Agricultural publications proclaim the good news: “Cattle Pastures May Improve Soil Quality”1 and “USDA Weighs In: Grazing Good for Soil & Environment.”2 Both headlines refer to the findings published in Franzluebbers and Stuedemann (2009)3—research demonstrating that soil sequesters more atmospheric carbon (C) as pasture managed under “low grazing pressure” (LGP) than as “unharvested” pasture (UH) left ungrazed by cattle, or as pasture subjected to “high grazing pressure” (HGP).
While this result may be correct, the study neglects to account for the amount of methane (CH4) (a short-acting, but potent greenhouse gas) that is emitted by the study’s cattle. Without such an accounting, it is impossible to conclude that any of Franzluebbers and Stuedemann’s grazing management prescriptions are superior to land management that does not include cattle. As I will argue below, this assertion is dependent upon a context that is not limited to that of the study’s localized soil (and the atmospheric C it sequesters) but has been expanded to the broader context of the global climate system.
Franzluebbers and Stuedemann arrived at their conclusion by evaluating the factorial combination of nutrient source and forage utilization on soil-profile distribution (0–150 cm) of soil organic carbon (SOC) during 12 years of
management on Typic Kanhapludult (Acrisol) in Georgia, USA. In measuring the concentration of SOC (p. 31, Table 2) and change rate of SOC (p. 33, Table 4), the authors found that grazing’s superiority (compared to ungrazed management) was greatest at soil depth 0–30 cm, with statistical significance decreasing as soil depth approached 150 cm. Therefore I will primarily focus on soil depth data in the range of 0–30 cm when comparing the change rate of SOC sequestration to CH4 emitted by cattle grazing on LGP test plots.
With management averaged over three nutrient source treatments accessed to a soil depth of 30 cm, I compute from the data (p. 33, Table 4) a rate of change for SOC of 1.40 Mg ha-1 year-1 for LGP management compared to only 0.797 Mg ha-1 year-1 for UH management. At face value this result lends significant support to the claim that grazing is superior to non-grazing in mitigating global climate change, as the grazed pasture sequesters approximately 1.76 times as much C as the ungrazed one.
But there’s a significant omission in this analysis as regards another greenhouse gas, one that in the short term has even greater potential than CO2 to produce global warming. That gas is methane (CH4) produced by cattle through enteric fermentation. Franzluebbers and Stuedemann address neither the mass of CH4 produced by the cattle in their experiments nor the mass of CH4 that is being absorbed from the atmosphere by the soil upon which their cattle graze. Despite the absence of this information in their article, reasonable estimates can be made.
Let’s first consider the mass of C emitted as CH4 by the steers that grazed the pasture. The LGP trials consisted of 5.8 steers per hectare grazing for 140 days per year during the first 5 years of the study, and for approximately 310 days per year during the remaining 7 years. Although CH4 emitted by a typical steer ranges from 60 to 71 kg per year,4 as a concession to ranching advocates, I’ll calculate CH4 emissions based on the low end of the range (60 kg year-1). And I’ll charge to the steers the CH4 emitted only for the time they were present on the test plots during the 12 years of the study.
The amount of CH4 emitted by the steers per hectare per year can be estimated by computing the weighted average of the annual CH4 emissions per steer over the two periods of the 12-year study, and then multiplying by the number of steers per hectare. This yields 0.228 Mg CH4 ha-1 year-1, of which about 75% by mass is C (0.170 Mg C ha-1 year-1). From the perspective of ranching advocates this still looks like a favorable result, as the mass of C sequestered by the soil (1.40 Mg C ha-1 year-1) is more than 8.2 times the mass of C emitted by the steers.
But this balance in favor of the soil sequestering atmospheric C may be less significant than it appears at first glance, as there are two additional factors about CH4 to consider.
First, CH4 has a property called Global Warming Potential (GWP) that denotes the relative ability of CH4 compared to CO2 to trap heat in the global climate system over a given time frame. Current studies peg the GWP of CH4 at “34” over a 100-year interval (GWP100) and at “86” over a 20-year interval (GWP20).5 Stated otherwise, over a 20-year interval, a given mass of CH4 would have the same effect in the global climate system as a mass of CO2 that is 86 times greater than that mass of CH4.
