Appendix E - Persistence of biochar in soils

It is well documented that carbonaceous residue from partial combustion or pyrolysis of biomass, termed biochar or Black Carbon (BC), can persist in soils for thousands of years. However, there is also strong evidence that charred organic material in soils can decay over time. Assessing the contribution that BC can make to carbon sequestration in Australia requires that BC decay rates be quantified under local field conditions.

The present review of the published literature was undertaken to gather and analyse information on the level of confidence that may be placed on claims made by various pyrolysis technology proponents of long persistence of BC in soils. The task force is aware that long-term measurements of BC persistence in soils for a range of well-characterised chars are underway in Australia.213,214 That data, when available, will provide a marked improvement in the confidence with which biochar technologies may be assessed in terms of permanence of carbon in soils.

A convenient simplification is to assume that the BC comprises one or more fractions that each decay at a constant rate215. The BC residual after time t years can then be described by eq. E.1:

Where Cremaining : fraction of original BC after t years
Ci : initial fraction of BC component i
ki : exponential decay constant for the ith fraction (per year)
t : decay period (years)
i : number of BC components with discrete decay rates

This allows the definition of alternative characteristic times for each fraction such as

  • Mean residence time: average time that the fraction exists (= 1/ki).
  • Half life: elapsed time to 50% reduction of fraction (Cremaining = 0.5).

Decay constants of BC have been assessed by a number of authors. Nguyen et al. (2008)216 measured BC carbon in 18 samples of agricultural soils from Kenya at a location with a mean annual temperature of approximately 19°C. The farms had been established on land cleared by slash and burn of original forests over periods of 2 to 100 years prior to sampling. BC in soil was correlated as a function of time by the relation BC = 3.51 + 9.16e−0.12.years. This indicates two fractions, one with negligible decay over the experimental period and one with an exponential decay rate of 0.12 per annum.

Nguyen and Lehmann (2009)217 measured decay rates for BC produced by slow heating to 350 and 600°C of corn stover and oak wood chips under nitrogen. Sized char particles (+ 0.5mm, −2 mm) were then incubated in sand doped with microbial inoculant and nutrient solution for one year under saturated, unsaturated and alternating moisture conditions. Carbon loss varied between 21.2% for 350°C corn stover BC under unsaturated conditions to 6.15 % for 350°C oak char under saturated conditions. In general the low temperature chars exhibited greater carbon loss than did the high temperature char, particularly in the case of corn stover char.

Nguyen et al. (2010)218 further investigated the effect of temperature on decomposition rates of BC produced from corn stover and oak chips as described by Nguyen and Lehmann (2009). The BC types were incubated for a period of one year in inoculated sand cultures, at six different temperatures and using eight replicates.

All BC types were found to mineralise to CO2 more rapidly as temperature increased, with the low temperature chars mineralising more rapidly than high temperature chars and corn stover char more rapidly than oak chip derived materials at any temperature. The impact of incubation temperature change was found to decrease as temperature increased.

Data is provided on residual carbon at one year for each char type and at six temperatures between 4 and 60°C in an electronic supplement to the paper. Residual carbon is correlated against temperature using eq. F.2 and the parameter values for each char type are tabulated.

f = Const + a.e−bT (E.2)
Where f : fraction residual carbon at end of experiment
T : temperature (°C)
a : pre exponential factor
b : exponential temperature weighting factor

Residual carbon at a temperature of 18°C is here calculated for each char using eq. E.2, and the equivalent exponential decay rate and half-life calculated. Results are listed in Table E.1.

Table E.1

Char Source Char temperature (°C) Charresid 18°C (%) Exponential decay constant (per annum) Half life (years)
Corn stover 350 83.8 −0.177 4
Corn stover 600 85.8 −0.153 5
Oak Chips 350 90.7 −0.098 7
Oak Chips 600 92.8 −0.075 9

Kuzyakov et al. (2009)219 have measured carbon loss from soil and loess samples doped with 14C labelled BC. The BC was produced by slow heating of perennial ryegrass followed by heat soaking (13 hours) at 400ºC. The BC was incubated for 1181 days at 20°C and saturation to 70% water holding capacity. CO2 emitted by the samples was measured at intervals by absorption into NaOH solution and microbial biomass determined at the end of the experiment. Samples were subject to agitation and to dosing with glucose solution (to stimulate microbiological activity) on a number of occasions.

