Published online 25 May 2007
Published in J Environ Qual 36:1105-1111 (2007)
DOI: 10.2134/jeq2006.0466
© 2007 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
Spatial Analysis of a Large Magnitude Erosion Event Following a Sierran Wildfire
Erin M. Carroll,
Wally W. Miller*,
Dale W. Johnson,
Laurel Saito,
Robert G. Qualls and
Roger F. Walker
Dep. of Natural Resources & Environmental Science, College of Agriculture, Biotechnology, and Natural Resources, Univ. of Nevada-Reno, 1000 Valley Rd., Reno, NV 89512
* Corresponding author (wilymalr{at}cabnr.unr.edu)
Received for publication October 26, 2006.
 |
ABSTRACT
|
|---|
High intensity wildfire due to long-term fire suppression and heavy fuels buildup can render watersheds highly susceptible to wind and water erosion. The 2002 "Gondola" wildfire, located just southeast of Lake Tahoe, NV-CA, was followed 2 wk later by a severe hail and rainfall event that deposited 7.6 to 15.2 mm of precipitation over a 3 to 5 h time period. This resulted in a substantive upland ash and sediment flow with subsequent down-gradient riparian zone deposition. Point measurements and ESRI ArcView were applied to spatially assess source area contributions and the extent of ash and sediment flow deposition in the riparian zone. A deposition mass of 380 Mg of ash and sediment over 0.82 ha and pre-wildfire surface bulk density measurements were used in conjunction with two source area assessments to generate an estimation of 10.1 mm as the average depth of surface material eroded from the upland source area. Compared to previous measurements of erosion during rainfall simulation studies, the erosion of 1800 to 6700 g m–2 mm–1 determined from this study was as much as four orders of magnitude larger. Wildfire, followed by the single event documented in this investigation, enhanced soil water repellency and contributed 17 to 67% of the reported 15 to 60 mm ky–1 of non-glacial, baseline erosion rates occurring in mountainous, granitic terrain sites in the Sierra Nevada. High fuel loads now common to the Lake Tahoe Basin increase the risk that similar erosion events will become more commonplace, potentially contributing to the accelerated degradation of Lake Tahoe's water clarity.
|
INTRODUCTION
|
|---|
THE THREAT of wildfire has long been a topic of discussion in the Lake Tahoe Basin. One reason for concern is the intense fire season due to an extremely dry summer climate and the increased likelihood of erosion from exposed, post-wildfire soils, therefore endangering Lake Tahoe water clarity. Areas affected by wildfire are frequently prone to flooding, landslides, and debris and sediment flows (Parrett et al., 2004; Gartner et al., 2004). Post-fire debris and sediment flows have been attributed to several factors including increased post-fire erosion due to lack of vegetation cover and fire-induced subsurface hydrophobic layers that increase the mass wasting potential of overlying wettable soil. The buildup of heavy fuels in Sierran watersheds due to fire suppression increases the likelihood of an intense fire capable of removing most aboveground biomass, leaving the watershed in a highly erosive condition.
The "Gondola" wildfire of 3 July 2002 was followed on 18 July 2002 by a severe hail and rainfall event. The result was a substantive upland ash and sediment flow with subsequent down-gradient riparian zone deposition. The goal of this study was to assess the degree and extent of upland watershed erosion following the first major post-wildfire precipitation event. Our objectives were: (i) to collect spatially explicit depositional data from the down-gradient riparian zone; and (ii) to apply Geographic Information Systems (GIS) for spatial analysis and the estimation of ash/sediment flow mass and depth of eroded upland source area.
|
MATERIALS AND METHODS
|
|---|
Study Site
The study site is located 2 km southeast of Lake Tahoe, near Stateline, NV (38° 57' N; 119° 55' W). Dominant overstory vegetation consists of decadent white fir (Abies concolor Gord. & Glend.), Jeffrey pine (Pinus jeffreyi Grev. & Balf.), and lesser amounts of sugar pine [Pinus lambertiana Dougl. (Strobus L. Mold.)]. The understory consists of Sierra chinquapin [Castanopsis sempervirens Dudl. (Castanea s. Kell)], currant (Ribes L.), and minor amounts of snow bush (Ceanothus velutinis Dougl. Ex Hook.) and bitterbrush [Purshia tridentata (Pursh) DC]. Soils have developed on granitic parent material, belong to the Cagwin-Rock Outcrop complex, and are classified as mixed Typic Cryopsamments on predominantly north-facing slopes of 10 to 40%.
