Agriculture Reference
In-Depth Information
Box 2.1 Use of Landsat dNBR to Determine Fire Severity and Ecosystem Response
in Crown Fire Shrublands (from Keeley
2009
)
In late October 2003 five large wildfires burned more than 200 000 ha in southern
California. A total of 250 tenth-hectare plots were sampled in these burned areas
to assess fire severity and vegetation recovery (Keeley
et al.
2008
). Fire severity
was assessed using the twig diameter method commonly used in crown fire
ecosystems (Moreno & Oechel
1989
) on multiple samples of the same shrub
(
Adenostoma fasciculatum
). Vegetation recovery was based on plant cover in the
first spring following fires. The Landsat TM assessment known as the differenced
Normalized Burn Ratio (dNBR) was provided by the U.S. Geological Survey
EROS data center (Sioux Falls, SD). This remote sensing index was strongly
correlated with field measurement of fire severity (
Fig. B2.1.1a
), explaining over a
third of the variation between these sites. However, when dNBR was used to
predict ecosystem response variables related to recovery there was little to no
relationship. Total vegetative recovery (
Fig. B2.1.1b)
was very weakly related to
dNBR and explained only about 1% of the variation, and there was no signifi-
cant relationship with woody cover (
P
0.94, not shown), or percentage of
resprouting by the dominant shrub (
Fig. B2.1.1c
).
ΒΌ
(a)
(b)
(c)
r
2
= 0.34
P
< 0.000
r
2
= 0.01
P
= 0.04
r
2
= 0.01
P
= 0.11
275
10 0
10 0
220
80
80
165
60
60
110
40
40
55
20
20
0
0
0
0.0 0.3 0.6 0.9 1. 2 1. 5 1. 8
Fire severity (log twig diam)
0
55
110 165 220 275
dNBR
0
55
110 165 220 275
dNBR
Fig. B2.1.1
Relationship between (a) field measurement of fire severity and early
assessment dNBR, (b) dNBR and first-year plant cover and (c) dNBR and postfire resprouting
by the shrub
Adenostoma fasciculatum
.
is the time required to burn the equivalent of a specified area, or
fire return
interval
, which is the spatially explicit time between fires in a specified area,
sometimes expressed as the inter-fire interval. For example, wildlands in southern
California have an average fire rotation interval of 36 yrs, but the actual fire
return interval can vary from fires every few years at some sites to fires every
100 yrs at other sites (Keeley
et al.
1999a
). The variance in fire return interval is
critical to ecosystem resilience and species differ markedly in tolerance to high and
low departures from the mean fire return interval (see
Chapter 3
).
One of the difficulties with precise usage of the term fire frequency is that
many landscapes have a mixed fire regime that comprises both surface and