It was found that BC decomposition rates decreased from an initial rate of 0.016%/day and 0.024%/day (soil and loess respectively) to a constant rate over the final two years of 0.0013 to 0.0015 %/day (both substrates). This decrease in rate with time was taken as evidence of depletion of more reactive components in the BC. Agitation of samples had minimal effect but stimulation of biological activity in the samples caused a short term increase (< one month) in decomposition by about a factor of 3 in the soil and somewhat higher in the loess. This was taken as confirming that biological activity is important in the decomposition of BC. It was concluded that under field conditions of 7°C mean temperature and mean precipitation of 600 – 700 mm, decomposition rate would decrease in proportion to decrease in biological activity to about 10% of that measured.

Chung et al. (2008)220 have measured decomposition of BC residue from 16 historical sites. The BC was originally produced by charring of woods such as oak, chestnut and hickory to produce a high strength, high density charcoal suitable for blast furnace use. Samples of BC containing soils were obtained from the historical storage sites plus from non BC containing soils adjacent the storage sites as reference material.

BC containing and non BC containing soils were incubated at 30°C and 60% water holding capacity for a total of 177 days. Separately, selected BC particles (+1mm −2mm) immersed in inoculated sand with added nutrient were incubated for 50 days. Evolved CO2 was collected in NaOH solution and analysed at intervals during the incubation period. Data from the soils was fitted by a double exponential decay model (a labile and a recalcitrant component) for the soils and a single exponential for the BC particles.

Mineralisation rates of the recalcitrant component of the soils was determined to range between 0.15 × 10−4 and 5.02 × 10−4 per day (mean of 0.43 × 10−4/day). While the quantity of BC in soils (assumed to be the recalcitrant material) was not determined directly, the mean difference in organic carbon between BC containing soils and their respective reference samples allowed a half-life of 80 years to be determined for the BC component. The data also indicated that an increase in temperature by 10°C would produce a 3.38 times increase in mineralisation rate and a half life at 10°C mean average temperature of 925 years is proposed. Cheng et al. note that the more labile fractions of the BC will likely already have been lost over the average 130 year life of the samples.

Decomposition rate constants determined for the BC particles were much higher than for the recalcitrant component of the soils. Measured decomposition was assumed due entirely to contamination of the particles with labile organic carbon. It is noted that, if the alternative assumption, that the measured mineralisation is from the BC particles, in made then decay rates and half-life of the BC particles can be calculated from the data provided. The half-life is here estimated to range between 23 and 206 years for these particles. Accepting labile carbon contamination in at least some of these particles then these data provide minimum values for half-life of the BC particles and varying contamination may account for the large spread in the data.

Bird et al. (1999)221 measured oxidative resistant elemental carbon (OREC) in soil plots in Zimbabwe that had been exposed to fire at regular periods of 1, 3 and 5 years respectively, or alternatively protected from fire since 1947. OREC in samples was determined by chemical oxidation of the labile proportion of carbon in the soil followed by combustion to determine residual (OREC) quantities.

Total OREC in the protected plots averaged 2.0 ± 0.5 mg/cm2 as compared to that of the burnt plots of 3.8 ± 0.6 mg/cm2, with only minor differences (± 10%) between different burning treatments. This data suggested a half-life for natural degradation of OREC of < 100 years. Further, larger char particles (>2mm) were common at the burnt sites but largely absent in the protected sites. This indicated a half-life of < 50 years for these particles which appear to progressively move to smaller size fractions. It was concluded that only a small fraction of OREC is likely to be sequestered into a slow cycling geological carbon reservoir at these well aerated savannah sites.

Baldock and Smernik (2002)222 investigated the chemical changes and resistance to decay of charred samples of red pine sapwood. Samples were heated in air to constant weight at temperatures of 150 to 350ºC. The resultant chars were then analysed for chemical structure and bioavailability of carbon. Bioavailability was assessed by incubating the samples in a sand medium in air at 25°C. Water, microbial inoculum and nutrients were added to the sand with the objective of determining the bioavailability of carbon in the heated samples relative to that of the original material and to glucose and cellulose. Residual carbon was determined after 120 days by combustion and the fraction of biological available material determined as the fraction of initial organic carbon respired.

