Seasonal and topographic effects on estimating fire severity from Landsat TM/ETM+ data

The maximum solar elevation is typically less than 50 degrees in the Alaskan boreal region and solar elevation varies substantially during the growing season. Because of the relatively low solar elevation at boreal latitudes, the effect of topography on spectral reflectance can influence fire severity indices derived from remotely sensed data. We used Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper (ETM) data to test the effect of changing solar elevation and topography on the Normalized Burn Ratio (NBR) and the differenced Normalized Burn Ratio (dNBR). When a time series of unburned pixels from black spruce forests was examined, we found that NBR values consistently decreased from June through September. At the stand level, dNBR-derived values from similar unburned and burned black spruce stands were substantially higher from September imagery relative to July or August imagery. Within the Boundary burn, we found mean post-fire NBR to consistently vary owing to topographic control of potential solar radiation. To minimise spectral response due to topographic control of vegetation and fire severity, we computed a dNBR using images from August and September immediately after a June–July wildfire. There was a negative bias in remotely sensed fire severity estimates as potential solar radiation decreased owing to topography. Thus fire severity would be underestimated for stands in valley bottoms dominated by topographic shading or on steep north-facing slopes oriented away from incoming solar radiation. This is especially important because highly flammable black spruce stands typically occur on such sites. We demonstrate the effect of changing pre- and post-fire image dates on fire severity estimates by using a fixed NBR threshold defining ‘high severity’. The actual fire severity was constant, but owing to changes in phenology and solar elevation, ‘high severity’ pixels within a burn ranged from 56 to 76%. Because spectral reflectance values vary substantially as solar elevation and plant phenology change, the use of reflectance-based indices to assess trends in burn severity across regions or years may be especially difficult in high-latitude areas such as the Alaskan boreal forest.

[1]  J. W. Wagtendonk,et al.  Comparison of AVIRIS and Landsat ETM+ detection capabilities for burn severity , 2004 .

[2]  B. Markham,et al.  Revised Landsat-5 TM radiometric calibration procedures and postcalibration dynamic ranges , 2003, IEEE Trans. Geosci. Remote. Sens..

[3]  John R. G. Townshend,et al.  The Landsat Tutorial Workbook: Basics of Satellite Remote Sensing , 1984 .

[4]  S. Howard,et al.  An Evaluation of Gap-Filled Landsat SLC-Off Imagery for Wildland Fire Burn Severity Mapping , 2004 .

[5]  A. McGuire,et al.  Topographic influences on wildfire consumption of soil organic carbon in interior Alaska: Implications for black carbon accumulation , 2007 .

[6]  Stephen R. Yool,et al.  Mapping Fire-Induced Vegetation Mortality Using Landsat Thematic Mapper Data: A Comparison of Linear Transformation Techniques , 1998 .

[7]  D. Civco.  Topographic normalization of landsat thematic mapper digital imagery , 1989 .

[8]  T. Landmann,et al.  Characterizing sub-pixel Landsat ETM+ fire severity on experimental fires in the Kruger National Park, South Africa. , 2003 .

[9]  E. Kasischke,et al.  Evaluating the potential of Landsat TM/ETM+ imagery for assessing fire severity in Alaskan black spruce forests , 2008 .

[10]  Jay D. Miller,et al.  Quantifying burn severity in a heterogeneous landscape with a relative version of the delta Normalized Burn Ratio (dNBR) , 2007 .

[11]  P. Rich,et al.  A geometric solar radiation model with applications in agriculture and forestry , 2002 .

[12]  E. Kasischke,et al.  Evaluation of the composite burn index for assessing fire severity in Alaskan black spruce forests , 2008 .

[13]  Sandra L. Haire,et al.  Lessons From the Fires of 2000: Post-Fire Heterogeneity in Ponderosa Pine Forests , 2003 .

[14]  Anthony C. Janetos,et al.  Using Remote Sensing to Assess Russian Forest Fire Carbon Emissions , 2002 .

[15]  C. Justice,et al.  An examination of spectral band ratioing to reduce the topographic effect on remotely sensed data , 1981 .

[16]  David Riaño,et al.  Assessment of different topographic corrections in Landsat-TM data for mapping vegetation types (2003) , 2003, IEEE Trans. Geosci. Remote. Sens..

[17]  R. Colwell.  Remote sensing of the environment , 1980, Nature.

[18]  Eric S. Kasischke,et al.  Constraints on using AVHRR composite index imagery to study patterns of vegetation cover in boreal forests , 1997 .

[19]  N. Benson,et al.  Landscape Assessment: Ground measure of severity, the Composite Burn Index; and Remote sensing of severity, the Normalized Burn Ratio , 2006 .

[20]  S. Flasse,et al.  Characterizing the spectral-temporal response of burned savannah using in situ spectroradiometry and infrared thermometry , 2000 .

[21]  S. Running,et al.  Remote Sensing of Forest Fire Severity and Vegetation Recovery , 1996 .

[22]  S. Ekstrand,et al.  Landsat TM-based forest damage assessment : correction for topographic effects , 1996 .

[23]  David P. Roy,et al.  Remote sensing of fire severity: assessing the performance of the normalized burn ratio , 2006, IEEE Geoscience and Remote Sensing Letters.

[24]  D. Verbyla,et al.  Evaluation of remotely sensed indices for assessing burn severity in interior Alaska using Landsat TM and ETM , 2005 .