The Gondola wildfire burned through a study area originally designed and instrumented to examine the effects of mechanical harvest and prescribed fire, and their interaction on forest health, nutrient cycling, and discharge water quality. The wildfire took place following the collection of baseline vegetation and soil samples from the overall study site. During the wildfire, which consumed approximately 272 ha (USDA-FS, 2002), wind speeds remained below 24.14 km h–1, relative humidity ranged from approximately 20 to 50%, and air temperature ranged from 7.2 to 23.9°C (Fire and Aviation Management Web Applications, 2002). Estimated burn severity in the vicinity of the study plots and erosion event was considered moderate (Fig. 1
); i.e., areas where the surface litter is consumed and the O horizons are deeply charred or consumed, the soils have reached temperatures of 100 to 200°C at the 1-cm depth but are not visibly altered, woody debris is mostly consumed, and lethal temperatures for soil organisms have occurred to depths of 3 to 5 cm (D. Downie, personal communication, 2007). At our location, most of the aboveground biomass was consumed changing from approximately 7% vegetative ground cover pre-fire to approximately 0% post-fire (Dr. R.F. Walker, University of Nevada, Reno; unpublished data, 2002), and a hydrophobic layer approximately 10 cm below the soil surface was created.
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 1. Rough estimate of burn severity for the Gondola wildfire compared to the potential source area as estimated from ocular assessment (D. Downie, Lake Tahoe Basin Management Unit, unpublished data, 2002).
|
|
The wildfire was followed by a major precipitation event on 18 July 2002. Data from the California Department of Water Resources, Division of Flood Management indicated that precipitation quantities ranged from 7.6 to 15.2 mm, and that the event lasted approximately 3 to 5 h. The variability of the precipitation quantities recorded throughout the area is attributed to the fact that mountainous regions of the Sierra Nevada commonly receive widely scattered, yet intense storms. The event was a mixture of hail and rain (Fig. 2
), and fresh ash and sediment from the burned, upland areas were washed down-gradient and deposited into a riparian area consisting of well-vegetated quaternary alluvium on slopes of 0 to 5%. There was no evidence of erosion from the nearby unburned area; hence the erosion event was strictly limited to the burned site. The depositional area contains an ephemeral drainage, with buried channel characteristics indicated (i.e., sediment accumulation from previous events has buried stumps and to some extent shrub branches), which discharges directly into Edgewood Creek and ultimately into Lake Tahoe.
View larger version (157K):
[in this window]
[in a new window]
|
Fig. 2. Erosion rills formed by runoff during severe precipitation event on 18 July 2002 at Gondola Wildfire field site (Photo by Jay Howard).
|
|
Data Collection
The perimeter of the ash and sediment deposition within the riparian zone was delineated by ocular assessment and recorded using a handheld Global Positioning System (GPS), after which eight transects were established that included a total of 17 individual transect points (Fig. 3
). Transects were chosen to best characterize the variability of the overall deposition within the constraints of accessibility associated with fallen timber and debris, and as allowed by the USFS fire team. A GPS reading was taken at each transect point. Ash and fresh sediment deposits were collected at each of the 17 point locations within the general ash/sediment depositional area. Ash and sediment deposits from the most recent erosion event were easily identified compared to those from previous events and underlying soil by color, consistency, and the boundary formed by live vegetative growth now rooted in previous alluvial depositions. Samples were collected using a standard core sampler for bulk density determination (Klute, 1986). In addition, ash/sediment depth was measured at points spaced every 2 m along the length of each transect. Deposition thickness, area of deposition, and bulk densities were then used to calculate an estimated volume and total mass of the ash/sediment flow. Estimated minimum and maximum source areas were delineated from an ocular assessment, and GPS points were taken of perimeters.
View larger version (47K):
[in this window]
[in a new window]
|
Fig. 3. Ash/sediment depositional area, transects, and data collection points in riparian zone of Gondola Wildfire field site.
|
|
Ash/sediment samples for nutrient analysis were taken immediately adjacent each bulk density sample along the transects. Pre-fire upland soil samples were taken at each of the 16 original study plots in May 2002. Additional upland soil samples were collected 1 yr post-fire (July 2003) (Murphy et al., 2006). Data from the A11 (0–10 cm depth) soil samples were used for purposes of constructing the N balance reported in this study. Samples were homogenized, oven-dried at 55°C, and passed through a 2 mm standard sieve. The <2 mm size fraction was sent to the Oklahoma State University Soil, Water, and Forage Analytical Laboratory (Stillwater, OK) and analyzed for total nitrogen (N) using a dry combustion C and N analyzer (LECO, St. Joseph, MI).
