Naturaland anthropogenic fires play a crucial role in the ecology of many terrestrialecosystems, such as boreal forests, temperate forests, grasslands, and thechaparral of Southern California, with significant effects on the carbon cycle,ecological succession, and atmospheric chemistry/aerosols of these ecosystemsand the global climate system.
1 For example, according to NobelPrize winning atmospheric chemist Paul Crutzen, the emission of carbon monoxideand methane from biomass burning can have an effect on the concentration of theOH radical and thus the oxidative capacity of the atmosphere.2 Moreover,aerosols ejected by biomass burning can have radiative forcing effects, and theburning itself can change the albedo of the land, causing a further climateforcing.3 Forest fires in boreal regions may have particularlyimportant implications for climate forcing as they can lead to the thawing ofsubterranean permafrost and the release of methane, an extremely powerful greenhousegas.4 Given the large and diverse consequences that fire can havefor the Earth System, it is important that researchers understand the causesand effects of biomass burning in the context of the variability of the past onboth a global and regional scale. Unfortunately, there is little historicaldata that documents biomass burning prior to the 20th century, soresearchers have instead used proxies, such as sedimentary charcoal records, todetermine global patterns in biomass burning over the past two millennia.
1These charcoal records indicate that over the last 150 years, there has been asignificant reduction in biomass burning, probably due to an increase in humanintervention.1 However, the decrease in moisture associated withclimate change has begun to predispose many forests to wildfires, and researchersproject that this will lead to an increase in biomass burning far beyond thevariability observed over the past 2000 years.1,6 In fact, thissharp rise may have already begun as early as the mid 1980’s, with aparticularly severe increase in the Western United States7.8. Unfortunately,the charcoal proxies traditionally used to quantify historical biomass burningare spatially constrained with some large geographic gaps, for example, incentral Eurasia an Siberia1,5.
Moreover, while the chronologies of somecharcoal records are well resolved, the chronologies of many others remainhighly uncertain.9 Thus, in addition to charcoal sediments,researchers have turned to examining the aerosolized organic combustionproducts of biomass burning which have been deposited and preserved in a moreeasily dated record, ice sheets. Two of the most promising biomass burningtracers are the phenolic acids, vanillic acid (VA) and para-hydrobenzoic acid(p-HBA).
These acids are produced by the pyrolysis of the polymer, lignin, animportant structural component of plants composed of cross linked phenols, andthey offer several advantages over other burning proxies found in ice cores,particularly for the determination of burning in high latitude boreal forests.5,10Namely, their intermediate atmospheric lifetime of several days (when protectedin aerosols) offers information unique from more reactive species such aslevoglucosan, a product of cellulose combustion, and longer lived speciesassociated with biomass burning such as methane, ethane, and carbon monoxide11-15.Moreover, unlike some inorganic compounds used as burning proxies such as blackcarbon, potassium, and nitrate, they do not have other known significantsources. Lastly, the ratio of VA to p-HBA produced depends largely on the typeof vegetation that is burning. This enables researchers to draw conclusionsabout the fuel sources that generate these phenolic acids with a relativelyhigh degree of confidence16,17.
The concentrations of VA and p-HBAfound in ice cores are affected by not only emissions from biomass burning butalso complex transport, deposition, and possibly post-depositional processesthat lend some complications; however, these complications may be overcome withfuture studies. While these species had been examined previously in 2007 and2012, the first high resolution multi-millennial study of these compounds asbiomass burning proxies in ice cores was published in 2017 by Grieman et al. ofUC Irvine.4,5,18VAand p-HBA enter the atmosphere primarily through the combustion of lignin.Lignin is a type of crosslinked phenolic polymer that is an importantconstituent of the cell walls in the vascular tissue of plants. It is composedof three major monomeric building blocks (see figure 1).
They are H-type(p-hydroxyphenyl), G-type (3-methoxy 4-hydroxyphenyl) and S-type (3,5-dimethoxy4-hydroxyphenyl); these monomers are linked by ether (C-O) or C-C bonds.19The exact composition of lignin by monomer varies according to plant species; (figure2 shows a macrostructure of lignin along with the structures of VA, and p-HBA.)thus, the products of the pyrolysis of lignin also change according to whatplant is burning. The lignin in conifer trees, the principal component of highlatitude boreal forests, has a disproportionately high amount of G monomers,making VA a particularly good tracer of conifer burning. The products of ligninpyrolysis are also affected by the temperature of burning.19 Uponcombustion at relatively low temperatures (200oC-400oC),the polymeric linkages are dissolved, liberating the phenolic monomers.