Approximately 85%, 74% and 20 % of the organic carbon in glucose, cellulose and original sapwood respectively were mineralised to CO2 over the period of incubation. This reduced to 13% for sapwood heated to 150°C and to <2% for sapwood heated to temperatures ≥200ºC respectively. Samples heated to ≥200°C all showed some loss of material but differences were not statistically different. This reduction in bioavailability coincided with significant changes in chemical composition of the sapwood observed at temperatures ≥200°C.

Harmer et al. (2004)223 measured decomposition rates of BC material produced by heating charred maize and rye straw to 350°C in closed stainless steel containers for two hours. Oak wood was charred at 800 oC for 20 to 24 hours. The BC samples were mixed with sand, microbial inoculum, nutrient solution and moisture sufficient to reach around 60% of water holding capacity and then incubated for 60 days at 20°C. Samples were primed by the additional of 14C labelled glucose at day 1 and day 25 and BC mineralised determined from measurement of CO2 released with adjustment of that derived from the labelled glucose.

Mean residence times of 39 and 76 years respectively were calculated for charred straw and charred wood respectively using a two component first order decay equation. Due to the short time of the experiment it may be expected that this will be a minimum value due to the likely early loss of easily degradable fractions. However, while high initial decay rates were observed, decay rates had stabilised by the end of the experiment. The addition of glucose was seen to accelerate decomposition of the char indicating a contribution to decay by microbial action.

Gavin et al. (2003)224 collected multiple core samples for 75 sites in forest in Vancouver Island and analysed these for BC that had been produced by past wildfire activity. Care was taken to identify sites where BC was unlikely to be affected by lateral transport mechanisms. BC content of samples was determined by manual isolation under a binocular microscope and weighing of separated BC particles. Ages of the BC samples were determined by radiocarbon dating, correlated against tree ring data where available. It was found that content of BC in soil varied greatly between sites and decreased only slightly with time, content of >0.5 mm particles in soil was expressed in first order exponential form with an exponential decay rate of −0.000151 %/year (R2 = 0.09).

Preston and Schmidt (2006)225 interpret this data as indicating an average half-life of 6623 years for the BC. However half-life, defined as time to 50% residual carbon, is 4590 years using the correlation provided in Gavin et al. (2003) while mean residence time, defined as the inverse of the decay constant is 6623 years. The low coefficient of determination (R2) for the original data also suggest high uncertainty in this conclusion.

Wardle et al. (2008)226 measured decay rates for laboratory produced char, forest litter and a 50:50 mix of char and litter, buried in mesh bags at three sites in a boreal forest. Average temperature for the sites are not given but are expected to be low. Weight loss measurements were carried out at intervals over 10 years. The char exhibited a small weight loss over the first four years but then showed a small weight gain for the final (10 year) measurement. The cumulative carbon release over 10 years for the char is indicated as of the order of 5 mg/g carbon with an uncertainty of ±5 mg/g.

It was noted that the char accelerated decomposition of the litter in the char/litter mix, suggesting that char additional to soil may result in increased mineralisation of labile carbon in soils. The char containing samples were both found to have accumulated nitrogen at 10 years.

Lehmann et al. (2008)227 used measured data on BC and soil organic carbon in two savannah regions in northern Australia as a basis for estimating the decay rate of BC produced by wild fire at these locations. Estimates of fire frequency, proportion of biomass converted to BC and decay rates for soil organic carbon allowed determination of BC decay rates that allowed predicted BC contents to match the measured values. Mean residence times of from 718 to 9259 years were determined, depending on assumptions made about fire frequency and proportion biomass conversion to BC.

Hammes et al. (2008)228 measured BC in samples of Russian steppe soil samples taken approximately 100 years apart. The steppes were previously frequently exposed to fire, however the initial sample was taken near to the time that regular fire was controlled by agricultural activity. The more recent sample was taken from the location identified in records as that of the initial sample while supporting samples were taken in the near vicinity. The topography and stability of soil type over large areas in this location is expected to reduce impact of potential differences in location of original and more recent samples.

The rate of decay of BC was modelled using the concept of turnover time as defined in eq. E.3, τ reduces to mean residence time ( = 1/exponential decay rate) when modern BC contribution is zero.