GIS Analysis
Data points acquired applying GPS technology were implemented to create initial point features in ArcView v.3.3 GIS software (Table 1). The attribute tables for all point features were updated to include fields for ash depth and bulk density measurements. Features were then built on to create more instrumental features that ultimately led to the development of polygon features containing all attribute information. Irregularities within the deposition as well as the logistically constrained layout of transects prohibited the application of kriging when interpolating the depositional depth and bulk density measurements from point data. Because of having to apply the spline interpolation method rather than kriging, error estimates were not available for the overall spatial analysis. Error estimates were also not available with the ocular assessments of source and depositional areas.
View this table:
[in this window]
[in a new window]
|
Table 1. Summary of features created in GIS, from what they were created, the process by which they were created, and the values that were added to the attribute tables.
|
|
While working with the data to characterize the natural system, it became necessary to make some assumptions and adjustments. For example, the zero points along the outskirts of the ash/sediment flow created an odd contour (weaving in and out), so the contour was adjusted to follow the flow of other adjacent contours. Some areas outside and gaps just within the zero contour had negative value contours. Since it is impossible to have a negative amount of deposition, and no erosion from the depositional area itself was apparent, these contours were deleted. Once these adjustments were implemented, a smooth, workable contour feature remained (Fig. 4A
). The bulk density contours that were created were then clipped to the shape of the ash depth contours using a polygon created from the outer or zero contour of the ash/sediment depth contours (Fig. 4B).
View larger version (64K):
[in this window]
[in a new window]
|
Fig. 4. (A) Ash/sediment depositional depth contours (cm) in riparian zone of Gondola Wildfire field site, as derived from field data and ESRI ArcView analysis. (B) Ash/sediment depositional bulk density contours (g cm–1) in riparian zone of Gondola Wildfire field site, as derived from field data and ESRI ArcView analysis.
|
|
The attribute table from the final intersection polygon feature was used to calculate additional information. The script [Shape].ReturnArea was used to calculate the area for each polygon created by the intersection, as well as for the minimum and maximum source area polygons. The following equation was applied to determine the mass of the ash and sediment deposition:
 | [1] |
where md is the total deposition mass, Ai is the area of each individual polygon within the intersection feature, di is the estimated depth of deposition within each individual polygon within the intersection feature, and ßi is the estimated bulk density within each individual polygon within the intersection feature.
Nitrogen Balance Analysis
Because we had the means (i.e., data availability), an alternate method of estimating the upland source area and the corresponding depth of erosion was applied to compare and test the reliability of the initial ocular assessment. Total nitrogen values from the source area before the wildfire, 1 yr post-fire, and from the erosion event were used to determine the net loss of total nitrogen from the O and A11 horizons (Murphy et al., 2006). Two assumptions were necessary in balancing the source area nitrogen with the depositional ash and sediment nitrogen: (i) all N in the consumed vegetation and O horizon was completely volatilized; and (ii) soil N in the mineral surface horizon remained unaltered. Hence, any loss of soil N from the upland burn area would be due to erosion, and N measured in ash and sediment deposition could be attributed to sediment, but not ash, deposition.
|
RESULTS
|
|---|
GIS Analysis
Delineation of the depositional zone resulted in an estimated deposition area of 0.82 ha. From Eq. [1], the total mass of ash and sediment deposited over the area was determined to be about 380 Mg. Although the ash and mineral deposition from upland runoff was now reasonably quantified, the specific source area from which it was derived was much less certain. The following equation was initially applied to estimate the depth of erosion from the upland source area:
 | [2] |
where As is the source area from ocular assessment, ßs is the measured bulk density of the source area, and ds is the depth of erosion. This approach produced an estimated erosion depth ranging from 11 to 14 mm (Table 2).
