Comment on: 'burning management and carbon sequestration of upland heather moorland in the UK'.

A recent paper by Farage et al. (2009) on burning management and carbon sequestration of an upland heather (Calluna vulgaris) moorland addresses an important subject that is currently under much debate in the UK. Many of the upland moorland systems in the UK have organic soils and store considerable quantities of carbon. Estimates of the total carbon stored in these soils vary depending on the assumptions made for soil depth, geographical area, bulk density and carbon content (see Chapman et al. 2001) but a recent reappraisal suggests they store in the region of 3000 Mt carbon (SEERAD 2007). Large areas of these systems are managed by burning, either for red grouse habitat or for sheep grazing (Yallop et al. 2006). The use of fire for managing this land is highly controversial (Tucker 2003; Pearce 2006; Davies et al. 2008a) and the debate has become polarised; the effect of burning on carbon budgets is an important component of this debate (Davies et al. 2008a). Research that estimates the impact of burning on carbon sequestration is urgently needed and the paper by Farage et al. is already being used as evidence that, to quote their abstract, 'careful burning management at this site does not have a major detrimental effect on the carbon budget'. Although Farage et al. are careful to point out that their results apply only to the one particular site, Mossdale Moor, where measurements were made, their conclusions are inevitably being used by landowners and moorland managers elsewhere in the country as evidence to defend the practice of management by burning and ultimately may be used as evidence in the policy decision-making processes (e.g. Scottish Government 2009). It is therefore crucial that a paper on such an important topic should be placed in an appropriate context.

Unfortunately, their estimates of the amount of carbon in vegetation that is consumed in a fire appear to be extreme and thus atypical of this type of management fire. For example, they report a very high amount of biomass remaining after the fires (given as 1262 g/[m.sup.2] in the text but as 1149 g/[m.sup.2] in their table 1), which is remarkable for a management fire in 15-year-old heather. This contrasts with our own data from 26 experimental fires, typical of those used in management, in similar vegetation in which we recorded a mean of only 277 g/[m.sup.2] of heather remaining with a maximum of 940 g/[m.sup.2] in one late-mature-phase stand (Table 1). This late-mature-phase stand was itself atypical in that the heather stems had collapsed such that a high proportion of the dead and fine fuel was close to the ground, un-aerated, and remained too damp to burn. The only situation in which more fuel than this could remain after a fire is burning in very old heather (late mature and degenerate phases; Gimingham 1988) in which the coarse woody stems comprise a significant proportion of the total biomass but remain unburned and in which a deep layer of mosses and litter has accumulated beneath the heather canopy which is too wet to burn. However, Farage et al. describe their burnt stands as '~15 years old' (peak to late building phase) and describe the harvested material as 'almost exclusively heather', which presumably did not include significant quantities of either very coarse stems or moss and litter.

Farage et al. calculate that the carbon released from aboveground biomass in two fires was 103 and 201 g/[m.sup.2] (16% and 24% of the above-ground biomass), respectively. This again contrasts with our own data in which the losses averaged 79% of above-ground biomass (Table 1). Using Farage et al.'s estimate for the carbon content of biomass as 48.1%, this gives a mean of 448 g C/[m.sup.2] (932 g/[m.sup.2] organic matter) consumed; the minimum fuel consumption from the 26 fires was 240 g C/[m.sup.2] (499 g/[m.sup.2] organic matter). In addition to our own measurements, Kayll (1966) reported a mean value of 1481 g/[m.sup.2] for fuel consumption (1032 g/[m.sup.2] of heather; the total includes consumption of what must have been an unusually dry moss/litter layer) in two fires burnt in 15-year-old heather in the autumn, and 650 g/[m.sup.2] of fuel consumed (all heather) for four fires burnt in 25-year-old heather in the spring. Hobbs and Gimingham (1984) reported similar values, with consumption ranging from 868 g/[m.sup.2] in a pioneer stand to >2000 g/[m.sup.2] in mature heather (mean 1404 [+ or -] 81 g/[m.sup.2], corresponding to 675 g C/[m.sup.2]). Indeed, from our experience, Farage et al.'s estimate of 103 g/[m.sup.2] of carbon (equivalent to 214 g/[m.sup.2] of organic matter) seems insufficient fuel to carry a fire.