τ = −T/In(FB/(F1)) (F.3)
Where τ : turnover time (years)
T : time elapsed between samples (years)
F : ratio of modern BC contribution to historical rate
B : fraction of original BC stock remaining after T years

Turnover time was calculated to range between 212 to 541 years depending on choice of times between sampling and soil bulk density. The best estimate, assuming a 90% reduction in BC contribution rate over the experimental period, was 293 years.

Liang et al. (2008)229 measured decay rates from BC in anthrosols collected from four sites in Brazil and from four adjacent non BC containing sites for comparison. Samples were incubated for 532 days at 30°C with moisture maintained at 55% of field capacity. Decay of carbon was determined by measurement of CO2 evolved during the experiment. A double exponential decay model was fitted to the data, assuming a large slow turnover pool (BC) and a smaller more labile component.

Turnover times ranging between 9 to 20 years were calculated for the control samples and between 44 to 52 years for the BC containing anthrosols. However, it was concluded the double curve fit approach overestimated the proportion of stable carbon in the soils. It was however concluded that the turnover time for the BC would be centuries to millennia based on a more than order of magnitude difference in decay rates of the labile and slow turnover components. In view that the double decay model was considered not to properly represent the decay of the different carbon fractions this conclusion must be considered tentative.

Biochar clearly decays with time. To recognise biochar as a carbon sink then requires that a minimum storage time period be ascribed, carbon lost due to decay prior to this time cannot be considered as sequestered. A precedent for minimum storage time is contained in the Australian carbon farming initiative draft guidelines230 which indicate a requirement for a minimum 100 years storage period.

A summary of the literature data on BC decay rate is given in Table E.2. Different authors have provided decay rate data in various forms. Here it is converted to a constant base of percent carbon remaining after 100 years as determined when applying the decay rates assessed by the various studies. Where incubations were carried out at temperatures above 25ºC, decay rates were corrected to 18°C231 using an average of correlations for different char types from Nguyen et al. (2004).

Care must be exercised in extrapolating short term incubations studies to life of BC materials under field conditions. In particular many incubations use high temperatures and constant moisture to accelerate decay rates, many use sand as a medium which maximises aeration but minimises the potential for soil components such as clays to bind with and protect BC particles from decay. The short term of these experiments will also mean that results will be biased toward properties of the more degradable fractions. Conversely, examination of aged soil and BC samples will be biased toward fractions that have high resistance to decay.

Table E.2 Literature data on BC decay

Reference Char type Cremaining, 100y
Nguyen et al. (2008) Char produced by slash and burn clearing of land 28%
Nguyen and Lehmann (2010) Corn stover and oak chips, charred at 350 and 600ºC, varying moisture of incubation <1%a
Nguyen et al. (2010) Corn stover and oak chips, charred at 350 and 600ºC, varying temperature of incubation 0a
Kuzyalov et al. (2009) Perennial rye grass char in soil and loess 60%b
Cheng et al. (2008) Old charcoal from blast furnace operations 56%c
Bird et al. (1999) Savannah combustion 29%
Baldock and Smirnik (2002) Pinus sapwood charred at 150 C 0%
Baldock and Smirnik (2002) Pinus sapwood charred >200ºC >0.2%
Harmer et al. (2004) Maise and Rye residue 350ºC 17%
Harmer et al. (2004) Oak wood residue ºC 40%
Preston and Schmidt (2006) Aged charcoal from forest fires, data from Galvin et al.(2003) 98.5%
Wardle et al. (2008) Lab charcoald 95%
Lehmann et al. (2008) BC from burning of savannah 87 – 99%
Hammes et al. (2008) BC from burning of Russian steppe 71%

a decay rate measured at 30°C adjusted to 18°C using data from Nguyen et al. (2010)

b estimate based on constant decay rates over final two years of experiment

c estimate based on half-life inferred by authors at 30°C, adjusted to 18°C using data from Nguyen et al. (2010)

d Temperature not given, expected to be low in boreal forest.

It must be further recognised that BC properties can depend strongly on the source biomass, it’s preparation including activation, grinding and previous exposure to weathering, and the pyrolysis conditions. These latter conditions could include, for example, the presence during heating of oxygen, moisture and other contaminants; heating rate; final temperature and the period of soaking at the maximum temperature. Few of the publications discussed above provide such details and it is likely that none of these chars is representative of that produced from commercial pyrolysis equipment.

