|
|
|
Field validation of Burned Area Reflectance Classification (BARC) products for post fire assessment
Hudak, A.T.; Robichaud, P.R.; Evans, J.B.; Clark, J.; Lannom, K.; Morgan, P.; and Stone, C. 2004.
Field validation of Burned Area Reflectance Classification (BARC) products for post fire assessment.
In: Remote Sensing for Field Users,
Proceedings of the Tenth Forest Service Remote Sensing Applications Conference,
Salt Lake City, Utah, April 5-9, 2004.
CD-ROM. 13 p.
Keywords: BARC
Links:
PDF [790K]
Abstract:
The USFS Remote Sensing Applications Center (RSAC) and the USGS EROS Data Center (EDC) produce Burned
Area Reflectance Classification (BARC) maps for use by Burned Area Emergency Rehabilitation (BAER) teams in
rapid response to wildfires. BAER teams desire maps indicative of soil burn severity, but photosynthetic and non-photosynthetic
vegetation also influences the spectral properties of post-fire imagery. Our objective was to assess
burn severity both remotely and on the ground at six 2003 wildfires. We analyzed fire-effects data from 34 field
sites located across the full range of burn severities observed at the Black Mountain Two, Cooney Ridge, Robert,
and Wedge Canyon wildfires in western Montana and the Old and Simi wildfires in southern California. We
generated Normalized Burn Ratio (NBR), delta Normalized Burn Ratio (dNBR), and Normalized Difference
Vegetation Index (NDVI) indices from Landsat 5, SPOT 4, ASTER, MASTER and MODIS imagery. Pearson
correlations between the 44 image and 79 field variables having an absolute value greater than 0.5 were judged
meaningful and tabulated in overstory, understory, surface cover, and soil infiltration categories. Vegetation
variables produced a higher proportion of meaningful correlations than did surface cover variables, and soil
infiltration variables the lowest proportion of meaningful correlations. Soil properties had little measurable
influence on NBR, dNBR or NDVI, particularly in low and moderate severity burn areas where unconsumed
vegetation occludes background reflectance. BAER teams should consider BARC products much more indicative of
post-fire vegetation condition than soil condition. Image acquisition date, in relation to time of field data collection
and time since fire, appears to be more important than type of imagery or index used. We recommend preserving
the raw NBR or dNBR values in an archived map product to enable remote monitoring of post-fire vegetation
recovery. We further recommend that BAER teams rely on the continuous BARC-Adjustable (BARC-A) product
(and assign their own severity thresholds as needed) more than the classified BARC product, which oversimplifies
highly heterogeneous burn severity characteristics on the ground.
Moscow FSL publication no. 2004n
|