Authors of climate-related articles have often chosen to consider CH4’s impact over a 100-year period. But in 2013, the IPCC noted that “there is no
scientific argument for selecting 100 years compared with other choices.”6 Moreover, the IPCC found that at the 20-year timescale, total global emissions of CH4 are equivalent to over 80% of global CO2 emissions.7 In that light, Howarth (2014) argued for focusing on the 20-year, rather than the 100-year, period based on “the urgent need to reduce methane emissions over the coming 15–35 years.”8
The second factor that must be considered in regard to Franzluebbers and Stuedemann’s research is the amount of sequestered SOC that originates from CH4, rather than from CO2. Although I know of no results reported from the same region as their research, other studies can provide a reasonable estimate for the quantity of this CH4.
One such study by Wang et al. (2015)9 examined the soil sequestration of CH4 emitted by sheep at the Guyuan State Key Monitoring and Research Station of Grassland Ecosystem (China). There, “moderately grazed” grassland was found to sequester the greatest amount of CH4 compared to other types of land management. Their data yields an average daily uptake by soil of 0.02781 kg CH4 ha-1 day-1. If this result is then extrapolated to the average duration per year spent by steers during the 12 years of the Franzluebbers and Stuedemann study (i.e., 239 days per year), the soil would sequester 0.00665 Mg CH4 ha-1 year-1, approximately 2.9% of the estimated average annual CH4 emitted by the steers through enteric fermentation (i.e., 0.228 Mg CH4 ha-1 year-1).
In view of the possibility that the result obtained by Wang et al. (2015) is uncharacteristically low, let’s consider the study by Allen et al. (2009)10 that measured soil sequestration of CH4 in three regions of Australia: temperate, Mediterranean, and subtropical, using a network of paired pasture–forest sites,
representing three key stages of forest stand development: establishment, canopy closure, and mid to late rotation. For pasture soils, sequestration ranged from -0.876 kg CH4 ha-1 year-1 (a CH4 emitter) to 2.628 kg CH4 ha-1 year-1, while in forest soils the sequestration ranged from 0.08 kg CH4 ha-1 year-1 to 4.38 kg CH4 ha-1 year-1. In all cases, the rate of CH4 sequestered by soil was less than that derived from findings of Wang et al. (2015).
Consequently, I’ll assume that the mass of CH4 sequestered by the soil in the Franzluebbers and Stuedemann study is well represented by the mass sequestered in the study by Wang et al., namely 0.00665 Mg CH4 ha-1 year-1. Subtracting this amount from that which is emitted by the steers yields 0.2214 Mg CH4 ha-1 year-1 (i.e., 0.228 – 0.00665) added to the atmosphere.
Using a CH4 GWP20 of 86, the atmospheric CH4 remaining from the steers per hectare has a CO2 equivalency of 19.04 Mg CO2 ha-1 year-1 (i.e., 86 × 0.221 Mg CH4 ha-1 year-1). But the C sequestered by the soil (1.40 Mg C ha-1 year-1) represents only 5.19 Mg CO2 ha-1 year-1 (as C represents only 27% of a CO2 molecule’s mass). On balance, the CH4 emitted by the steers and the CO2 contributing to SOC yields a net atmospheric loading equivalent to 13.85 Mg CO2 ha-1 year-1.11
What other sources might annually produce 13.85 Mg of atmospheric CO2 pollution? For answers, I consulted the U.S. Environmental Protection Agency’s website12 that provides a number of possibilities. Among them we find that this quantity of CO2 is equivalent to consuming 32.2 barrels of oil, or burning 14,876 pounds of coal, or driving an average passenger vehicle 32,976 miles. And this is the air pollution generated by merely 5.8 steers grazing in accord with the experimental design of Franzluebbers and Stuedemann on only one hectare of land over the course of approximately two-thirds of a year. This is the prescription these authors tout as an environmentally beneficial land use to be replicated on 13.8 Mha of pasture across the eastern coastal and southeastern states of the U.S. (p. 28). Were such replication to occur, annual CO2-equivalent pollution of 191,130,000 Mg would ensue, which the just-cited EPA website equates to the CO2 pollution annually spewing from more than 50 coal-fired power plants.