View this table:
[in this window]
[in a new window]
|
Table 2. Summary of variable input and results from GIS analysis, N balance analysis, and an average of the two methods.
|
|
Nitrogen Balance Analysis
The results from the total N analysis for the depositional area were applied to calculate a total N mass for the deposit. An average total N of 3.9 g kg–1 and a depositional mass of 380 Mg were found to contain approximately 1496 kg total N (Table 3). By assuming that all of the N content of the burned vegetation and O horizon materials was volatilized and left the system, and that no N in the mineral soil was volatilized, the total N content of the ash/sediment deposition could be considered as derived from upland soil erosion. Using the pre-fire and post-fire N values reported by Murphy et al. (2006), the estimated total N loss from the A11 horizon due to erosion was about 291 ± 170 kg ha–1. When the following equation was applied:
 | [3] |
where mn is the mass of N in the depositional area and Ln is the loss of N from the source area, a down-gradient deposition of 1496 kg N indicated a source area of 3.2 to 12.4 ha, much larger than that which was estimated from the ocular assessment technique. Again applying Eq. [2] now produced an estimated depth of surface material lost from the eroded upland area ranging from 3 to 13 mm (Table 2). The findings from both scenarios overlap (11–14 mm ocular and 3–13 mm nutrient budget) and yield an overall mean of about 10.1 mm.
|
DISCUSSION
|
|---|
Although overlapping, the ocular assessment of source area produced a much smaller range (11–14 mm) in upland erosion than did the N balance analysis (3–13 mm). This is due, in part, to the large standard error associated with the measured N values from the A11 soil horizon (Murphy et al., 2006). When considering an individual N balance for each of the affected upland study plots (Table 4), however, the amount of erosion appears quite irregular across the overall study area and suggests that the source area actually consists of a series of blocks and rills that exhibit different amounts of erosion. This premise is supported by what is observed in Fig. 2. Hence, the larger range of depths of eroded surface material associated with the N balance analysis is probably a more accurate portrayal of what actually took place at the study site.
The estimated erosion rates from this study were first compared to single event erosion rates on disturbed plots from previously reported rainfall simulation studies in the Sierras (Guerrant et al., 1991; Naslas et al., 1994) and following a wildfire in New Mexico (Johansen et al., 2001). In general, we found the estimated erosion from the post-wildfire erosion event in this study to be two to four orders of magnitude larger than those estimated using rainfall simulation techniques (Table 5). This highlights the fact that small plots used in rainfall simulation studies typically do not generate the same flow velocities and erosive power as do the natural events that generate overland flow over an entire watershed. It is also possible that the source area in this study was actually much larger than estimated.
View this table:
[in this window]
[in a new window]
|
Table 5. Summary of literature comparison for single storm event erosion, on disturbed soils, using rainfall simulators.
|
|
Consequently, we next compared our erosion estimates to erosion rates reported in the literature for undeveloped upland Sierran soils receiving average annual precipitation (about 60 to 110 cm yr–1) over a 1000-yr span. Based on the results from both source area assessments (ocular and N balance), the single erosion event following wildfire contributed a minimum of 5% to a maximum of 93% of the erosion expected to occur over a 1000-yr span (Table 6). Applying the overall mean of 10.1 mm for estimated erosion, the respective single event contribution ranged from a minimum of 17% to a maximum of 67%. In either case, it appears that the immediate impact of a single erosion event following a wildfire can be quite severe relative to the expected long-term erosion from 1000 yr of average annual precipitation.
View this table:
[in this window]
[in a new window]
|
Table 6. Summary of results for differing methods of assessing source area and depth of erosion from the source area, and their comparison to reported erosion rates in the Sierra Nevada (Riebe et al., 2001; Granger et al., 2001) to determine the percent of total expected erosion for 1000 yr that occurred in just one storm.
|
|
|
CONCLUSIONS
|
|---|
GIS is a useful tool that is increasingly becoming a necessity for the analysis of spatially explicit data. In this study, we were able to: (i) apply spatial analysis tools to quantify the deposition mass from a large runoff and erosion event 2 wk post-wildfire; (ii) estimate the amount of upland surface material lost; (iii) compare our estimates to other post-disturbance erosion quantities reported in the literature; and (iv) bracket the magnitude of a single event erosion episode relative to the expected erosion over 1000 yr of average annual precipitation in the Tahoe Basin and vicinity. Although quantification of the source area is paramount to understanding the actual scale of erosion, we were unable to delineate it within a high degree of certainty. It is worthy of note, however, that the measurements we made were completely free of any artifacts that could be introduced by barriers, weirs, or other collection devices (i.e., this was a completely natural event). Furthermore, we are unaware of any similar or more quantitative data for post-wildfire erosion within the Tahoe Basin watershed. The estimated impact of this single erosion event following wildfire was at least an order of magnitude greater than the expected average annual erosion based on the 1000 yr projections (i.e., 5% vs. 0.1% annually, respectively) reported in the literature.