Farage et al. appear to have estimated carbon losses in the fire by comparing the material remaining after a fire with that present in a neighbouring unburnt stand of vegetation. No evidence is presented to show that the pre-fire biomass was measured in the burnt stands, nor indeed that the stands had similar biomass before the fires other than that they were both '~15 years old'. The simplest explanation for the very low estimates of fuel consumed is, therefore, that the pre-fire biomass in the burnt stands was actually significantly higher than that in the neighbouring unburnt stands and the calculated fuel consumed is therefore underestimated. The description of their experimental design suggests that they only collected samples from a single burnt and unburnt stand in each of the two years. If true, this would be pseudoreplication (sensu Hurlbert 1984), which invalidates the use of any statistics to estimate confidence intervals on the estimates of carbon lost due to burning. If this is not the case then their fires must have been of exceptional behaviour with a very low intensity and should not be considered representative of moorland management burning. Pseudoreplication could also explain the apparently anomalous 50% reduction in below-ground biomass between the 12-15-year-old stand and the 25-year-old stand which, by implication, they attribute to stand age (their table 2) but could more easily be explained by environmental differences between sites (e.g. due to peat depth or peat quality such as humification, aeration or water content).

Other aspects of the carbon budget for Mossdale Moor come from published data and from published models that are supported by on-site measurements of gaseous efflux of C[O.sup.2]. We are concerned that some of these relationships are taken out of context and are of limited value in the present situation but may create a false sense of confidence in the conclusions. For example, their estimation of soil respiration is based on models published by Chapman (1979) from heaths in Dorset in the south of England. These lowland heaths represent a very different climate from Mossdale Moor and are on well-aerated mineral soils or sand dune systems which contrast strongly with the peat-based upland soils of this study. The thermal response, moisture content, carbon content, bulk density and pH of these two soil types are likely to be very different; these are key drivers of response of soil respiration. Farage et al. rightly point out the sensitivity of soil respiration to temperature. We therefore hope that the lapse rate used for converting weather data at a meteorological station 20 km away from the field site was 0.61[degrees]C/100m rather than the 0.061[degrees]C/100m quoted in their paper. The respiration figure (0.281 [+ or -] 0.063 g C[O.sub.2]/[m.sup.2]) quoted by Farage et al. appears to be within reasonable limits (c.f. figures quoted in SEERAD 2007) and the estimate of error associated with this figure implies a degree of replication, but the details of methodology are inadequate to allow any evaluation of the spatial and temporal variation that this figure represents. Again we suspect that the data are pseudoreplicated or at the very least auto-correlated through time. Thus, the figures and the modelling approach derived from them should be treated with some caution.

In the UK the data in relation to the effect of fire on carbon processes in upland moorlands and peatlands are sparse. Garnett et al. (2000) conclude that burning reduced carbon storage from the Hard Hill designed burning/grazing experiment at Moor House in the Northern Pennines. Ward et al. (2007) examined gaseous carbon fluxes at the same site nine years after burning in the l0-year burning cycle and found that both photosynthetic and respiration rates appeared greater in burnt plots, though, on balance, the burning treatment sequestered more C[O.sub.2] than the unburnt treatment. They also found that methane fluxes were lower in the burnt treatment than the unburnt. This contrasts with data from laboratory experiments by Hogg et al. (1992), who suggest that ash deposition may increase methane fluxes, and with a pilot study on blanket peat in Forsinard in northern Scotland suggesting that methane fluxes can be heightened in the immediate post-fire period (Gray 2006). These data therefore imply that post-fire methane fluxes are likely to vary significantly through time; this may also be true for other trace gas fluxes. The overall effect of this temporal variation on the carbon balance for moorlands in relation to fire is unknown at the present time but it is an important additional factor that should be included in the simplified carbon balance calculations used by Farage et al.

Farage et al. estimate that about half of the carbon lost from a site due to burning may be due to surface erosion following the fire. This is based on estimates by Kinako and Gimingham (1980) and from Imeson (1971). Both of these studies measured reduction in the level of the soil surface following burning, but neither of them measured bulk density of the peat and did not, therefore, consider the possible effects of peat compaction and shrinkage due to drying. It is therefore likely that they significantly overestimated surface losses due to erosion.

