Carbon sequestration in soils

Carbon sequestration in soils Jimes I? Bruce, Michele Frome, Eric Haites, Henry Janzen, R a ~ m La/:

Keith Pawtian

T

James P. Bruce is the Canadian Policy Representative, Soil and Water Conservation Society, Ottawa; Michele Frome is former Director of Poliq Programs, Soil and Water Conservation Sociq, Washington, DC;Eric Haites r j with Margaree Consultants, Toronto; Henry Janzen i s with Agriculture and AgriFood Can&, Letbbdge9 Alberta; Ramn L.al is with the School of Natural Resources9 The Ohio State Uniomity, Columbus, OH 432101085; Keith Paustian is with the Natural Resource Ecology Laboratory, Colorado State Univmig. *Correpondingauthor.

382

a1 production increases, much progress may be possible. Atmospheric concentrations of CO, and other greenhouse gases (GHGs) can be lowered by reducing emissions or by taking CO, from the atmosphere via photosynthesis and sequestering it in different components of terrestrial, oceanic, and freshwater aquatic ecosystems. In 1992, nearly all countries of the world signed the Framework Convention on Climate Change, and more than 160 nations have subsequently ratified this agreement, Its long-term goal is to “stabilize atmospheric concentrations of GHGs at a level that wilI prevent dangerous anthropogenic interference with the climate system.” As a step toward this goal, countries adopted the Kyoto Protocol in December 1997. To come into force, this protocol must be formally ratified by 55 Parties to the Convention, including enough Parties listed in Annex 1 (industrialized countries) to account for 55% of CO, emissions in 1990. Only Annex 1 Parties have commitments to reduce emissions between 2008 and 2012 by various amounts compared to 1990 emissions. The United States has accepted a target of 7% below 1990 levels and Canada a target 6% below 1990 levels. However, because with no emission reduction actions both countries would experience significant increases by 2010, the reductions below the projected “business as usual” emissions by that year are expected to be 25 to 30%. In the Kyoto Protocol, the GHGs of particular concern in exchanges with soils are CO,, methane (CH,), and nitrous oxide (N,O). This discussion paper focuses on CO,, by fir the most important GHG influenced directly by human activities. A. Creditsfbr increasing agricultural sinks. Sinks are defined in the Framework Convention as processes or activities that remove a G H G from the atmosphere. They have been included as potential offsets for emissions in the Kyoto Protocol in a limited manner. Agricultural soils are specifically recognized in the list of potential sources of GHG, which Parties must identify in their 1990 base year emissions and try to reduce. Farm management

JOURNAL OF SOIL AND W A T E R CONSERVATION

Copyright © 1999 Soil and Water Conservation Society. All rights reserved. Journal of Soil and Water Conservation 54(1):382-389 www.swcs.org

he purpose of this article is to examine (a) the magnitude of the potential for carbon sequestration in the soil as a means of reducing carbon dioxide (CO,) in the atmosphere, (b) some of the measures that might be used to achieve this potential, (c) the methods available for estimating carbon sequestration on a farm or regional level, (d) what is needed to achieve international consensus, and (e) additional information needs. This article is not presented as a definitive document but rather as an overview of where scientific opinion converges and where more work is needed. In addition, it aims to provoke discussion of the measures that can increase soil carbon sequestration and the policies that might be used to implement those measures. O n May 21-22, 1998, the Soil and Water Conservation Society conducted a workshop on carbon sequestration in soils with a broad group of scientists, policy analysts, and practitioners, Workshop participants reviewed and discussed the original draft of this paper. This version reflects modifications from the workshop. One of the goals of this article is to provide a more definitive consensus on the extent to which carbon sequestration in soil can contribute to greenhouse gas mitigation and to help Canada,and the United States meet their targets under the Kyoto Protocol to the Framework Convention on Climate Change of December 1997. An equally important goal is to identify governmental and private sector policies that can promote soil carbon sequestration. If similar actions can achieve both carbon sequestration and agricultur-

