Intense forest wildfire sharply reduces mineral soil C and N: the first direct evidence
Bernard T. Bormann, Peter S. Homann, Robyn L. Darbyshire, and Brett A. Morrissette. 2008. Intense forest wildfire sharply reduces mineral soil C and N: the first direct evidence. Can. J. For. Res. 38: 2771–2783 (2008).
Full text [here]
Selected excerpts:
Abstract:
Direct evidence of the effects of intense wildfire on forest soil is rare because reliable prefire data are lacking. By chance, an established large-scale experiment was partially burned in the 2002 Biscuit fire in southwestern Oregon. About 200 grid points were sampled across seven burned and seven unburned stands before and after the fire. Fire-related soil changes — including losses of soil organic and inorganic matter — were so large that they became complicated to measure. The 51 Mg ha–1 of loose rocks on the soil surface after fire suggests erosion of 127 Mg ha–1 of fine mineral soil, some of which likely left in the fire plume. After accounting for structural changes and erosion with a comparable-layers approach, combined losses from the O horizon and mineral soil totaled 23 Mg C ha–1 and 690 kg N ha–1, of which 60% (C) and 57% (N) were lost from mineral horizons. Applying a fixed-depth calculation — commonly used in previous fire studies — that disregards structural changes and erosion led to underestimates of loss of nearly 50% for C and 25% for N. Although recent debate has centered on the effects of postwildfire forest management on wood, wildlife habitat, and fuels, this study indicates that more consideration should be given to the possible release of greenhouse gases and reduction of future forest productivity and CO2 uptake.
Introduction
Forest ecologists think of wildfire as an important natural process that regulates fuel accumulation and successional patterns across most western US forests (DeBano et al. 1998). Forest wildfires also have great societal consequences. Rural communities and firefighters are well aware of the dangers of high-intensity (high-temperature) large-scale fires. Given dry conditions and sufficient fuels, these fires can make their own weather, spread at alarming rates, and often become nearly unstoppable. The monetary and human costs of fighting such fires — loss of property, timber, wildlife habitat, water quality, C stocks, and other resource values, and remediation expenses — can be substantial (Neuenschwander et al. 2000; Dombeck 2001). The direct cost of fighting wildfires nationally in 2002 was $1.6 billion, with nearly a tenth of this budget spent on a single fire, reported on here, called the Biscuit complex fire, in southwestern Oregon, USA (Government Accounting Office 2004). …
Current understanding of the ecological effects of intense wildfire is severely limited by the lack of detailed knowledge of soils before wildfire. To date, only one study has reported effects of wildfire based on before and after soil sampling, but fire intensity was unknown (Murphy et al. 2006; Johnson et al. 2007). …
When the 2002 Biscuit fire serendipitously burned through part of a 150 ha Long-Term Ecosystem Productivity (LTEP) experiment (Bormann et al. 1994; Homann et al. 2008), we were given an opportunity to examine soil changes in paired pre- and post-wildfire samples. Because we had clear evidence that these plots had burned at high intensity, our objective was to determine the effects of high-intensity fire on the loss of soil C and N mass. Volumetric soil sampling, which has not been conducted in previous wildfire studies, allowed us to rigorously evaluate changes on a per-area basis. The parallel data from unburned stands provided important evidence on soil development that was not attributable to the fire. Having archived prefire samples also allowed us to reduce analytical biases. By combining these data, we provide an assessment of the potential effects of wildfire that is the most rigorous to date. Our assessment of soil physical and chemical changes provides unique information about the effects of high-intensity forest wildfire on soils and greenhouse-gas emissions, along with important implications for long-term productivity and future C sequestration. …
Results
The clearest effect of intense wildfire on our plots, which has also been widely noted across the Biscuit fire,2 was a substantial increase in the amount of near-surface rocks on the burned plots — there was 51 ± 8 M ha–1 (mean ± 95% CI) of loose rocks on the soil surface after the fire. (Fig. 5). …
The second notable effect of high-intensity fire was the major loss of soil organic matter at the soil surface that extended into the mineral soil, and corresponding losses of soil C and N. In the uppermost comparable layer 1 (the O horizon and mineral soil to 3.7 cm), soil C decreased by 19 ± 2 Mg ha–1 from the prefire sampling value in 1992 (Fig. 7 left side, Table 4 method 1). A small but significant (p <0.05) amount of C (2.5 Mg ha–1) was lost from the two deepest layers combined (4 and 5). If all C in the prefire O horizon (9 ± 1 Mg ha–1) was combusted, then 60% of the soil C loss came from mineral soil layers.