Although my calculation of CH4-equivalency to CO2 has been performed with a value (86) associated with GWP20, it is worth noting that were this calculation performed with a value (34) associated with the frequently used GWP100, the atmospheric impact of CH4 would be reduced but still not offset by the atmospheric C sequestered by the soil upon which the steers grazed.
For completeness, I’ll consider the net C change rates for management under LGP and under UH assessed to the maximum soil depth investigated by Franzluebbers and Stuedemann: 150 cm. With the measurements averaged over three nutrient source treatments, I compute from data in Franzluebbers and Stuedemann (p. 33, Table 4) an LGP change rate of 0.796 Mg SOC ha-1 year-1 compared to a UH change rate of 0.28 Mg SOC ha-1 year-1.13 Again, as C represents only 27% of the mass of a CO2 molecule, the LGP value of 0.796 Mg SOC ha-1 year-1 yields a soil sequestration value of 2.95 Mg CO2 ha-1 year-1. When balanced against the CH4 emitted by the steers, the LGP treatment yields an atmospheric increase in CO2 equivalency of 16.09 Mg CO2 ha-1 year-1 (i.e., 19.04 – 2.95) when the impact of CH4 is assessed with GWP20 at 86.
In summary, the CH4 emitted by cattle in Franzluebbers and Stuedemann’s study contributed to global climate change far in excess of the increased rate of C sequestration achieved by the soil on which the cattle grazed relative to the rate of C sequestration with the UH management. Mitigation of global climate change would have been achieved within the experimental design only by foregoing even light grazing (LGP) and instead settling for the lower rate of soil C sequestration afforded by the UH management. This conclusion prevails regardless of whether the change rate of SOC sequestration is measured to a depth of 30 or 150 cm or whether the impact of CH4 is measured over 20 or 100 years.14, 15
Maximizing C Sequestration Depends on the Original Biome
Even selecting the UH option begs a larger question regarding land use if the highest objective is to reduce greenhouse gases through soil sequestration of C. Although the region of the Franzluebbers and Stuedemann test plots has been cropland since the early 19th century, from what sort of biome was that land converted? Was it grassland, somewhat comparable to the pasture studied in these experiments? Or was the land originally forest? The close proximity (only 16 km distant) of the Oconee National Forest strongly suggests the latter. Presumably, with sufficient time (and perhaps encouragement) that land would revert to forest. And if so, how much atmospheric C might the land then sequester?
That question has been addressed in research reported in Huntington (1995),16 conducted less than 75 km distant and approximately due west from the site of the Franzluebbers and Stuedemann experiments. For abandoned
cropland regenerating as forest over a 70-year period, Huntington found the rate of soil C sequestration to range from 0.34 to 0.79 Mg ha-1 year-117 (1.06 to 0.61 Mg ha-1 year-1 less than results of Franzluebbers and Stuedemann for LGP management averaged over three nutrient treatments and measured to a depth of 30 cm). But both rates of forest soil C sequestration are certainly superior to net C sequestration of LGP management when CH4 emissions of the cattle are considered.18 If the goal is to maximize the long-term sequestration of atmospheric C, rather than maximizing the rate of C sequestration, then additional data from Huntington strongly suggests that regenerating the landscape as native forest is superior to any management of the landscape as pasture.
Consider that in that same regenerating forest, Huntington found 82.1 Mg C ha-1 (soil depth 0–100 cm)19 compared to that of 69.9 Mg SOC ha-1 obtained with the best grazing management (i.e., LGP) evaluated by Franzluebbers and Stuedemann on pasture, averaged over three nutrient treatments to the greater (by 50%) depth of 150 cm (p. 31, Table 2). When above- and below-ground tree biomass was included, the total forest ecosystem sequestration soared to 185 Mg C ha-1,20 more than double the amount stored in Franzluebbers and Stuedemann’s pasture ecosystem that includes C in the below-ground21 and above-ground biomass22 annually remaining from the LGP management.