practices, especially conservation tillage, have reduced the amount of soil CO, emissions since 1990 in both Canada and the United States, and they will help achieve the targets. However, the only specifically identified sinks accepted in the ProtocoI are for land use changes and forests and those “verifiable changes in stock” due to afforestation, reforestation, and deforestation in the commitment period of 2008 to 2012 (Article 3.3). Considerable controversy has arisen over the ambiguous wording of Article 3.3 on how forestry carbon sink offsets are to be determined. At its session in Bonn, Germany, in June 1998, the Subsidiary Body for Science and Technology Advice (SBSTA) was offered several alternative interpretations. Because this Article requires countries to provide estimates of their 1990 carbon stock, one interpretation was t o measure the cumulative change in carbon stock as equal to the average stock in the 2008-2012 period minus the carbon stock in 1990. Another interpretation was to determine “the net change in emissions, as measured by changes in carbon stock from afforestation, reforestation, and deforestation activities, that may be used to ofiet emissions in the commitment period = carbon stock on 31 December 2012 minus carbon stock on 1 January2008.” The SBSTA delegates adopted the latter interpretation. As critics have pointed out, this interpretation means countries could take any actions they wished in the decade up to 2008-in the extreme, cutting down their forests (releasing much CO,) and then receiving credit for the regrowth and reforestation from January 2008 to December 2012 (within the commitment period). Whether this interpretation prevails at the formal meeting of the Conference of Parties in the h i r e (Buenos Ares and others) remains to be seen. It is also uncertain whether this interpretation of Article 3.3 (forestry) will also apply to Article 3.4, which addresses agricultural soil sinks. If it does, it will tend to reduce incentives for immediate action on agriculturd soil sinks; in addition, the short, 5year commitment period might make the measurement of a slow but steady gain in soil carbon difficult to verifjc As noted, agricultural sinks are referred to somewhat differently from forestry sinks, Article 3.4 of the Kyoto Protocol requires that the Conference of Parties (COP) “shall at its first session or as soon as practicable thereafter decide upon modalities, rules, and guidelines as to how (carbon) removals in agricultural soils and

measures, a significant portion of this amount may be realized over the next 20 to 30 years. B. Current Rates of Soil Carbon Change. Much of the carbon loss from agricultural soils occurs during the first decade after cultivation. With time, the rates of carbon loss have abated, because of both depletion of the readily decomposable carbon pools and gradual improvement in soil management practices. Consequently, most agricultural soils in the United States and Canada are now almost neutral with respect to emissions; they are neither large sources nor significant sinks of CO,. For example, computer simulations (Smith et al. 1997a) suggest that, on average, carbon loss from agricultural soils in Canada was only about 40 kglhdyr in 1990 and that rates of loss are declining. In a similar evaluation of soils in the central United States (Donigan et al. 1997), it was concluded that carbon losses had diminished and that soils are now beginning to accumulate carbon again. These findings, along with others from direct soil analyses, suggest the potential for the reversal of historical C trends (i.e,, the transformation of soils from a source to a significant sink for atmospheric CO,.

II, potentiabr &n

SeQUestration

A. Metbods of inmeusing soil cur&on. Changes in soil carbon content reflect the net result of carbon input (via plant litter) and carbon loss (via decomposition). To elicit a gain in carbon storage, therefore, a new management practice must (a) increase the amount of carbon entering the soil as plant residues or (b) suppress the rate of soil carbon decomposition. The former is a function of the net primary production (i.e., plant yield) and the proportion of the plant yield that is eventually returned to the soil in the form of plant litter or crop residues. The rate of decomposition is controlled by soil conditions ( e g , moisture, temperature, and oxygen SUKciency), composition of the organic material, placement of the material within the soil profile, and the degree of physical protection (e.g., within soil aggregates). Soil carbon storage can also change through erosion, which redistributes carbon across the landscape. Thus, some parts of the landscapes may lose carbon while others may gain carbon. Aggregate breakdown leads to rapid mineralization of carbon previously encapsulated within the aggregates, Because some of the eroded material is deposited elsewhere on the FIRST Q U A R T E R 1999