Soil N losses were also large, 547 ± 79 kg ha–1. No significant subsurface soil N losses were seen, but an increase of 40 ± 32 kg ha–1 was observed in layer 3. If all N in the prefire O horizon (226 ± 21 kg ha–1) was volatilized, then 57% of the soil N loss came from mineral soil layers. …
Discussion
Comparison to other studies
Our estimated loss of 23 Mg C ha–1 from organic and minerals soil layers is higher than most previous estimates. Our losses of 500 to 700 kg N ha–1 fell in the upper range of reported values. Comparing our results with others is challenging, however, given the variety of assumptions, sampling methods and depths, and analyses used. Uncertainties with these studies come from multiple sources and cloud our knowledge of the effects of intense fire on soils. Because of the lack of opportunities to directly measure soil changes before and after intense wildfire, researchers have had to rely on estimates obtained in retrospective studies or extrapolated from laboratory and lower-intensity prescribed fire studies. …
We cannot rule out a bias in retrospective studies because they assume that unburned areas can be used to represent the preburn conditions (Baird et al. 1999). Inherent differences between burned and unburned areas with respect to moisture, site conditions, and burn history can influence soil properties, as has been demonstrated for part of the Biscuit fire (Thompson et al. 2007). Different site histories confound interpretations and may lead to incorrect conclusions about soil dynamics (Yanai et al. 2003). …
Several mechanisms may explain the loose surface rocks after fire (Fig. 3): postfire erosion of fines, small-scale resorting of soil constituents, and atmospheric losses during the fire. Most of the soil organic matter in the O horizon was burned, and the products of combustion, including CO2 and volatilized nutrients, were exported as gas or smoke particles, leaving behind over seven times more rocks above the mineral soil surface. Losses of fine mineral soil from upper mineral soil layers are usually attributed to postfire waterdriven erosion, and our erosion-box estimates support this explanation to a point. Water-driven erosion for the 2003–2004 water year on burned soils in erosion boxes, placed across a range of slopes, averaged 57 m3 ha–1 compared with 0 m3 ha–1 on unburned soils. Below a 15 ha catchment with two burned LTEP stands, only a tiny fraction of the estimated 850 m3 of moving sediment (extrapolated from the erosion-box data) appeared in the ditches along logging roads. The export of the 127 Mg ha–1 of missing soil, estimated by the difference between pre- and post-fire soil sampling (Fig. 6) and extrapolated to this catchment, would be about 1900 m3 (assuming a sediment bulk density of 1 Mg m–3). The complex microtopography — partly created by windthrows, downed logs, tree trunks, and needles cast off after the fire — appeared to capture much of the moving soil. The fire also created soil voids — where decayed stumps and roots burned deeply into the soil — that filled over time. This vertical sorting mechanism does not appear to be responsible for the increase in the amount of surface rocks because we observed no corresponding drop in the
rock concentration at deeper soil layers (Fig. 5).