Finally, consider Huntington’s report of C sequestration in nearby native forest (Fernback Forest, Atlanta, GA) that has been only mildly disturbed. There, soil C was measured at 122 Mg C ha-1. Carbon in above- and below-ground tree biomass was 203.9 Mg C ha-1, and C sequestered in the total ecosystem was 326 Mg C ha-1.23 Would managed pasture (that was once native forest, such as that described in Franzluebbers and Stuedemann) ever sequester C in this amount? The likely answer is no, as the conversion of a grassland to a coniferous forest (being the “Potential Natural Vegetation” of the region) has been estimated to yield an increase within standing biomass of 157.5 MT (Mg) C ha-1.24
Comparing the findings of Franzluebbers and Stuedemann to those of Huntington reveals that if the highest objective is to reduce the greenhouse gas impacts of atmospheric C, then forestland should be left undisturbed. Forestland that has now become unproductive cropland should be returned to
forest if possible, not maintained as pasture. And because the heat-trapping properties of enteric fermentation-emitted CH4 will far outweigh any benefits associated with increased soil-sequestered C, the least desirable option would be to manage unproductive cropland as cattle-grazed pasture, even under the best grazing management.
The author thanks E. Patch, T. Shuman, and E. Walsh for valuable comments on earlier versions of this essay.
3. A. J. Franzluebbers and J. A. Stuedemann, “Soil-Profile Organic Carbon and Total Nitrogen During 12 Years of Pasture Management in the Southern Piedmont USA,” Agriculture, Ecosystems and Environment 129 (2009): 28–36.
4. K. A. Johnson and D. E. Johnson, “Methane Emissions from Cattle,” Journal of Animal Science 73(8) (1995): 2483–92.
9. Xiaoya Wang, Yingjun Zhang, Ding Huang, Zhiqiang Li, and Xiaoqing Zhang. “Methane Uptake and Emissions in a Typical Steppe Grazing System during the Grazing Season,” Atmospheric Environment 105 (2015): 14–21.
10. D. E. Allen, D. S. Mendham, Bhupinderpal-Singh, A. Cowie, W. Wang, R. C. Dalal, and R. J. Raison, “Nitrous Oxide and Methane Emissions from Soil are Reduced Following Afforestation of Pasture Lands in Three Contrasting Climatic Zones” Australian Journal of Soil Research 47(5) (2009): 443–58.
11. In this and subsequent calculations of net atmospheric C-based greenhouse gases due to steers (and the land upon which they graze) in the Franzluebbers and Stuedemann experiments, I have chosen an approach that biases the results in favor of graziers. I have subtracted from the steer-emitted CH4 a reasonable estimate of the quantity sequestered by the soil to allow for the possibility that the method used by Franzluebbers and Stuedemann to measure soil C was unable to account for sequestered C derived from atmospheric CH4. On the other hand, if the study measurements of soil-sequestered C include C derived from atmospheric CH4, as well as from CO2, then my approach reports the Franzluebbers and Stuedemann experiment as sequestering more atmospheric C than likely occurred. In either case, my numbers for virtual net atmospheric CO2 represent an upper bound on the efficacy of the Franzluebbers and Stuedemann experiments to sequester atmospheric C.
13. The lower change rates of SOC to depth of 150 cm relative to 30 cm result from a small but significant decline in SOC with depth below 30 cm. The authors suggest a few possible causes for this phenomenon, but none can be accepted with certainty.
14. Considering the cattle-emitted CH4 impact with GWP100 at 34 (under LGP management averaged over three nutrient treatments) yields a CO2 equivalency of 7.514 Mg ha-1 year-1. The atmospheric C (as CO2) sequestered by the soil to depth of 30 cm (i.e., 5.19 Mg ha-1 year-1) yields an atmospheric CO2 equivalency gain of 2.324 Mg ha-1 year-1. If the atmospheric C sequestered by the soil is considered to depth of 150 cm, then CO2 removed from the atmosphere is 2.95 Mg ha-1 year-1, yielding an atmospheric CO2 equivalency gain of 4.564 Mg ha-1 year-1.