Farage et al. state that the quantity of material combusted during routine burning depends primarily on the intensity of the fire. Our experimental fires (Davies et al. 2009) suggest that the quantity of fuel consumed in heather fires is only very weakly related to Byram's fireline intensity (the rate of energy release per unit length of fire front) within the range of conditions of normal management fires. Fireline intensity is the product of the quantity of fuel consumed and the rate of spread of the fire. The rate of spread depends largely on wind speed and the structure of the fuel (associated with stand age), but the proportion of fine fuel consumed varies surprisingly little between fires of different intensities in similar fuels; in effect, almost all of the fine heather fuel, but none of the coarse woody stems, is consumed in the majority of fires. Only a very small proportion of the unburnt stems remains as charcoal; the majority remains as wood that will decompose rapidly once incorporated into the litter layer. Forbes et al. (2006), reviewing data from fires in various habitats round the world, suggest that <3% of the carbon consumed by fires is transformed to black carbon. Black carbon (which includes but is not exclusively charcoal) represents a continuum of carbon compounds from partial pyrolysis of organic material to graphite and soot particles, some of which may be more readily decomposable than is generally assumed (Schmidt and Noack 2000; Gonzalez-Perez et al. 2004; Preston and Schmidt 2006; Knicker 2007). However, Forbes et al. (2006) acknowledge a wide variation in black carbon production; their data are not accompanied by fire characteristics such as intensity or severity and, crucially, these data are not from Calluna-dominated vegetation. In well-managed, relatively low-intensity burns we might expect the proportion of black carbon produced to be small, most of the unburnt stems remaining as easily decomposable wood.

Of much greater importance for carbon budgets than fire intensity is the moisture content of the moss and litter beneath the heather canopy. In most management fires in spring this is too wet to burn, but if dry (e.g. as in accidental summer fires; Legg and Davies 2009) then there may be a considerable additional loss of carbon. If the surface horizons of organic soil are also below a critical threshold (Rein et al. 2008) then there is a risk of very significant carbon loss through smouldering of the peat.

In theory, well-managed burning that ensures adequate regeneration of Calluna should be carbon neutral with respect to above-ground carbon stocks. However, it is the fate of the massive below-ground stores of carbon that is important. Research to determine whether management by burning and the maintenance of this carbon store are mutually compatible is urgently needed. This would of course need to account for the long-term effects that any change in fire regime might have on land use and vegetation type; there may be an increased potential for severe accidental fires that damage peat where fuels are no longer adequately managed or where vegetation and fire regimes change as a result of climate change (Legg and Davies 2009). Undoubtedly this will require a landscape-scale interdisciplinary approach to the question. Given the importance of the issue to landowners, conservationists, and policy makers, we need to ensure that the decision-making process is underpinned by a sound evidence base.

Additional keywords: Calluna vulgaris, rotational burning, upland heath, fire severity, carbon budgets.

10.1071/SR09166

Comment received 21 August 2009 accepted 27 November 2009

References

Chapman SB (1979) Some interrelationships between soil and root respiration in lowland Calluna heathland on southern England. Journal of Ecology 67, 1-20. doi:10.2307/2259333

Chapman SJ, Towers W, Williams BL, Coull MC, Paterson E (2001) 'Review of the contribution to climate change of organic soils under different land uses.' (Scottish Executive Central Research Unit: Edinburgh)

Davies GM, Gray A, Hamilton A, Legg CJ (2008a) The future of fire management in the British uplands. International Journal of Biodiversity Science and Management 4, 127-147.

Davies GM, Hamilton A, Smith A, Legg CJ (2008b) Using visual obstruction to estimate fuel toad and structure: the 'FuelRule' technique. International Journal of Wildland Fire 17, 380-389. doi: 10.1071/WF07021

Davies GM, Legg CJ, Smith AA, MacDonald AJ (2009) Rate of spread of fires in Calluna vulgaris-dominated moorlands. Journal of Applied Ecology 46, 1054-1063. doi:10.1111/j.1365-2664.2009.01681.x

Farage P, Ball A, McGenity TJ, Whitby C, Pretty J (2009) Burning management and carbon sequestration of upland heather moorland in the UK. Australian Journal of Soil Research 47, 351-361. doi: 10.1071/ SR08095

Forbes MS, Raison RJ, Skjemstad JO (2006) Formation, transformation and transport of black carbon (charcoal) in terrestrial and aquatic ecosystems. The Science of the Total Environment 370, 190-206. doi:10.1016/ j.scitotenv.2006.06.007

Garnett MH, Ineson P, Stevenson AC (2000) Effects of burning and grazing on carbon sequestration in a Pennine blanket bog, UK. The Holocene 10, 729-736. doi: 10.1191/09596830094971

Gimingham CH (1988) A reappraisal of cyclical processes in Calluna heath. Vegetatio 77, 61-64. doi:10.1007/BF00045751

Gonzalez-Perez JA, Gonzalez-Vila JF, Almendros G, Knicker H (2004) The effect of fire on soil organic matter a review. Environment International 30, 855-870. doi: 10.1016/j.envint.2004.02.003

Gray A (2006) The influence of management on the vegetation and carbon fluxes of blanket bog. PhD Thesis, The University of Edinburgh.

