Impact of soil movement on carbon sequestration in agricultural ecosystems

Abstract

Recent modeling studies indicate that soil erosion and terrestrial sedimentation may establish ecosystem disequilibria that promote carbon (C) sequestration within the biosphere. Movement of upland eroded soil into wetland systems with high net primary productivity may represent the greatest increase in storage capacity potential for C sequestration. The capacity of wetland systems to capture sediments and build up areas of deposition has been documented as well as the ability of these ecosystems to store substantial amounts of C. The purpose of our work was to assess rates of sediment deposition and C storage in a wetland site adjacent to a small first-order stream that drains an agricultural area. The soils of the wetland site consist of a histosol buried by sediments from the agricultural area. Samples of deposited sediments in the riparian zone were collected in 5 cm increments and the concentration of ¹³⁷Cs was used to determine the 1964 and 1954 deposition layers. Agricultural activity in the watershed has caused increased sediment deposition to the wetland. The recent upland sediment is highly enriched in organic matter indicating that large amounts of organic C have been sequestered within this zone of sediment deposition. Rates of sequestration are much higher than rates that have occurred over the pre-modern history of the wetland. These data indicate the increased sedimentation rates in the wetland ecosystem are associated with increased C sequestration rates.

Introduction

From the period 1850 to 1998, it is estimated that 270 Gt C was emitted from industrial activity, with 176 Gt C accumulating in the atmosphere as CO₂ and 120 Gt C being assimilated by the ocean. Closure of this C budget requires a net terrestrial emission component of 26 Gt C (IPCC, 2000). In this same time period, it has been estimated that land use change by human activity has resulted in an emission of 136 Gt C. The gap between the estimated net terrestrial emission (26 Gt C) and land use change emissions (136 Gt C) has been termed the “missing C sink” (110 Gt C). In large part, the substantial size of this missing terrestrial C sink reflects our current lack of understanding of the terrestrial component. In this respect, it is noteworthy that the terrestrial component in most global C budgets is estimated primarily by the residuals after the other more easily better-characterized components have been evaluated.

Based on data from the previous decade (1990s), the Intergovernmental Panel on Climate Change (IPCC) complied a current annual global C budget as depicted in Fig. 1 (IPCC, 2000). Average annual emissions of CO₂ during this period were 6.3 Gt C year⁻¹ from fossil fuel and cement production, accumulation in the atmosphere was 3.3 Gt C year⁻¹ and net ocean uptake as 2.3 Gt C year⁻¹. To close this budget requires net terrestrial assimilation of 0.7 Gt C year⁻¹. It is noteworthy that the current global C budget estimates a net terrestrial uptake term for C, whereas the long term global C budget (1850–1998) estimate a net terrestrial emission for the period. This difference may be consistent with the output of two global C budget models that used different combinations of atmospheric, oceanic, and ice core CO₂ data and have indicated that the dynamics of the sink/source capacity of the terrestrial ecosystem have substantially changed within the last century with a transition from a C source to sink occurring in the period 1930–1940 (Bruno & Joos, 1997, Joos & Bruno, 1998). One of the models indicates that the global terrestrial biosphere turned from being a net source of 0.5 Gt C year⁻¹ into a net sink of 1.0 Gt C year⁻¹ during this period and estimates that between 1950 and 1990 the terrestrial biosphere had a net assimilation of 43 Gt C (Joos and Bruno, 1998). Interestingly, a recent report (Fan et al., 1998) indicated that the North American continent south of 51° N is an important, if not the major, terrestrial sink in the global C budget.

The finding of a major transition from source to sink in the recent global terrestrial ecosystem C budget and the possibility of a major terrestrial C sink residing within the North American continent indicates the need for research on the possible dynamics within the continental ecosystem that can account for the recent development of major C sinks. Potential causes for development of the source/sink transition include CO₂ fertilization increasing the biomass accumulation of the biosphere due to increased atmospheric content, but the magnitude of the transition is considered too large for this mechanism (Joos and Bruno, 1998). Other possible causes include increased usage of fertilizer N in agriculture that may stimulate C sequestration in soil because of the need for N to stabilize soil C.