The data reviewed above shows little consistency, although it is clear that some chars do indeed persist for periods well in excess of 100 years. There appears to be a trend with high temperature chars exhibiting greater decay resistance that low temperature chars, and chars sourced from wood appear to be more persistent than BC from more friable sources.

213 Singh B.P., Cowie A.L., Smernik R.J. (2011) “A novel 13C natural abundance approach for measuring biochar’s stability and priming effect on ‘native’ soil carbon. 11th Australasian Environmental Isotope and 4th Australasian Hydrogeological Research Conference, Cairns, July 12–14.

214 Singh B., Singh B.P., Cowie A.L. (2010) “Characterisation and evaluation of biochars and their application as a soil amendment”. Australian Journal of Soil Research, 48(7), 516–525.

215 Lehmann J., Czimczik C., Laird D. and Sohi S. (2009) “Stability of Biochar in the Soil” in Biochar for Environmental Management, Lehmann J and Joseph S. (eds) Earthscan London

216 Nguyen B.T., Lehmann J., Kinyangi J., Smernik R., Riha S.J. and Engelhard M.H. (2008) “Long term black carbon dynamics in cultivated soil” Biogeochemistry 89: pp 295–308

217 Nguyen B.T. and Lehmann J. (2009) “Black carbon decomposition under varying water regimes” Organic Geochemistry 40: pp846–853.

218 Nguyen B.T., Lehmann J., Hockday W.C., Joseph S., and Masiello C. (2010) Temperature Sensitivity of Black Carbon Decomposition and Oxidation” Environ. Sci. Technol. 44; pp 3324–3331.

219 Kuzyakov Y., Subbotina I, Chen H, Bogomolova I and Xu X (2009) “Black carbon decomposition and incorporation into soil microbial biomass estimated by 14C labelling”. Soil Biology & Biochemistry 41:pp 210–219

220 Cheng C.H., Lehmann J., Thies J.E., Burton S.D. (2008) “Stability of black carbon in soils across a climatic gradient” Journal of Geophysical Research V113; G02027.

221 Bird M.I., Moyo C., Veenendaal E.M., Lloyd J. and Frost P. (1999) “Stability of elemental carbon in a savanna soil” Global biogeochemical cycles 13(4); pp923-932

222 Baldock J.A. and Smernik R.J. (2002) “Chemical composition and bioavailability of thermally altered Pinus resinosa (Red pine) wood” Organic Geochemistry 33; pp 1093–1109.

223 Hamer U., Marschner B., Brodowski S.and Amelung W. (2004) “Interactive priming of black carbon and glucose mineralisation” Organic Geochemistry 35; pp 823–830

224 Gavin D.G., Brubaker L.B. and Lertzman K.P. (2003) “Holocene fire history of a coastal temperate rain forest based on soil carbon radiocarbon dates” Ecology, 84(1) pp. 186–201.

225 Preston C.M. and Schmidt M.W. I. (2006) “Black (pyrogenic) carbon: a synthesis of current knowledge and uncertainties with special consideration of boreal regions” Biogeosciences, 3; pp 397–420.

226 Wardle D.A., Nilsson M. and Zackrisson O. (2008) “Fire-Derived Charcoal Causes Loss of Forest Humus” Science 320(2) pp629.

227 Lehmann J., SkjemstaD J., Sohi S., Carter J., Barsons M., Falloon P., Coleman K., Woodbury P. and Krull E. (2008) “Australian climate–carbon cycle feedback reduced by soil black carbon” Nature Geoscience Letters 1; pp832-835.

228 Hammes K., Torn M.S., Lapenas A.G. and Schmidt M.W.I. (2008) “Centennial black carbon turnover observed in a Russian steppe soil” Biogeosciences Discussion 5; pp 661–683.

229 Liang B., Lehmann L., Solomon D, Sohi S., Thies T.E., Skjemstad J.O., Luiza F.J., Engelhard M.H., Neves E.G., Wirick S. (2008) “Stability of biomass-derived black carbon in soils” Geochimica et Cosmochimica Acta 72; 6069–6078

230 Dept of Climate Change and Energy Efficiency “Carbon farming initiative - Draft Guidelines for Submitting Methodologies. Commonwealth of Australia ( – accessed 12/7/2011))

231 Based on mean of long term annual average maximum and minimum temperatures at 13 locations across NSW, Vic, SA and WA. Farming regions. Data from Australian bureau of meteorology ( assessed 12/7/2011)