Hobbs RJ, Gimingham CH (1984) Studies on fire in Scottish heathland communities: I Fire characteristics. Journal of Ecology 72, 223-240. doi: 10.2307/2260015

Hogg EH, Lieffers VJ, Wein RW (1992) Potential carbon losses from peat profiles: Effects of temperature, drought cycles, and fire. Ecological Applications 2, 298-306. doi: 10.2307/1941863

Hurlbert SH (1984) Pseudoreplication and the design of ecological field experiments. Ecological Monographs 54, 187-211. doi:10.2307/ 1942661

Imeson AC (1971) Heather burning and soil erosion on the North Yorkshire moors. Journal of Applied Ecology 8, 537-542. doi: 10.2307/2402889

Kayll AJ (1966) Some characteristics of heath fires in north-east Scotland. Journal of Applied Ecology 3, 29-40. doi: 10.2307/2401664

Kinako PDS, Gimingham CH (1980) Heather burning and soil erosion on upland heaths in Scotland. Journal of Environmental Management 10, 277-284.

Knicker H (2007) How does fire affect the nature and stability of soil organic nitrogen and carbon? A review. Biogeochemistry 85, 91-118. doi: 10.1007/s10533-007-9104-4

Legg CJ, Davies GM (2009) What determines fire occurrence, fire behaviour and fire effects in heathlands? In 'Managing Heathlands in the Face of Climate Change. Proceedings of the 10th National Heathland Conference'. 9-11 September 2008, University of York. (Ed. I Alonso) pp. 45-55. Natural England Commissioned Report Number 014. Available at: http://naturalengland.etraderstores.com/NaturalEnglandShop/ Product.aspx?ProductID=5d76122d-3a43-4bc5-b28d-04d27flafed6 (accessed 14/9/2009).

Legg CJ, Davies GM, Kitchen K, Marno P (2007) A fire danger rating system for vegetation fires in the UK. The FireBeaters Phase I Report. The University of Edinburgh School of GeoScienees. Available at: http://hdl. handle.net/1842/3011 (accessed 14/9/2009).

Pearce F (2006) Grouse-shooting popularity boosts global warming. New Scientist 2564, 14. doi:10.1016/S0262-4079(06)60204-2

Preston CM, Schmidt MWI (2006) Black (pyrogenic) carbon in boreal forests: a synthesis of current knowledge and uncertainties. Biogeosciences Discussions 3, 211-271.

Rein G, Cleaver N, Ashton C, Pironi P, Torero JL (2008) The severity of smouldering peat fires and damage to the forest soil. Catena 74, 304-309. doi: 10.1016/j.catena.2008.05.008

Schmidt MWI, Noack AG (2000) Black carbon in soils and sediments: analysis, distribution, implications, and current challenges. Global Biogeochemical Cycles 14, 777-793. doi:10.1029/1999GB001208

Scottish Government (2009) Wildlife and Natural Environment Bill: Consultation Document. The Scottish Government, Edinburgh. Available at: www.scotland.gov.uk/topics/environment/wildlife-habitats/ wildnatenvbill (accessed 14/9/2009).

SEERAD (2007) 'ECOSSE--Estimating Carbon in Organic Soils Sequestration and Emissions.' Scottish Executive Environment and Rural Affairs Department, Edinburgh. Available at: www.scotland. gov.uk/Resource/Doc/170721/0047848.pdf (accessed 14/9/2009).

Tucker D (2003) Review of the impacts of heather and grassland burning in the uplands on soils, hydrology and biodiversity. English Nature Research Report No. 550, Peterborough.

Ward SE, Bardgett RD, McNamara NP, Adamson JK, Ostle NJ (2007) Long-term consequences of grazing and burning on northern peatlaud carbon dynamics. Ecosystems 10, 1069-1083. doi: 10.1007/s10021-007-9080-5

Yallop AR, Thacker JI, Thomas G, Stephens M, Clutterbuck B, Brewer T, Sannier CAD (2006) The extent and intensity of management burning in the English uplands. Journal of Applied Ecology 43, 1138-1148. doi: 10.1111/j.1365-2664.2006.01222.x

Colin Legg (A,D), G. Matt Davies (B), Alan Gray (C)

(A) The University of Edinburgh School of GeoSciences, King's Buildings, Edinburgh EH9 3JN, Scotland, UK.

(B) University of Washington, School of Forest Resources, Merrill Hall, Box 354115, Seattle, WA 98195-4115, USA.

(C) Centre for Ecology and Hydrology, Bush Estate, Penicuik, Midlothian, EH26 0QB, Scotland, UK.

