Charcoal and pyrogenic carbon (PyC) are general terms given to organic materials that have been subject to partial combustion during biomass burning events (Forbes et al., 2006; Zimmerman and Mitra, 2017). There is increasing interest in the role that PyC plays in long term carbon (C) storage in terrestrial and aquatic ecosystems, yet there is only limited understanding of the rate at which PyC accumulates in forest ecosystems during natural and prescribed (Rx) fire events. Of the PyC formed during forest fire events, much remains onsite as stored C (Lynch et al., 2004), eventually lost to depositional areas due via surface erosion (Abney et al., 2019) and if incorporated into mineral surface soils, the PyC may remain in surface soils an indefinite length of time (Abiven and Santín, 2019). Given this longevity in soil ecosystems, PyC has been identified as being an important C storage medium (Lehmann et al., 2006; DeLuca and Aplet, 2008; Santín et al., 2015) and receives a great deal of attention across a variety of research fields (Carcaillet, 2001; Gavin, 2004; Gavin et al., 2006; Power et al., 2006).

The role of PyC in the soil ecosystem extends beyond its capacity to serve as stored C, PyC can directly and indirectly influence soil microbial processes (Thies et al., 2015), increase rates of organic matter decomposition (Wardle et al., 2008; Pluchon et al., 2015) alter soil nutrient availability (Lehmann et al., 2011; Gul and Whalen, 2016; Gao and DeLuca, 2019; Gao et al., 2019), modify soil nutrient cycling (DeLuca and Sala, 2006; DeLuca et al., 2006, 2015; Gao et al., 2017; Gao and DeLuca, 2018), and adsorb organic compounds in plant root exudates, litter decomposition products, and microbial byproducts (Keech et al., 2005; Gundale and DeLuca, 2006; Quilliam et al., 2013). These facets of PyC have led to a significant increase in the number of studies devoted to the role of naturally formed PyC in terrestrial ecosystems (Pingree and DeLuca, 2017).

Despite the fact that PyC has widely been recognized as a highly chemically stable form of ecosystem C that is formed in all biomass burning events and is of great importance to global C cycle, to date there has been limited effort to quantify forest fire generated PyC and in particular that affiliated with Rx fire events in forest ecosystems (Alexis et al., 2012; Pingree et al., 2012; Wiechmann et al., 2015). In addition to wildfire events that commonly occur as a result of local climatic conditions at the seasonally dry forest ecosystems of the Rocky Mountain West (Agee, 1996), Rx fires are applied widely as a forest management practice throughout the region to reduce surface fuel loading and restore the historical fire regime (Ryan et al., 2013). Quantifying the rate of PyC production during fire events in this region is essential to estimating the potential of ecosystem C storage in other forest systems with similar climatic conditions and fire management practices. The objectives of the work reported here were to: (1) Evaluate the PyC generation rate and the percent fuel conversion to PyC during wildfire or Rx fire events of forest ecosystems through synthesizing existing literature; (2) Determine the amount of PyC present in soil following forest restoration treatments that include Rx fire in an empirical study in western Montana, and evaluate the relationship between PyC generation and fuel consumption using data collected on the same study sites. By combining a synthesis and an empirical study that both focus on the evaluation of PyC generation in forest fire events, our goal was to advance the understanding of total ecosystem C budgets and the C storage potential of post-fire forest systems. We believe the findings from this research could help inform more rational strategies toward assessment of the influence of fire and fuel treatments on ecosystem C storage.

In this study, we combined an assessment of previously published data on PyC production in forest wildfires or prescribed (Rx) fires with an empirical study from a Rx fire study in western Montana. The data synthesis from existing literature provides a range for biomass conversion to PyC during Rx fire and wildfire events with Rx fire events producing similar amounts of PyC, but a far lower degree of variation than wildfire events (Figure 1). This is somewhat expected given the relative low severity of most Rx fires. Previous studies have estimated PyC generation per unit fuel consumption in coniferous forest ecosystems to range from <1% to nearly 20% (Tinker and Knight, 2000; Lynch et al., 2004; Forbes et al., 2006) and more commonly reported as being between 1 and 5% (Preston and Schmidt, 2006). In a more recent detailed analysis of PyC generation in a boreal forest wildfire with pre and post-fire sampling canopy, understory, and surface soils, Santín et al. (2015) reported an average ecosystem conversion rate of 27% of forest biomass with a conversion rate of 50% for downed woody debris. These estimates are far higher than the average value of 5.2% of fuel consumption observed in our data synthesis, but may reflect the fact that most studies have not had the benefit of pre-and post-fire sampling (Miesel et al., 2018). Further, visual estimates can overestimate PyC generation rates compared to wet chemical digestion methods.