An intriguing alternative explanation for most of the missing fine soil is transport via the massive smoke plume. The elevation of the smoke column and the spread of the plume provide a plausible convective erosion process for off-site transport of substantial material. Large plumes of smoke, some more than 1500 km long, were visible most days during the months of the fire from the NASA MODIS satellite (Fig. 9). Fine soil particles have been detected in smoke (Palmer 1981; Samsonov et al. 2005), and wind speeds near the soil surface — driven by extremely strong vortices resulting from fire-driven atmospheric convection (Palmer 1981; Banta et al. 1992) — can carry smoke to the lower stratosphere (Trentmann et al. 2006). The possibility that a substantial mass of fine particles, including mineral soil, was transported high into the atmosphere raises questions about the effects of intense fire on radiation interception, water-droplet nuclei, and off-site terrestrial and ocean fertilization.
Implications of intense-fire-induced soil changes on climate, forest productivity, and management decisions
Many previous estimates of fire contributions to greenhouse gasses (e.g., Crutzen and Andreae 1990) are based on biomass combustion alone and fail to consider mineral soil losses. Although Campbell et al. (2007) considered soil C losses from the entire Biscuit fire, a concern about the lack of prefire soils data in their estimates is expressed in the range of their C-emission estimates, 0.7 to 1.2 Tg C for the portion of the fire with vegetation damage classes similar to those of our plots. If we extrapolate our results to this area of the Biscuit fire, the resulting soil C loss would be about 1.6 Tg and N loss about 45 Gg. Mineral soil (<4 mm) particulate losses (Fig. 6), extrapolated to the same area, sum to nearly 9 Tg.
Our soil C loss is greater than the high end of the estimates of Campbell et al. (2007); this discrepancy may be related to bias from their unburned controls or to our small sample of the Biscuit fire area. To the extent that our estimates might apply more broadly to other intense fires, climate models may need to be recalibrated to account for effects of intense fire, including fire-induced greenhouse gases and emissions of particulates.
The intensity of wildfires and magnitude of losses of fine soils and soil C and N have additional implications for soil fertility and subsequent rates of plant production and C sequestration. Soil C losses lead to increased bulk density and reduced soil water-holding capacity, cation-exchange capacity, and sources of energy for microbial communities. To the extent that soil N, soil C, and soil structure control productivity, these changes should result in major declines that will last as long as it takes to return to prefire conditions. …
Any potential loss in productivity is relevant to the US National Forest Management Act of 1976, where the Secretary of Agriculture is required, “through research and continuous monitoring, to ensure that management systems will not produce substantial and permanent impairment of the productivity of the land”. The US Endangered Species Act of 1973 is also relevant to the management of high-intensity fires, for example, in the case of the northern spotted owl that nests primarily in stands of large trees averaging only 32 large trees ha–1 (Hershey et al. 1998). When soils can no longer produce such trees, the area of suitable habitat that could redevelop after fire is also lessened.
Much of the recent debate has centered on the effects of postwildfire management on tree regeneration, wildlife habitat, and future fire risk (Donato et al. 2006; Newton et al. 2006; Shatford et al. 2007; Thompson et al. 2007). In light of the first direct evidence of major effects of intense wildfire on soils — based on extensive and detailed pre- and post-fire soil sampling — we think that soil changes, especially the potential loss of soil productivity and greenhousegas additions resulting from intense wildfire, deserve more consideration in this debate. In forests likely to be affected by future intense fire, preemptive reduction of intense-fire risks can be seen as a way to reduce losses of long-term productivity and lower additions of greenhouse gases. Preemptive strategies may include reducing fuels within stands but also improving fire-attack planning and preparation and changing the distribution of fuels across the landscape to reduce the size of future fires. Practices can include thinning and removing or redistributing residues and underburning.
In forests already affected by intense fire, amelioration to increase C sequestration, tree growth, and eventually late successional habitat should be strongly considered. Amelioration practices might include seeding or planting N2-fixing and other plants, fertilizing, and managing vegetation and fuels through time. To the extent that receipts from pre- and post-wildfire logging are the only means of paying for these practices, such logging should be balanced against other management objectives and concerns. Harvesting before and after fire to generate revenue, if done improperly, has the potential to harm soils, but this outcome needs to be weighed against the outcomes resulting from increased high-intensity fire and from not ameliorating after soils have been burned intensely.