383


petagram = 1 billion metric ton) in soil organic carbon (SOC) and 800 to 1,000 Pg as soil inorganic carbon (SIC) or carbonate carbon (Post et al. 1982, Eswaran et al. 1993). The SOC content is generally high in virgin soils under grass or forest vegetation. Conversion of grass and forest land to cropland and pastures leads to losses of SOC. Grassland and forest soils tend to lose fiom 20 to 50% of the original SOC content in the zone of cultivation within 40 to 50 years after land use change (3, 4). Historical losses of SOC have been accentuated by low production levels, intensive tillage, inadequate use of fertilizers and organic amendments, removal of crop residue and biomass burning, and lack of soil protection against erosion and other degradative processes (Cole et al. 1993, d 1995). Estimates of historic loss of SOC from the cultivated soils of the world (croplands) range from 41 Pg (Houghton and Skole 1990) to 55 Pg (Cole 1996)). These estimates of historical losses of SOC provide a reference level for the potential of world soils to recover and sequester carbon through improved management. Assuming recovery of 50% of the historic soil loss, the carbon sequestration potential of world cropland over the next 50 to 100 years may be on the order of 20 to 30 Pg (Cole 1996). This equals 7 to 11% of emissions from fossil he1 combustion at 1990 levels, over 50 years. 2. Noritb America. Soils of N o r t h America account for about 22% of the terrestrial carbon pool. The total SOC pool to a depth of 1 meter is 267 Pg for North America, 52 Pg for the United States (contiguous 48 states), 13.5 Pg for Alaska, 190 Pg for Canada, and 11 Pg for Mexico (Waltrnan et al. 1997). Other estimates of the SOC pool of the 48 contiguous America states are on the order of 60 to 80 Pg. (Kern 1994). The cropland area in the United States is about 170 million ha (Mha), or about 19% of the total land area (Kellogg et al. 1994). It is estimated that the historic loss of SOC from U.S.cropland may be as much as 5 Pg (LAet al. 1998a). The improved cropland area in Canada is 45 Mha (Food and Agriculture Organization 1995), with a 1. Carbon in agricultural soils SOC pool of about 6 Pg to a depth of 1 meter (excluding pastureland). The historic loss of SOC from cropland in CanaA. Past carbon losses da may be about l Pg. Therefore, absolute potential of carbon sequestration with improved management of cropland 1. A global perspective in the United States and Canada is about World soils constitute a principal car- 5.5 Pg, if the losses could be Mly recovbon pool of 1,500 to 2,000 Pg (1 Pg = ered (see Section 3.2). With appropriate

land use changes shall be taken into account.’’ Article 3.4 also asks the C O P meeting in autumn 1998 (Buenos fires) to take advice on this matter from its SBSTA and from the scientific assessments and methodologies developed by the Intergovernmental Panel on Climate Change (IPCC). In general, these guidelines would apply to the second commitment period (after 2008-2012), but “A Party may choose to apply a decision ... [on this matter] for its first commitment period ...provided the activities to achieve the increased sinks have taken place since 1390.” Thus, a key to gaining international acceptance of carbon sequestration in agricultural soils is to have (a) a confident projection of the potential, narionally and internationally, and (b) an agreed methodology for determining “verifiable changes in stock,” B. A ~ r u f p m d w c t i u i i yIncreasing . soil organic matter (SOM) is widely recognized as a means to increase agricultural production. Principal processes of carbon sequestration in soil include hurnification of organic materials, aggregation by formation of organomineral complexes, deep placement of organic matter beneath the plow zone, deep rooting, and calcification. In contrast, leading causes of decline in SOM content include different soil degradative processes (e.g., erosion, compaction, decline in soil structure, minerdization, or oxidation of humic substances). These soil degradation processes are set-in-motion by anthropogcnic activities that include plowing, biomass burning, drainage of wetlands, improper grazing practices, and mining of soil fertility by low productivity subsistence agricultural practices. The SOM content is closely linked with soil quality and soil productivity. The objective of judicious soil management practices is to enhance soil quality through improvement of SOM content and to thereby increase agricultural productivity. The SOM content can be enhanced by adding biomass to the soil and curtailing or mitigating soil degradative processes (see Section 3).

landscape or in water systems, not all the carbon lost by erosion can be considered a net contribution to atmospheric CO,, For the same reason, soil carbon gains resulting from a practice that reduces erosion cannot be entirely equated to removal of atmospheric CO,.

