facility. Specifically, Australia Geosciences assisted the effort 
by providing near real time access to Landsat 5 and 7 
acquisitions for the burned area. Appropriate Landsat scenes 
were identified using the image browsing capabilities of the 
ACRES Digital Catalogue. Scenes were then transferred by file 
transfer protocol (FTP) to EROS/RSAC staff in the United 
States for generation of the burn mapping products. Once 
generated, products were FTP transferred or emailed directly to 
BAER team members in Australia for use in field assessments 
and planning activities. 
2. METHODOLOGY 
2.1 Study Area and Products 
In 2009, EROS/RSAC received requests from US BAER teams 
deployed to Victoria to provide burn severity maps for several 
fires (or fire complexes). BAER team staff worked closely with 
local land and fire managers. Specifically, assessments for the 
following fires were requested: 
Beechworth 
Bunyip 
Churchill-Delburn 
East Tyer-Thompson 
Murrindindi 
Wilson Promontory 
The spatial relationship of individual mapped fires to the 
Melbourne urban area and a Landsat image mosaic that 
provided coverage for path 92 and rows 85-87 is shown in 
Figure 1. Data layers generated for all Victoria fires included 
the following: a pre-fire image, a post-fire image as shown in 
Figure 2, continuous burn severity data in the form of a 
differenced normalized burn ratio (dNBR) (Eas. (1) and (2)) or 
the differenced normalized difference vegetation index 
(dNDVI) (Eqs. (3) and (4)) image, and a classified dNBR or 
dNDVI image displaying low, moderate and high soil burn 
severity as shown in Figure 3. 
2.2 Satellite Image Processing 
Burn mapping products were generated for the 2009 Victoria 
bushfires using the same methodology that has been used in the 
United States to support BAER teams since the 2003 fire 
season. These mapping techniques are based upon the 
normalized burn ratio (NBR) and dNBR (Key, 2006). The NBR 
is a remote sensing image derivative that exploits the 
characteristics of the near-infrared and short-wave infrared 
portions of the electromagnetic spectrum as they have proven to 
be good discriminators of burn scars and the mosaic of 
severities that typically occur within a burn perimeter. The 
dNBR compares NBR imagery acquired before the fire with 
imagery of the same area acquired immediately after the fire to 
identify the location and magnitude of changes in vegetation. 
The NBR is computed using Landsat Enhanced Thematic 
Mapper (ETM) or Thematic Mapper (TM) near-infrared (NIR) 
and short-wave infrared (SWIR) spectral bands (4 and 7) 
respectively. The NBR is calculated as follows: 
NBR = (NIR — SWIR) /(NIR + SWIR) (1) 
52 
  
  
  
62 Kiometers 
i ; 
Figure 1. Overview of 2009 mapped fires (bright red) and their 
proximity to the Melbourne urban area and Landsat path 92 and 
rows 85-87 (yellow line grid). 
  
> 1 2 3 Xoetere 
eurer anebrnds bis ene 
Figure 2. Delburn post-fire image acquired 02 Feb 2009. 
Figu 
By co 
the da 
the NI 
image 
follow 
dNBR 
After 
and d 
the U 
pre-fii 
chang 
heavil 
veget: 
increa 
with | 
loss a 
pre-fu 
veget: 
such 
mana; 
Lands 
spectr 
the d 
and R 
indice 
(NDV 
veget: 
early 
comp 
time. 
(NIR) 
respei 
NDVI