One noteworthy occurrence, however, during the period of apparent source to sink transition (1930s & 1940s) was massive soil erosion associated with agriculture, which resulted in major redistribution of surface soil resources within the continental ecosystem. It is noteworthy that the current IPCC global C budget (Fig. 1) contains a term for export of terrestrial C to the ocean which is approximately equal to the net terrestrial uptake indicating that the terrestrial ecosystem is at steady state relative to total C storage. This implicates soil erosion as an important process influencing terrestrial C cycling via terrestrial C movement and ultimate export to oceans.

Stallard (1998) estimated annual soil erosion to be 9.6 t ha⁻¹ year⁻¹ (3.4 t ha⁻¹ year⁻¹ water and 6.2 t ha⁻¹ year⁻¹ wind) with total soil erosion of about 7.5 Gt year⁻¹ within the United States. Lal et al. (1998) estimated that erosion from cultivated cropland in the US has decreased from 1.9 Gt year⁻¹ in 1982 to 0.6 Gt year⁻¹ in 1995. This decrease in erosion was mostly the result of increases in crop land area under conservation tillage, conservation reserve programs, and afforestation programs. But even with these recent reductions, soil erosion remains the dominant force in redistributing soil C within agricultural landscapes.

A recent high profile debate over accuracy of the commonly used erosion models has brought attention to the concern about validity of current estimates of soil erosion in the United States (Trimble, 1999, Pimentel et al., 1999, Nearing et al., 2000, Trimble & Crosson, 2000). This concern is based on an argument that current erosion models measure soil detachment and not the amount exported from the watershed. With this argument much of the eroded soil predicted by these models is redeposited within the watershed and as a result the estimated “soil loss” is largely accounted for by redistribution on landscapes (Trimble and Crosson, 2000).

The influence of soil erosion and redistribution on terrestrial C storage is poorly understood. Some assessments have indicated that soil erosion has an overall negative influence on C storage in the terrestrial ecosystem (Lal et al., 1998). The expectations for increased greenhouse gas emissions due to erosion are based on (1) the decrease in the capacity of eroded soils to produce biomass; (2) the breakdown of soil structure that exposes C locked in aggregates to oxidation; and (3) the mineralization of exposed C by microbial decomposition and other oxidative processes. Some have estimated that 20% of organic C associated with erosion is mineralized to CO₂ because of increased biodegradation with loss of soil structure (Lal, 1995). The magnitude of C loss from a land surface due to erosion can be substantial and the dominant loss mechanism. For example, Harden et al. (1999) found at a Mississippi site cropped for more than 100 years that nearly 100% of the original organic C in the agricultural soil had been lost. They attributed as much as 80% of the reduction in organic C to soil erosion.

Soil erosion results in redistribution of soil C across the landscape and into riparian zones, waterways, lakes, and oceans where it may be sequestered over geologic time. Ritchie, 1988, Ritchie, 1989 provided an assessment of C storage in reservoirs and navigation pools of the Upper Mississippi River within the United States and concluded that C sequestration in reservoirs alone may constitute a global C sink of 0.2–0.3 Gt year⁻¹. Others have estimated that globally 20 Gt year⁻¹ of sediment is deposited in world oceans which translate into large exports of terrestrial C into ocean sediment with likely geologic storage. Sediment deposition in lakes, reservoirs, and ocean represents passive storage of C and such deposition is not expected to stimulate rates of net C sequestration per se.

Typically only 10% of eroded soil is exported from a watershed (Lal et al., 1998), and therefore, in large part the eroding soil represents redistribution of soil resources on the landscape. This redistribution of soil resources can lead to deep burial of terrestrial C (Lal et al., 1998) that is a mechanism for removing organic C from more reactive components of the biosphere and may constitute a largely passive form of C storage. However, consideration should also be given to influence of soil resource redistribution on active mechanisms of C sequestration within landscapes. An emerging concept is that soil resource redistribution within the terrestrial biosphere may establish states of ecosystem disequilibria that may promote C sequestration.

