Ecological Archives E094-227-A3

Brian J. Harvey, Daniel C. Donato, William H. Romme, Monica G. Turner. 2013. Influence of recent bark beetle outbreak on fire severity and postfire tree regeneration in montane Douglas-fir forests. Ecology 94:2475–2486. http://dx.doi.org/10.1890/13-0188.1

Appendix C. Information on fire severity measurements.

Surrogate measurements of fire severity

We also quantified fire severity for each plot from a satellite-derived index of fire severity: the Relative differenced Normalized Burn Ratio (RdNBR) downloaded from the Monitoring Trends in Burn Severity website (www.mtbs.gov). This commonly used, surrogate metric provides a reliable measure of fire severity by generating an index of change from pre- to post-fire based on reflectance in Landsat TM bands 4 and 7, which represent chlorophyll content and moisture in vegetation/soils, respectively (Miller and Thode 2007). Lower RdNBR values represent less pre- to post-fire change in live green vegetation, moisture content, and soil conditions (i.e., less-severe fire), whereas higher RdNBR values represent more change (i.e., more-severe fire). A single RdNBR value was extracted for each plot in ArcGIS 10.1 based on the pixel encompassing the plot center.

We attempted to reduce the potential redundancy of many response variables through a factor analysis; however this procedure did not produce interpretable synthetic variables. While RdNBR is, in a sense, a synthetic variable that should broadly represent important metrics of fire severity (Miller et al. 2009), we found modest Spearman rank correlations between RdNBR and other metrics of fire severity (Table C1). Therefore, we present analyses for RdNBR and field fire severity metrics.

Table C1. Spearman rank correlations (rs) between fire severity response variables used to test for effects of beetle-killed basal area on fire severity. All correlations are significant (P < 0.001).

 

RdNBR

Litter +
duff depth (mm)

Charred
surface
cover (%)

Mean
char
height (m)

Mean
bole
scorch (%)

Basal
area killed
by fire (%)

Trees
killed
by fire (%)

 

 

 

 

 

 

 

 

RdNBR

1.00

-0.57

0.58

0.49

0.46

0.47

0.41

Litter + duff depth (mm)

 

1.00

-0.62

-0.64

-0.52

-0.59

-0.57

Charred surface cover (%)

 

 

1.00

0.45

0.44

0.45

0.44

Mean char height (m)

 

 

 

1.00

0.77

0.79

0.76

Mean bole scorch (%)

 

 

 

 

1.00

0.83

0.80

Basal area killed by fire (%)

 

 

 

 

 

1.00

0.97

Trees killed by fire (%)

 

 

 

 

 

 

1.00

Quantitative field measurements of fire severity across fire severity classes

Soil charring (if present) was measured to the nearest mm at every 3 m along the main axis of the plot (20 points per plot), but was excluded from analysis due to the potential confounding effects of erosion/deposition that may have occurred between the time of fire and sampling. We recorded whether there was deep charring (charring into the sapwood) on the bole of the tree using the following four categories: no deep char (0), deep char on < 50% of the bole circumference (1), deep char on > 50% of the bole circumference but only on the lower portion of the bole (not into the crown) (2), and deep char on > 50% of the bole circumference and into the crown (3). Charring scores were averaged across all trees in a plot. Fire-severity classes were defined following (Turner et al. 1999): ‘light surface fire’ – fire is not stand-replacing with some surviving canopy trees; ‘severe surface fire’ – fire is stand-replacing (100% overstory mortality) but canopy-tree needles are not consumed in fire; ‘crown fire’ – fire is stand-replacing, canopy-tree needles are consumed by fire, mineral soil is exposed, and char heights reach top of tree boles. Fire severity classes were statistically distinct for quantitative measures of fire effects (Table C2).

Table C2. Differences in quantitative fire severity metrics across fire-severity classes. Values are means (SE). Significant differences among classes for each variable (Tukey’s HSD test, P < 0.05) are denoted using subscript letters (a,b,c).

 

Fire-severity class

 

Light surface

 

Severe surface

 

Crown

Fire severity metric

n = 23

 

n = 31

 

n = 31

 

 

 

 

 

 

 

 

 

Mean char height (m)

3.3a

(0.5)

 

12.4b

(0.9)

 

20.6c

(0.3)

Mean deep char level (trees) (0–3)

0.15a

(0.03)

 

0.47b

(0.06)

 

0.81c

(0.09)

Litter + duff depth (mm)

15.8a

(1.7)

 

7.3b

(0.9)

 

3.8c

(0.4)

Charred surface cover (%)

16a

(2)

 

22b

(2)

 

31c

(2)

Literature cited

Miller, J. D., E. E. Knapp, C. H. Key, C. N. Skinner, C. J. Isbell, R. M. Creasy, and J. W. Sherlock. 2009. Calibration and validation of the relative differenced Normalized Burn Ratio (RdNBR) to three measures of fire severity in the Sierra Nevada and Klamath Mountains, California, USA. Remote Sensing of Environment 113:645–656.

Miller, J. D., and A. E. Thode. 2007. Quantifying burn severity in a heterogeneous landscape with a relative version of the delta Normalized Burn Ratio (dNBR). Remote Sensing of Environment 109:66–80.

Turner, M. G., R. H. Gardner, and W. H. Romme. 1999. Prefire heterogeneity, fire severity, and early postfire plant reestablishment in subalpine forests of Yellowstone National Park, Wyoming. International Journal of Wildland Fire 9:21–36.


[Back to E094-227]