384

J O U R N A L OF S O I L A N D WATER C O N S E R V A T I O N

(Nyborg et al. 1997). Other potential opportunities for increasing SOC include irrigation and reseeding with improved species or varieties (Table 1). On extensively managed grasslands (e.g., rangeland), grazing intensity and frequency are the main management variables that can affect soil carbon levels. Grazing can influence plant species composition, net primary productivity, above-ground and below-ground allocation in plants, and nutrient cycling pathways (Milchunas and Lauenroth 1993, Burke et al. 1997). The general conclusions that can be drawn horn these and other studies (Bauer et al. 1987, Frank et al. 1995, Manley et al. 1995) is that where the vegetation cover and production capacity of grasslands are not adversely affected by grazing, there is little change in SOM.However, in areas where overgrazing has seriously degraded vegetation coyer and primary production, soil carbon will be lower due to increased erosion losses and reduced carbon inputs. In these situations, improved management to restore productivity levels could lead to concomitant increases in soil carbon, 4. On degraded land. Soil quality, its agronomic productivity and environmental moderating capacity (Lal 1997), relies heavily on SOC content. Drastic reduction in SOC content leads to decline in soil quality and vice versa. Soil degradation is a more severe problem in arable land than in rangeland and foresdwoodland. O n a global scale, the problem of soil (and environmental) degradation is more severe in the tropics than in temperate regions, in dry rather than moist ecoregions, and warm rather than cold climates. Available statistics on soil degradation (Oldeman 1994) are vague and subjective, especially for the tropics and subtropics. Of a total estimated area of about 2 billion ha of .degraded soils worldwide, as much as 75% may be in the tropics. Total land area prone to soil degradation in North America is estimated at 96 Mha (Oldeman 1994), of which 70 Mha may be in the United States and Canada et al. 1995, Wicherek and Laverdiere 1993). Of the 70 Mha, degraded cropland in the United States and Canada may be 53 Mha (US.Department of Agriculture 1995). Soil degradation may be due to several processes, including accelerated soil erosion, salinization, drastic disturbance by mining and urban activities, overstocking and grazing land, decline in soil structure by vehicular traffic, and soil contamination by industrid pollutants (lal 1997). Restoration of degraded soils involves

wall


(Paustian et al. 1997b). Addition of organic amendments like livestock manure also promotes soil carbon by adding carbon directly, although this carbon is merely a recycling of crop carbon and does not necessarily represent a new input. Other agronomic options that may hrnish higher yields include improved crop varieties, better pest control, more efficient fertilizO n cultivated land er practices, and improved water manageNumerous strategies for increasing car- ment (including irrigation). These higher bon in cultivated soils have been identi- yields will translate into higher soil carbon fied (Table 1). These can be broadly clas- contents, provided the higher residue sified into four main approaches: (i) amounts are returned to the soil. reduction in tillage intensity; (ii) intensifi2. On revegetated or ‘j,t-usi&” hnd Percation of cropping systems; (iii) adoption haps the most effective way of restoring of yield-promoting practices, including soil carbon content on land that has been improved nutrient amendment; and (iv) cultivated is to reestablish and maintain reestablishment of permanent perennial perennial vegetation. Soil carbon increases vegetation. have been observed for both managed Tillage can promote soil carbon loss by (e.g., pasture establishment) and unmanseveral mechanisms: It disrupts soil aggre- aged (e.g., old-field successions) convergates, which protect SOM from decom- sions of cultivated lands. These increases position; it may stimulate microbial activ- in soil carbon can be attributed to the abity through enhanced aeration; and it sence of physical disturbance due t o mixes fresh residues into the soil, where tillage, increased carbon inputs resulting conditions for decomposition are often from less r e m o d of carbon in harvested more favorable than on the surface. Fur- crops, and greater allocation of carbon thermore, tillage can leave soils more below ground, particularly with perennial prone to erosion, resulting in further loss- grasses (Paustian et al. 1997b). Rates of es of soil carbon. As a result, adoption of accumulation vary, depending on climate practices with reduced tillage can result in and soil conditions. For example, mesic significant accumulation of soil carbon environments with high productivity (La et d. 1998a, Paustian et al. 1997b). show greater rates of accumulation (JasThis accrual of soil carbon can be further trow 1996) than less productive semi-arid enhanced if reduced tillage also increases grasslands (Burke et al. 1995, Dormaar yield (e.g., through improved moisture re- and Smoliak 1985). tention). With recent advances in seeding Although revegetation of cultivated equipment and weed control, reduced- lands with grasses or trees may achieve the and no-tillage systems are now practicable highest carbon gain per unit of land area, in many regions and cropping systems. it requires removing that land from annuMany cropping systems can be intensi- al crop production and does not therefore fied by increasing the duration of photo- apply on as wide a scale as do some of the synthetic activity. For example, greater use other carbon sequestration options. of perennial forages often enhances soil 3. On pastures and rangeland. Managecarbon, because these crops have extended ment factors that can impact soil carbon periods of active growth and allocate levels on grasslands in general include grazgreater proportions of their carbon below ing management, use of fire, species selecground (Paustian et al. 1997a). Other op- tion, and use of production inputs (e.g., portunities for intensification of cropping fertilizer, irrigation). In general, soils that systems include the use of winter cover have the highest capacity for increased carcrops (Kuo et al. 1997) and the elimina- bon levels are those that have been depleted tion of summer fdow. Intensification of of carbon in the past due to poor managecropping systems not only increases the ment or having been cultivated and used amount of carbon entering the soil, it for annual crop production. may suppress decomposition rates by On intensively managed grasslands cooling the soil through shading and by (e.g., pastures), where productivity and d y n g the soil. management inputs are relatively high, Application of nutritive amendments, there are opportunities for increasing soil including commercial fertilizers and or- carbon. Good pasture management has ganic amendments, favors soil carbon by the potential to increase soil carbon increasing yields and, consequently, the through improved practices such as rotaamount of residues returned to the soil tional grazing and application of fertilizers