Importantly, there is little consistency between individual studies and the method employed to estimate PyC content generated during fire events, therefore the PyC production rate across studies was highly variable which creates challenges, particularly when modeling at larger spatial scale (Wei et al., 2018). By compiling published data of soil PyC content in the whole soil profile to estimate global PyC stock across a variety of environmental constraints (e.g., climate, soil pH, site net primary production), Reisser et al. (2016) found that soil properties, rather than fire characteristics best explained PyC content at global scales where high clay soils (clay content over 50%) and alkaline soils had more PyC content than low clay soils and acidic soils. The relatively small sample size of this data synthesis limits our ability to extrapolate the results; however, we report an average value of 2.2 Mg PyC ha⁻¹ for a single fire event. From these data it was hard for us to further generalize correlations between environmental variables and PyC production per single fire event. It is possible that the differences in soil and site properties across geographically distinct sites contributed to the variation in PyC persistence; however, it is more likely that fire and fuel characteristics including burn temperature, fuel composition and loading, and fuel moisture were the strongest drivers of variation in PyC production across different studies (Huang et al., 2018). This is supported by our data synthesis which showed that the higher amount of PyC production was observed under “wildfire” compared with “Rx fire” (Figure 1A), and under “thinning + Rx fire” when compared with “Rx fire only” (Figure S1). The average value of 2.2 ± 1.7 Mg C ha⁻¹ generated per fire event in the form of PyC is a close fit with the Rx fire data in our empirical study and suggests that PyC generation during Rx events is a significant contributor to the stable C pool in western forest ecosystems.

In our empirical study, the Rx fire treatments generated PyC in the O horizon at a rate of about 12% of fuel consumed in the thin plus burn treatments which is slightly higher than the average of our synthesis findings and midway between the estimate of Preston and Schmidt (2006) and that of Santín et al. (2015). Alexis et al. (2007) reported relatively lower PyC generation rates (3.2% of stem, leaf and litter converted to PyC post fire) in a sub-tropical scrub oak ecosystem and indicated that PyC accumulates at lower rates than conservative estimates of stable organic matter formation (7% of stem, leaf and litter). Insights from that ecosystem led the authors to suggest that fire and PyC generation lead to a long-term decline in soil C. However, this argument does not account for differences in the longevity of stable organic matter and PyC, which may be over 10 times greater for PyC compared to soil organic matter (DeLuca and Aplet, 2008). Further evidence for longevity of PyC in soils and an indication of the limited impact of a single burn event on soil PyC stocks is captured in the recent study by Santos et al. (2017) which investigated PyC stocks in litter layers (significantly different from a 100-years post-burn site to an unburned site) and dissolved PyC leaching from soil horizons (not significantly different from burned to unburned). At our FFS sites, the lack of differences in mineral soil PyC content across treatments is likely a product of historical fires, including a fire in the 1930's that covered portions of Block 1 and 3 (Gundale et al., 2005), whereas recent fire activity resulted in measurable changes in the O horizon PyC content.

Recent studies have pointed to PyC generation as offsetting C emissions during wildfire (Wei et al., 2018; Jones et al., 2019) requiring us to further assess the influence of fire on the global C balance. Importantly, these studies demonstrate that PyC accumulation in ecosystems is most directly related to fire severity and frequency (Sawyer et al., 2018; Adkins et al., 2019). So, although PyC generation in individual fire events in Savanah ecosystems is low, the net accumulation can be high. The same thinking must go into how we approach Rx fire in western forest ecosystems of the Inland Northwest, where multiple Rx fires over a 100 years period will store more PyC than a single high severity fire over the same time period (DeLuca and Aplet, 2008).

This study provides a combination of original empirical data on PyC generation as well as a data synthesis from prior studies that address PyC generation in forest wildfire and Rx fire events. The results presented here support previous studies showing that fires generate a significant quantity of PyC, a relatively passive form of organic C. Our empirical results further demonstrate that the majority of PyC mass resides in the mineral soil, even though PyC storage from recent fires resides in the O horizon (Brimmer, 2006; DeLuca and Sala, 2006; Kurth et al., 2006). Charcoal left on site likely disintegrates into smaller particles over time further affording incorporation into the mineral soil as a result of physical and biological pedoturbation (DeLuca and Aplet, 2008). This fine particulate C may also leach into deeper soil horizons over time (Santos et al., 2017). The reintroduction of fire using these restoration treatments in the region represents the first potential addition of PyC to these ecosystems over a 100-years period. Although the primary reasons for implementing restoration or fuel reduction treatments are for reducing wildfire risk and restoring historical ecosystem structure and function, there is clearly an additional benefit in that Rx fire generates a surprisingly large pool of highly stable C, particularly when it is applied in combination with forest thinning. And while the immediate effect of Rx fire is to release C, in the long term this approach may promote C sequestration. Future efforts to restore forest ecosystems in the Inland Northwest should incorporate Rx fire to further reduce surface fuel loading, reestablish more natural fire regimes, and support long-term ecosystem C storage.

The datasets generated for this study are available on request to the corresponding author.

TD, MG, and RB were involved in designing the empiric study, conducting field experiments, and laboratory analyses. TD and SG conducted data synthesis and analysis. TD, SG, and MG drafted the manuscript.