Theside chains of these monomers may then be converted to a carboxyl group in asecondary reaction to form VA from G monomers or p-HBA from H monomers (orsyringic acid from S monomers). Figure 3 shows a mechanism proposed by Kawamotofor the low temperature combustion of G monomers.19 However, athigher temperatures (400oC-600oC) the cleaving of methoxygroups to form catechols, cresols, and coke is favored, and at still highertemperatures(>600oC) the formation of inorganic black carbon andcarbon monoxide is more likely.
5,19 The temperature dependence oflignin pyrolysis products has interesting implications for the utility of VAand p-HBA as biomass burning tracers as different regions often have differenttypes and intensities of wildland fires. There are no other documented sourcesfor these compounds in the atmosphere, and as such, their highest observedatmospheric concentrations are in biomass burning derived aerosols near burningsources (the potential for unknown sources will be addressed later). However,once these compounds have reached the atmosphere, the picture becomes morecomplicated.
VA and p-HBA aresemi volatile organics, so, in the atmosphere, according to Grieman, they canexist in both the gas phase and in aerosols “depending on temperature, aerosolwater content, pH, and cation concentrations.”5 In the gas phase,like most aromatics and other unsaturated hydrocarbons, these molecules arehighly susceptible to oxidation by the OH radical, and will have an atmosphericlifetime of approximately 1 day.5,11 If these molecules only existedin the gas phase then, the potential for long distance transport would beminimal, and, unless the biomass burning was occurring directly adjacent to theice core site (unlikely), their usefulness as a biomass burning tracer would beseverely limited. Fortunately, however, in the atmosphere, VA and p-HBA mostlyexist inside of aerosols, where they are protected from such rapid oxidation bythe OH radical. In 2013, Donahue used models to show that heterogeneous phaseoxidation by the OH radical can take over ten times longer than gas phaseoxidation.11 This means that the atmospheric lifetime of VA andp-HBA is greatly affected by the fraction dissolved in aerosols versus the gasphase, and, when the aerosol phase predominates, the atmospheric lifetime forthese compounds can easily reach five days or more allowing for long rangetransport.
Observations of these molecules in aerosol particles derived fromburning in both coarse and fine modes over the ocean (away from any burning) ata variety of latitudes offer evidence that these compounds do in fact havelonger lifetimes than gas phase oxidation would permit and that long distancetransport is possible.20,21 This intermediate atmospheric lifetimedistinguishes VA and p-HBA from other biomass burning tracers in that itconstrains the possible source location of these combustion products as zonaltransport is likely (on the order of 1000 to 2000 km from the site of the fire),but significant meridional transport is not. Grieman’s study illustrates theutility of this lifetime by using the HYSPLIT model to show that the mostlikely source for VA and p-HBA in the Akademii Nauk ice core is Siberian borealforest (see figure 4).
5 A similar exercise with longer lived biomassburning tracers like ethane or carbon monoxide, both of which have anatmospheric lifetime of about two months, would be futile as the potentialexists for these compounds to be transported much further. In fact, in 2016Nicewonger et al. demonstrated that the levels of these compounds found in highlatitude ice cores is mostly affected by burning in distant tropical latitudes.15On the other end of the spectrum are species such as levoglucosan, a product ofcellulose combustion which has also been used as a biomass burning proxy in icecore analyses. However, laboratory experiments have shown that the lifetime oflevoglucosan can be as little as two days, even when protected in aerosols;moreover, Donahue suggests that because of its chemical composition, a largefraction of levoglucosan resides in the gas phase where it is more susceptibleto OH radical oxidation.
11-13 This limits the potential fortransport, and while levoglucosan has also been observed in long-distancetransported aerosols derived from biomass burning, further studies to reconcilethese conflicts are warranted. Like all atmospheric hydrocarbons, VA and p-HBAare either removed from the atmosphere by wet or dry deposition or ultimatelyoxidized to carbon dioxide through a complex series of reactions with the OHradical.5 Perhapsthe most interesting and one of the most useful facets of using VA and p-HBA asa biomass burning tracers is that the ratio of the two compounds may be used todetermine what type of vegetation was the fuel source. While the emission of otherbiomass burning tracers such as ethane, carbon monoxide, and levoglucosan isaffected only by burning conditions and not vegetation type, the ratio oflignin pyrolysis products is unique to a given type of vegetation because eachtype of plant has a different specific lignin composition, and these ratioshave been studied extensively in laboratory settings. For example, laboratorystudies have shown that the combustion of conifers in North America generatesmore VA than p-HBA while grass fires in similar regions produce exclusivelyp-HBA.16,17 Grieman’s study identified three pre-industrial periodsand one post-industrial period of elevated concentrations of VA and p-HBA inthe Akademii Nauk ice core. The ratio of the two compounds was similar for allthree pre-industrial periods and consistent with conifer forest and woodlandburning, that is there is significantly more VA than p-HBA.