reversion to natural vegetation cover, establishment of rapidly growing perennials and annuals, and application of inorganic fertilizers and organic amendments. If SOC content is severely depleted but the soil's resilience characteristics are functional, restorative measures can improve SOC content. The rate of carbon sequestration through soil restoration depends on antecedent properties, restorative measures, ecoregional characteristics, and the initial SOC pool under natural conditions.

rate of carbon accumulation during the first decades to 2020, we make the following two assumptions: a. The best possible carbon-consewing practices are adopted on all nondegraded cultivated land in the United States and Canada by the year 2000. In most cases, these practices will indude reduced tillage in conjunction with improved crop rotations, fertilization techniques, and other favorable practices, as appropriate for the given soil and climate (Table 1). tibc next two & c d s b. During the first two decades after adoption, the rate of carbon gain in re1. On cultivated hnd. Depending on Total potential carbon storage in the amount of potential carbon gain, it sponse to these practices averages 0.2 agricultural soils may take about a century for soils to reach Mglhdyear in soils where productivity is While there is strong evidence that their maximum carbon content. Rates of constrained by cool temperatures or aridiagricultural soils can become a net sink carbon accrual, however, are usually high- ty and 0.4 Mg/ha/year in other soils. for CO,, the eventual size of that poten- est in the first two decades after the man- Based on these assumptions, the gain of tial sink has a range of uncertainty. Soils agement change, then rapidly diminish, carbon in cultivated soils would be nearly cannot accrue carbon indefinitely. Based The rate of carbon gain after adoption 765 Tg C over two decades (see Table 2). on ecological principles (Odurn 1369), of an improved practice varies among 2. On revegetated or Sct-asdde"land+ soil carbon eventually reaches some equi- soils. For example, results were summa- Conversion of previously cultivated land librium value that cannot be exceeded eas- rized from 27 studies, primarily in the to perennial grassland usually results in ily. The maximum potentid gain, there- United States, which measured the in- high rates of soil carbon gain, The U.S. fore, is the difference between the current crease in soil carbon under no-till relative Conservation Reserve Program (CRP), carbon status and that eventual equilibri- to conventional tillage (usually moldboard started in 1985 and currently including um value. plowing) after a period of 5 to 20 years about 14 Mha of land planted to perenniA simple way to estimate potential car- (Paustian et al. 1997b). Carbon gains al grasses or trees, provides an estimate of bon gain is to predict the proportion of under no-till ranged from - 4 to +10 these rates of carbon accrual. Analysis of previously lost carbon that can be recov- Mg/ha (mean = + 3 Mg/ha). A similar soils on CRP lands in the western and ered with improved management. For range has been reported in a review of 17 central U.S. shows rates of