Scaling methane emissions from vegetation Anthony J. Parsons 1 , Paul C.D. Newton 1 , Harry Clark 1 and Francis M. Kelliher 2 1 AgResearch, Private Bag 11008, Palmerston North, New Zealand 2 Landcare Research, PO Box 69, Lincoln, New Zealand In a recent highly publicised paper, Keppler et al. [1] report the remarkable observation that methane is emitted directly by plants, and in an aerobic environment. Pre- viously, emissions from vegetation were attributed to anae- robic microbial associations (e.g. in water-logged soils). But how important is this new source of what is a potent green- house gas? Keppler et al. [1] make a calculation that would estimate plant emissions to comprise 11–46% [2] of the current total annual methane emissions. We contend that this estimate needs to be revised, as the scaling used by Keppler et al. might not be the most appropriate. Keppler et al. grew plants in chambers and published emission rates expressed per unit of enclosed ‘standing’ biomass [1]. Different rates of emission were observed for intact material in the light, compared with in the dark. However, when scaling up to global ecosystems, Keppler et al. multiplied up by the NPP (net primary production of both shoot and root), that is, on the basis of the growth rate rather than on the standing biomass of each biome. We find this calculation questionable for two reasons: first, there is no evidence in their paper for a link between emissions and growth rate. Second, the authors calculate annual emis- sions by multiplying their observed emission rate per day per gram of standing biomass, by the annual (i.e. accumu- lated over a year) growth rate of vegetation, and then again by the length of the growing season. This is the equivalent of multiplying twice by the length of the growing season. As their measurements are taken from predominantly leafy (or young) example species (see Online Supplemen- tary Information for [1]), we suggest that biome leaf mass is a more appropriate scaling factor, and we have recalcu- lated their global figures on this basis [Table 1, column (4)]. These figures reduce the estimate of mean total global methane emissions from vegetation by 72%. However, the reduction is not of the same order across all biomes, because the relationship between standing biomass and NPP differs. In some cases, the value of NPP can be similar to the standing biomass; however, in biomes with low turnover rates, such as tropical forests, NPP is substan- tially smaller than standing biomass; in those with high turnover rates, such as temperate grasslands, NPP is much larger than standing biomass. There remains the question of whether non-leafy tissues release methane; if this were the case, then the potential release from systems such as tropical forests would increase enormously, because the leaf mass of this biome is 500 g m 2 whereas the total standing biomass is 30 400 g m 2 [3]. Although Keppler et al. present no data to answer this question, the rates of methane emission are likely to be lower than for leaves, otherwise the calcu- lated value for tropical forests would give a source of 1213 Tg CH 4 yr 1 , approximately double the current esti- mate for all known methane sources [2]. For interest, we assigned the dark emission rates used by Keppler et al., for example for leaf litter, to the non-leafy fraction as one possible estimate of this flux should it exist. The results in column (5) of Table 1 show only a small further increase over the values calculated for leaf fraction only. Our re-calculated values do not diminish the signifi- cance of Keppler et al.’s finding or its relevance to some apparent anomalies in the global methane budget, such as the mounting evidence of the need to increase estimates of Update TRENDS in Ecology and Evolution Vol.21 No.8 423 Table 1. Calculated global methane emissions from vegetation as presented by Keppler et al. [1] using biome annual NPP, our re-calculated estimates based on standing leafy biomass, and additional contributions from non-leafy biomass a Vegetation type or biome Estimated by Keppler et al. (Tg CH 4 yr 1 ) b LAI c (m 2 m 2 ) SLA (m 2 kg DW 1 ) c Leafy biomass (Tg CH 4 yr 1 ) Non-leafy biomass (Tg CH 4 yr 1 ) Tropical forests 78.2 6 12.0 15.6 (80) 7.3 Temperate forests 17.7 6 8.5 8.0 (55) 2.0 Boreal forests 3.0 3.5 7.7 3.6 (+20) 0.4 Mediterranean shrublands 2.7 2 6.9 0.8 (75) 0.1 Tropical savannas and grassland 29.2 5 16.9 8.0 (73) 0.8 Temperate grasslands 7.4 3.5 16.9 2.0 (67) 0.004 Deserts 3.8 1 6.9 2.2 (42) 0.02 Crops 7.2 4 24.5 2.2 (69) 0.04 Total 149 N/A N/A 42 (72) 10.7 a Growing season and day length are identical to those used by Keppler et al. [1]. Leafy mass derived from biome LAI is taken from the same source as Keppler et al.’s NPP [3] and SLA [5]. Values in brackets are the difference between columns (1) and (4) in %. b From [1]. c Abbreviations: LAI, leaf area index; SLA, specific leaf area. Corresponding author: Newton, P.C.D. (paul.newton@agresearch.co.nz). Available online 15 June 2006 www.sciencedirect.com