg C/m2) was derived from post-fire field measurements of carbon stocks completed in 2016 at 47 stands that burned during three 2015 Saskat...

Full description

Bibliographic Details
Main Authors: Potter, S., Rogers, B.M., Dieleman, C.
Format: Article in Journal/Newspaper
Language:English
Published: ORNL Distributed Active Archive Center 2020
Subjects:
BIOSPHERE > ECOSYSTEMS > TERRESTRIAL ECOSYSTEMS > FORESTS
BIOSPHERE > VEGETATION > CARBON
HUMAN DIMENSIONS > NATURAL HAZARDS > WILDFIRES > BURNED AREA
LAND SURFACE > SOILS > CARBON
Environmental Modeling > Computer
forest fires
carbon combustion
gridded data
carbon emissions
wildfires
Arctic
Canada
Online Access:https://dx.doi.org/10.3334/ornldaac/1787
https://daac.ornl.gov/cgi-bin/dsviewer.pl?ds_id=1787
Description
Summary:This dataset provides spatial estimates of carbon combustion from all 2015 wildfire burned areas across Saskatchewan, Canada, on a 30-m grid. Carbon combustion (kg C/m2) was derived from post-fire field measurements of carbon stocks completed in 2016 at 47 stands that burned during three 2015 Saskatchewan wildfires (Egg, Philion, and Brady) and at 32 unburned stands in adjacent areas. The study areas covered two ecozones (Boreal Plains and Boreal Shield), two stand-replacing history types (fire and timber harvest), three soil moisture classes (xeric, mesic, and subhygric), and three stand dominance classifications (coniferous, deciduous, and mixed). To spatially extrapolate estimates of combustion to all 2015 fires in Saskatchewan, a predictive radial support vector machine model was trained on the 47 burned stands with associated environmental variables and geospatial predictors and applied to historical fire and harvest areas. The dataset also includes uncertainty estimates represented as per pixel standard deviations of model estimates derived using a Monte Carlo analysis.

Similar Items