However, in thepost-industrial peak, there is more p-HBA than VA. This is evidence that thesource of this peak is more likely to be tundra grass burning or peat burning.This is an example of one type of information that is missed by usinglevoglucosan or other proxies as biomass burning tracers. Becauseof the relative novelty of using VA and p-HBA in ice cores as biomass burningtracers and the complexity of the many factors that determine theirconcentrations in ice, there is a great deal of room for further research inthe field before this method can be used as a quantitative biomass burningtracer (Grieman suggests that it should be used qualitatively in conjunctionwith other tracers).5 As I mentioned previously, there are no otherknown sources for these phenolic acids in the atmosphere. However, to myknowledge, nobody has examined the possibility that plants cannot emit VA orp-HBA or their precursors nor that these compounds cannot enter the atmospheredirectly from the soil.
Many plants are known to emit aromatic compounds (i.e.the chemicals that give aromatic plants their scent), and in fact, the aldehydeof VA, vanillin, derives its name from the vanilla genus of orchids which emitvanillin in large quantities. So, it is possible that some chemistry may occurin the atmosphere that forms VA or p-HBA from these aromatic emissions. Infact, in a recent study of aromatic Antarctic aerosol constituents, Zangrandoet al found that in aerosols that had been transported long distances and wereassociated with biomass burning, there was a relatively high concentration ofVA; however, in coastal, marine-derived aerosols, the researchers foundsignificant amounts of Vanillin and trace amounts of VA.
This suggests that theocean may generate Vanillin containing aerosols which may then be oxidized toVA in the atmosphere.20 Until these questions of other potentialsources have been examined, we cannot say with certainty that the level of VAand p-HBA found in ice cores is correlated only to biomass burning. Similarly,atmospheric transport processes for VA and p-HBA should be studied furtherbefore they can be used as quantitative biomass burning tracers. While modelshave suggested that long distance transport is possible, and these compoundshave been observed in aerosols away from the site of any biomass burning, it isdifficult to quantify the proportion of these chemicals generated by burningthat will be oxidized by the OH radical or deposited quickly. In fact, the samestudies that observed the presence of these compounds in aerosols away frombiomass burning also noted that they are present in concentrations 1-2 ordersof magnitude higher nearer to fires.
20,21 Therefore, if we wish totake a more quantitative approach to biomass burning, we must determine whatproportion of these chemicals travel a given distance in the atmosphere. One ofthe central questions is what proportion of these acids will be in the aerosolphase (somewhat shielded from the OH radical) versus the gas phase where theacids are easily susceptible to oxidation, and, as Donahue’s study demonstrated,small perturbations can have large effects on gas/aerosol partitioning fororganics and therefore the atmospheric lifetimes of these species.11In 1994, Subramanyam et al. used an annular denuder sampling system followed byconcentration and HPLC-UV analysis to determine the gas/aerosol phasepartitioning for polycyclic aromatic hydrocarbons and phenols, compounds thatare chemically similar to phenolic acids, in a polluted atmosphere inLouisiana.22 It would be interesting and informative to perform asimilar experiment to quantify the partitioning of VA and p-HBA at varyingdistances from a fire site.
In short, open questions remain regarding the speedof degradation and distance of transport of these compounds. Furthermore,in order to relate the concentrations of VA and p-HBA in ice to those in theatmosphere, the deposition of VA and p-HBA on ice sheets should be examined.The importance of depositional processes has been demonstrated by other studiesexamining biomass burning tracers in ice cores. For example, McConnel et al.were able to further constrain the source of black carbon found in Greenlandice cores using models that showed black carbon is primarily deposited by wetprocesses.18 Similarly, Legrand explains differences in ammoniumconcentrations in two Greenland ice cores by citing differences inprecipitation. That is, burning derived ammonium is more likely to be depositedby rain in the summer than snow in winter, so even though net accumulationrates at the two sites are similar, summer precipitation rates are significantlyhigher at one of the ice core cites causing he discrepancy.23 In thesame vein, Fischer et al.
attempted to measure the ammonium in the NGRIP andGRIP ice cores from Greenland to examine North American biomass burning. Theyused a simple method developed from the wet deposition scavenging ratio ofsulfate and from (admittedly sparse) observations of dry deposition velocity tomodel the deposition of ammonium both during transit and at the ice core site.24This enabled the researchers to relate concentrations of ammonium in the icecore to those in the atmosphere both directly above the ice and at the sourceof combustion. This is an important step in moving towards the quantificationof biomass burning, and Grieman suggests that a similar technique may be usedto evaluate VA and p-HBA concentrations in the atmosphere in the future.