While the bulk of the ash and sediment remained in the riparian zone rather than flushing from the watershed, it is seen as an important happenstance for two reasons. First, if the event had occurred in a watershed with a damaged stream environment zone or with direct drainage discharge into Lake Tahoe, a large quantity of sediments and associated nutrients would have emptied into the Lake and undoubtedly affected water clarity. Second, while the sediment and ash were not lost entirely from the watershed, they were lost to the upland soils. Because Sierran soils are typically thin with low nutrient content, the type of erosion exhibited in this study can greatly inhibit post-fire revegetation and overall watershed health. As fuel loads continue to increase in the Lake Tahoe Basin and vicinity, so does the risk of wildfire, erosion events of similar magnitude, and the potential consequence to lake clarity.
|
ACKNOWLEDGMENTS
|
|---|
We gratefully acknowledge the participation and support of the Nevada Agricultural Experiment Station, McIntyre-Stennis and the U.S. Forest Service Lake Tahoe Basin Management Unit and Joint Fire Sciences Program.
|
NOTES
|
|---|
A contribution of the Department of Natural Resources & Environmental Science. Research supported in part by Nevada Agricultural Experiment Station, publication # 52066959.
|
REFERENCES
|
|---|
- Fire and Aviation Management Web Applications. 2002. Approved Incident 209 Reports for CA-TMU-09828 [Online]. National Interagency Fire Center, Boise, ID. Available at http://famweb.nwcg.gov/hist_209/r_list_209s?button=Northern+California&v_gaid=NO &v_209_number=CA-TMU-09828 (verified 4 Apr. 2007).
- Gartner, J.E., E.R. Bigio, and S.H. Cannon. 2004. Compilation of post-wildfire runoff-event data from the Western United States. Open File Rep. 2004-1085. USGS, Reston, VA.
- Granger, D.E., C.S. Riebe, J.W. Kirchner, and R.C. Finkel. 2001. Modulation of erosion on steep granitic slopes by boulder armoring, as revealed by cosmogenic 26Al and 10Be. Earth Planet. Sci. Lett. 186:269–281.[CrossRef]
- Guerrant, D.G., W.W. Miller, C.N. Mahannah, and R. Narayanan. 1991. Site-specific erosivity evaluation of a Sierra Nevada forested watershed. J. Environ. Qual. 20:396–402.[Web of Science]
- Johansen, M.P., T.E. Hakonson, and D.D. Breshears. 2001. Post-fire runoff and erosion from rainfall simulation: Contrasting forests with shrublands and grasslands. Hydrol. Processes 15:2953–2965.[CrossRef]
- Johnson, D.W., J.D. Murphy, R.F. Walker, D.S. Glass, and W.W. Miller. 2007. Wildfire effects on forest carbon and nutrient budgets. Ecol. Eng. (in press).
- Klute, A. (ed.). 1986. Methods of soil analysis: Part 1–Physical and mineralogical methods. 2nd ed. ASA, CSSA, and SSSA, Madison, WI.
- Parrett, C., S.H. Cannon, and K.L. Pierce. 2004. Wildfire-related floods and debris flows in Montana in 2000 and 2001. Water Resources Investigations Rep. 03-4319. USGS, Reston, VA.
- Murphy, J.D., D.W. Johnson, W.W. Miller, R.F. Walker, E.M. Carroll, and R.R. Blank. 2006. Wildfire effects on soil nutrients and leaching in a Tahoe Basin watershed. J. Environ. Qual. 35:479–489.[Abstract/Free Full Text]
- Naslas, G.D., W.W. Miller, G.F. Gifford, and G.C.J. Fernandez. 1994. Effects of soil type, plot condition, and slope on runoff and interrill erosion of two soils in the Lake Tahoe basin. Water Resour. Bull. 30:319–328.
- Riebe, C.S., J.W. Kirchner, D.E. Granger, and R.C. Finkel. 2001. Minimal climatic control on erosion rates in the Sierra Nevada, California. Geology 29:447–450.[Abstract/Free Full Text]
- USDA-FS. 2002. Burned area report. Reference FSH 2509.13. United States Department of Agriculture Forest Service, Washington, DC.
TOP
ABSTRACT
INTRODUCTION