Stallard (1998) developed a model framework for linking terrestrial sedimentation to the C cycle. The model was built on two core hypotheses: (1) significant amounts of C are buried during the deposition of sediments in terrestrial environments because of human acceleration of erosion and modification of hydrological and nutrient cycles; and (2) pedogenic C that is buried is replaced by newly fixed pedogenic C at sites of erosion or deposition. Elements of the model include (1) the consideration of enhanced mobilization of sediments and nutrients and the processes that limit their delivery to the ocean; (2) the consideration of whether newly fixed C is being added at the site of erosion or deposition and whether organic C is preserved with burial; and (3) the consideration of organic sedimentation in wetlands. With this model framework, Stallard developed a wide-ranging set of scenarios to evaluate the impacts of terrestrial sedimentation on the global C sequestration. The results ranged from 0 to 2.3 Gt C year⁻¹ of C sequestration on a global basis, averaging 0.8 Gt C year⁻¹ and with estimates for human-induced burial of 0.6–1.5 Gt C year⁻¹ by terrestrial sedimentation using conservative assumptions. Stallard (1998) noted substantial uncertainty because of large information gaps concerning rates and locations of terrestrial sedimentation. This led to an inability to more rigorously test the hypothesis linking terrestrial sedimentation to the global C cycle and to assess magnitude of this sink. The information gap was largely attributed to the lack of good methods for tracking movement of soil redistribution, sedimentation, and C in terrestrial ecosystems.

Harden et al. (1999) tested elements of this hypothesis by measuring the potential for dynamic replacement of soil C in a highly erosive Mississippi loess soil. Using soil C, N, “bomb” ¹⁴C, and measurements of CO₂ flux, in conjunction with the Century model and spreadsheet models, permitted characterization of C storage and dynamics for different erosion and tillage histories. This study found that about 100% of prehistory soil C had been lost from sites of erosion with more than 127 years of land use and subsequently that about 30% of the C at sites of erosion had been replaced since 1950. This provided strong evidence for the principle of dynamic replacement of soil C and indicates that soil erosion can induce formation of a local C sink. Harden et al. (1999) conclude, however, that the ultimate source/sink influence of soil erosion is highly dependant on fate of sediment C that includes degree and time length of burial and relative rates of organic C decomposition in the receiving ecosystem. This conclusion further emphasizes the need for improved methods to track the movement of eroded soils in terrestrial ecosystems and assess C dynamics within the ecosystem receiving eroded soils.

Based on work by Stallard (1998) and Harden et al. (1999), a conceptual relationship can be developed for the interaction between soil erosion, C sequestration and net primary productivity (NPP) at the site of erosion (Fig. 2). Within the concept, increasing soil erosion induces C sequestration by pedogenic disequilibria caused by soil C loss and this influence may at least partly be counteracted by loss of NPP. Low to moderate rates of soil erosion may stimulate C sequestration whereas severe erosion would be detrimental to C sequestration in the ecosystem because of the loss of NPP. This model addresses sites of erosion but does not account for C dynamics at zones of deposition within terrestrial ecosystems.

Watershed-scale studies are needed to more fully assess impact of soil resource movement on terrestrial C storage. We test the hypothesis that agricultural activities within a watershed can markedly increase rates of C storage within the riparian buffer by increasing sediment deposition and contributing nutrients to wetland ecosystems, which may increase primary productivity.

Research on a small gauged headwater watershed at BARC (Beltsville Agricultural Research Center) in Maryland provided a means for assessing agricultural influence at a watershed scale (Fig. 3, Fig. 4). Characterization of the site includes an intensive topographic survey of the agricultural field and riparian ecosystem (Fig. 3). The agricultural field was sampled on a 25 m grid for assessing spatial distribution of soil C within the field (Fig. 5). Soil cores (15 cm dia.×35 cm) were taken in a series of transects across the agricultural field and the riparian ecosystem for measurement of the relationship between soil C and ¹³⁷Cs within the catchment (Ritchie and McHenry, 1990). Deeper profile samples (0–180 cm) were also obtained within the wetland to inventory C storage within the ecosystem. For a control site, soil samples were obtained from a pristine (undeveloped) first-order catchment located at the Patuxent Wildlife Research Center within 2.5 km of the agricultural study site. Soil samples were collected along transects within this catchment and they were analyzed for C and ¹³⁷Cs. Soil samples were analyzed for ¹³⁷Cs using a Canberra Genie-2000 spectroscopy system and analyzed for soil C by a Leco CNS 2000 dry combustion instrument. Radiocarbon (¹⁴C) measurements were performed by Beta Analytic Inc. (Miami, FL).