Ofthe four major known factors that determine VA and p-HBA levels in ice cores(emissions, transport, depositional processes, and post-deposition processes)we have addressed uncertainties and potential future research directions forall except post-depositional processes including photochemistry, meltwatercontamination, and re-volatilization of the deposited phenolic acids. To myknowledge, these processes as they relate to VA and p-HBA have not beenstudied. However, it has been shown conclusively that other organic acids suchas glycolate disappear from ice cores over time, limiting its use as a tracerto more recent times.23 Similarly, Grannas et al assert thatpost-depositional chemistry in snow is not trivial for organic species andtheir study found that the concentration of benzopyrene decreased by more than90% from the surface to the bottom of a 3 m deep snow pit in Greenland.25Benzopyrene is a polycyclic aromatic hydrocarbon that is also a product oforganic combustion and is fairly chemically similar to VA and p-HBA. Additionally,semi-volatile organic compounds like VA and p-HBA, while not as likely to besubject to re-volatilization as more highly volatile species, could welldissolve in any meltwater, causing changes in the chemical stratigraphy withinthe ice core, presenting obvious problems for the purposes of biomass burningdetermination. These post-depositional processes must be understood as well inorder to determine the quantitative relationship between the levels of VA andp-HBA in ice cores and biomass burning. Inaddition to these four major uncertainties, I would like to address a couplemore minor ones.
The first relates to method validation. Thus far, to myknowledge, Grieman’s study is the only literature that uses VA and p-HBA foundin ice cores as biomass burning tracers (over a period of longer than a fewcenturies). The need for doing so largely rose out of the paucity of charcoalsediments in Central Eurasia and Siberia. Despite Siberia being the largestforested area in the Northern Hemisphere and contributing significantly toburning emissions, the Global Charcoal Database contains only 11 records fromSiberia. While this makes the utility of this new method obvious in thiscontext, it is still important to compare the results of this method with moreestablished methods.
The only Siberian charcoal record that is resolved wellenough to compare with the Akademii Nauk Ice Core is the Bolshoe Bog recordwhich does exhibit similarly elevated levels of these phenolic acids.5Moreover, the 2007 McConnel study of a Greenland ice core showed that blackcarbon and VA levels were well correlated between 1790 and 1850 after whichanthropogenic sources of black carbon became too prevalent to determine biomassburning.18 Interestingly, in 2012, Kawamura found decent agreementbetween VA, p-HBA and levoglucosan peaks in a 300 year ice core taken from theKamchatka peninsula. While these early results are promising, furthervalidation should be undertaken by performing a similar analysis on an ice coretaken from a region where charcoal records are more numerous.
High latituderegions of both North America and South America would nicely meet thisrequirement. Thefinal uncertainty that I wish to address is how changes in burning conditionsmay affect the lignin combustion products. In addition to temperature, humidityand available oxygen may also change how lignin burns. Modeling of the pastmillennia and contemporary observations indicate that most burning in Eurasian borealforests consists of relatively low intensity ground fires.26Meanwhile, in North American conifer forests, fires tend to burn with high temperature,above the ground in the tree foliage. It would be interesting to investigatethe effects these different types of fires could have on which phenolic acidsare present in ice cores and their relative abundance. Currently, I am analyzingthe levels of these phenolic acids in an ice core taken from the EclipseIcefield in the Alaska Range, and I am hopeful that this work may answer someof the questions outlined above.
Inconclusion, VA and p-HBA are promising biomass burning tracers found in ice cores,and they may prove particularly useful for analyzing burning in high latitudeboreal forests. Their atmospheric lifetime and lack of other documented sourcesenables researchers to derive information unique from the information gleanedfrom other biomass burning tracers. However, there is ample room to improve thescientific understanding of several processes that affect the concentrations ofthese phenolic acids in ice cores, namely the potential for other emissionsources, complex atmospheric transport processes including aerosol/gas phase partitioning,depositional processes, and post-depositional processes. Currently, research isunder way to examine some of these questions and to improve the understandingof the quantitative relationship between biomass burning and the levels of VA andp-HBA found in ice cores.