15. A more recent paper in the same vein as Franzluebbers and Stuedemann is that by Megan B. Machmuller, Marc G. Kramer, Taylor K. Cyle, Nick Hill, Dennis Hancock, and Aaron Thompson, “Emerging Land Use Practices Rapidly Increase Soil Organic Matter,” Nature Communications, (2015) doi:10.1038/ncomms7995, http://www .nature.com/ncomms/2015/150430/ncomms7995/abs/ncomms7995.htm (accessed 26 July 2015). These authors examined the conversion of Georgia cropland to dairy pasture, but unlike Franzluebbers and Stuedemann, accounted for CH4 emissions from the ruminant animals (cows, in this case) that grazed the land.
Machmuller et al. performed a “whole farm C sequestration calculation” based upon the method reported in Jeff B. Belflower, John K. Bernard, David K. Gattie, Dennis W. Hancock, Lawrence M. Risse, and C. Alan Rotz, “A Case Study of the Potential Environmental Impacts of Different Dairy Production Systems in Georgia,” Agricultural Systems 108(C) (2012): 84–93 that used a CH4 GWP100 of 25 (based on the IPCC 2007 document, Climate change 2007: The physical Science Basis: Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, http://www.ipcc.ch/pdf/asses sment-report/ar4/wg1/ar4-wg1-chapter2.pdf (accessed 26 July 2015)). Based on that analysis, Machmuller et al. report that the farms they studied would be net C sinks for “at least an initial 5-year period following land use change.” Had the Machmuller et al. study been performed with the CH4 GWP20 of 86 (based on the more recent report by the IPCC, Climate Change 2013: The Physical Science
Basis, 714, Table 8.7, https://www.ipcc.ch/report/ar5/wg1/ (accessed 26 July 2015)) their claimed 5-year period as a net C sink would certainly have been reduced, if not eliminated. But even under the best case scenario acknowledged by Machmuller et al., the dairy operations under study would not be sustainable over the long term with regard to greenhouse gas emissions.
16. Thomas G. Huntington, “Carbon Sequestration in an Aggrading Forest Ecosystem in the Southeastern USA,” Soil Science Society of America Journal 59(5) (1995): 1459–67.
17. “Carbon Sequestration,” p. 1463.
18. Even the Huntington (1995) lower value for forest soil C sequestration (i.e., 0.34 Mg ha-1 year-1) is superior to Franzluebbers and Stuedemann’s LGP management which produces net atmospheric greenhouse gas loading equivalent to 12.3 Mg CO2 ha-1 year-1 (CH4 GWP20 at 86) or 0.634 Mg CO2 ha-1 year-1 (CH4 GWP100 at 34).
19. “Carbon Sequestration,” p. 1463, Table 1.
21. Franzluebbers and Stuedemann do not report the weight of below-ground biomass of their forage (i.e., Coastal Bermuda Grass) in LGP management, but note that their soil samples included roots (p. 30). In any case, below-ground temperate grassland biomass is estimated at only 6.3 metric tons (MT) ha-1 (equivalently 6.3 Mg ha-1) which provides an upper bound on the quantity of sequestered C. See T. M. Sobecki, D. L. Moffitt, J. Stone, C. D. Franks, and A. G. Mendenhall, “A Broad-Scale Perspective on the Extent, Distribution and Characteristics of U.S. Grazing Lands,” in The Potential of U.S. Grazing Lands to Sequester Carbon and Mitigate the Greenhouse Effect, ed. R. F. Follett, J. M. Kimble, and R. Lal (Boca Raton, Florida, Lewis Publishers, 2001), 44, Table 2.4.
22. Franzluebbers and Stuedemann report that the LGP management includes leaving 3 Mg ha-1 of forage, mostly Coastal Bermuda Grass. As this grass consists of approximately 44.6% C by weight (as calculated from data presented in “Broad-Scale Perspective,” p. 53; see Note 21 for details), the above-ground annual residual vegetation sequesters, at most, 1.34 Mg C ha-1.
23. “Carbon Sequestration,” p. 1463.
24. T. M. Sobecki, D. L. Moffitt, J. Stone, C. D. Franks, and A. G. Mendenhall, “A Broad-Scale Perspective on the Extent, Distribution and Characteristics of U.S. Grazing Lands,” in The Potential of U.S. Grazing Lands to Sequester Carbon and Mitigate the Greenhouse Effect, ed. R. F. Follett, J. M. Kimble, and R. Lal (Boca Raton, Florida, Lewis Publishers, 2001), 49 (Table 2.5).