(D) Corresponding author. Email: [e-mail]

Table 1. Pre-fire and post-fire biomass estimated for
26 experimental heather management fires

Fuel components for experimental fires at C: Crubenmore
Estate, near Dalwhinnie, Cairngorms National Park
(04[degrees]15'W, 56[degrees]57'N); and B: Black Hill, the
Pentlands near Edinburgh (03[degrees]20'W, 55[degrees]51'N).
Heather fuel and fine fuel are the pre-fire total shrub fuel
and fuel <2 mm diameter in the canopy estimated using the
non-destructive 'FuelRule' methodology which has been
robustly calibrated for use in heather moorlands (Davies et
al. 2008b). Moss-litter is similarly estimated by the
FuelRule method from the depth ofthe moss-litter layer
calibrated against harvested samples. Fuel consumed is
calculated by subtraction ofthe mass of heather stems
remaining after the fire measured by destructive harvesting
of two or three 50 cm by 50 cm quadrats randomly located
within each burn. Carbon consumed uses the ratio of carbon :
organic matter of 0.481 given by Farage et al. (2009). A
full description of the experimental design, plot set-up and
behaviour of these fires is given in Legg et al. (2007) and
Davies et al. (2009)

Site   Fuel            Date          Canopy    Heather   Fine
       category        burnt         height     fuel     fuel
                                      (cm)
                                                (kg/[m.sup.2])

C      Building        16/10/2003      19         0.72   0.54
C      Building        5/3/2004        27         1.17   0.96
C      Building        9/3/2004        12         0.42   0.40
C      Building        1/4/2004        21         0.86   0.67
C      Building        22/4/2004       20         0.76   0.62
P      Building        4/4/2006        16         0.78   0.72
P      Building        18/4/2006       17         0.90   0.69
C      Building        14/10/2006      19         0.86   0.67
C      Building        2/4/2007        26         1.22   0.76
                       Mean           19.5        0.85   0.67
                       s.d.           4.8         0.24   0.15

C      Late building   16/10/2003      28         1.17   0.87
C      Late building   5/3/2004        28         1.22   0.79
C      Late building   9/3/2004        30         1.07   0.82
C      Late building   1/4/2004        29         1.27   0.91
C      Late building   22/4/2004       27         1.02   0.73
P      Late building   4/4/2006        20         1.07   0.81
P      Late building   18/4/2006       29         1.44   1.00
C      Late building   14/10/2006      30         1.40   0.86
C      Late building   2/4/2007        29         1.39   0.95
                       Mean           27.8        1.23   0.86
                       s.d.           3.1         0.16   0.08

C      Mature          16/10/2003      45         1.67   0.80
C      Mature          5/3/2004        44         1.49   1.06
C      Mature          9/3/2004        47         1.55   0.80
C      Mature          1/4/2004        34         1.32   0.76
C      Mature          22/4/2004       39         1.34   0.89
P      Mature          4/4/2006        30         1.34   0.81
P      Mature          18/4/2006       23         0.99   0.65
C      Mature          14/10/2006      47         1.95   1.06
C      Mature          2/4/2007        51         1.46   0.59
                       Mean           40.0        1.46   0.82
                       s.d.           9.3         0.27   0.16

Site       Moss/         Fuel        Carbon
          litter        consumed    consumed

                       (kg/[m.sup.2])

C          0.59           0.62        0.30
C          0.77           0.95        0.46
C          0.52           (A)          (A)
C          0.79           0.71        0.34
C          0.76           0.61        0.29
P          0.56           0.53        0.25
P          0.57           0.76        0.37
C          0.65           0.78        0.38
C          0.79           1.11        0.53
           0.67           0.76        0.36
           0.11           0.19        0.09

C          0.79           1.00        0.48
C          0.88           1.03        0.50
C          1.08           0.91        0.44
C          1.04           1.03        0.50
C          0.99           0.89        0.43
P          0.61           0.89        0.43
P          0.89           1.19        0.57
C          1.38           1.24        0.60
C          1.11           1.24        0.60
           0.97           1.05        0.50
           0.22           0.14        0.07

C          1.50           1.17        0.56
C          1.84           1.11        0.53
C          2.20           1.06        0.51
C          1.82           1.08        0.52
C          1.57           1.08        0.52
P          1.17           0.50        0.24
P          0.99           0.85        0.41
C          1.93           1.01        0.49
C          4.23           0.87        0.42
           1.92           0.97        0.47
           0.95           0.21        0.10

(A) Plot failed to burn despite repeated ignition attempts.