Global Change Biology (2010) 16, 427–438, doi: 10.1111/j.1365-2486.2009.01981.

x

Contribution of trees to carbon storage in soils of silvopastoral systems in Florida, USA
S O L O M O N G . H A I L E *w , V I M A L A D . N A I R w and P. K . R A M A C H A N D R A N N A I R *
*Center for Subtropical Agroforestry, School of Forest Resources and Conservation, PO Box 110410, University of Florida,

Gainesville, FL 32611, USA, wSoil and Water Science Department, PO Box 110510; Institute of Food and Agricultural Sciences, University of Florida, Gainesville, FL 32611, USA

Abstract
Silvopastoral systems that integrate trees in pasture production systems are likely to enhance soil carbon (C) storage in lower soil layers due to the presence of deep tree roots. To quantify the relative soil C contribution from trees (C3 plants) and warm season grasses (C4 plants) in silvopastoral systems, soil samples were collected and analyzed from silvopastures of slash pine (Pinus elliottii) 1 bahiagrass (Paspalum notatum), and adjacent open pasture (OP), at six depths down to 125 cm, at four sites representing two major soil orders (Spodosols and Ultisols) of Florida. The plant sources of C in whole (nonfractionated) and three soil fraction sizes (250–2000, 53–250, and o53 lm) were traced using stable C isotope signatures. The silvopasture sites contained higher amounts of C3derived soil organic carbon (SOC) compared with OP sites, at all soil depths. Slash pine trees (C3 plants) seemed to have contributed more C in the silt 1 clay-sized (o53 lm) fractions than bahiagrass (C4 plants), particularly deeper in the soil profile. Spodosols sites contained more C in the o53 lm fraction at and below the spodic horizon (occurring between 15 and 50 cm) in silvopasture compared with OP. The results indicate that most of SOC in deeper soil profiles and the relatively stable o53 lm C fraction were derived from tree components (C3 plants) in all the sites, suggesting that the tree-based pasture system has greater potential to store more stable C in the soil compared with the treeless system.

Nomenclature:
OP 5 SOC 5 SP-A 5 SP-T 5 open pasture soil organic carbon center of the alley in a silvopasture in-between trees in a row of a silvopasture

Keywords: agroforestry, bahiagrass pasture, carbon sequestration, slash pine, stable carbon isotope

Received 10 December 2008; revised version received 21 April 2009 and accepted 7 May 2009

Introduction
Agroforestry systems (AFS) that combine trees and/or shrubs with crops and/or livestock production are planned and managed agroecosystems (Garrett et al., 2000). Increasing the overall productivity and efficiency of the land-use system are major goals of agroforestry (Nair, 2005). AFS have the potential to enhance carbon (C) sequestration in soil compared with treeless (agri-

Correspondence: Vimala D. Nair, tel. 1 352 392 1804, fax 1 352 392 3399, e-mail: vdn@ufl.edu

cultural) systems (Montagnini & Nair, 2004). Such claims are based on the premise that the tree components in AFS can be significant sinks of atmospheric C due to their long-term storage of high amounts of C in biomass, especially in the deep root systems. Furthermore, higher diversity of grassland species and specific plant functional traits were reported to increase uptake of C into the soil system through resource partitioning (Steinbeiss et al., 2008). Similarly Saha et al. (2009) found that homegardens with higher, compared with those with lower, number of plant species, as well as higher species richness and tree density had higher soil C, especially in the top 50 cm of soil.
427

r 2009 Blackwell Publishing Ltd

428 S . G . H A I L E et al.
Silvopasture – the integration of trees into forage or/ and livestock – has been practiced in the southeastern USA as ‘tree-pasture’ or ‘pine-pasture’ since the early 1950s (Nair et al., 2005). Indeed, silvopasture is the most common form of agroforestry in North America (Garrett et al., 2000; Nair et al., 2008). These agroecosystems are usually established by incorporating trees in existing, managed pastures of bahiagrass (Paspalum notatum ´ ), common bermudagrass [Cynodon dactylon (L.) Fluegge Pers. var. dactylon] or other similar grasses (Workman et al., 2003). Functional consequences of integration of trees into grass-dominated vegetation include changes of aboveand belowground productivity (Scholes & Hall, 1996; Archer et al., 2001), modifications to rooting depth and distribution (Gill & Burke, 1999), and changes in the quantity and quality of litter inputs (Connin et al., 1997; ´ gy & Jackson, 2000). Soil Jackson et al., 2000; Jobba organic matter (SOM) is extremely vulnerable to landuse change (Paul et al., 1997; IPCC, 2001), as well as to intensification of agricultural practices (Matson et al., 1997). Thus, in order to quantify the strength and longevity of the C sink in tree-based pasture systems, it is important to understand the mechanisms and processes associated with C transformation and storage. SOM has a very complex and heterogeneous composition, and is associated with mineral soil constituents to form soil aggregates. The nature and extent of turnover of soil organic carbon (SOC) is intimately linked to organic matter size fractions as well as to soil structure and extent of aggregation (Martens, 2000). Different components of SOC have different residence times, ranging from labile to stable forms (Carter, 1996). Thus, to evaluate changes in soil C and SOM dynamics correctly, it is necessary to separate out functionally different SOM fractions. Soil size fractionation helps to differentiate these different SOC fractions, (Christensen, 1992); it is based on the premise that SOC associated with sand-size aggregates (or macro-organic matter 4250 mm) is often more labile than SOC in the clay and silt fractions (Tiessen & Stewart, 1983). Stable C isotope-ratio analysis in SOC studies emerged as a tool to trace the source of SOC to C3 and C4 components in vegetation (Stout et al., 1981). Numerous studies (Bernoux et al., 1998; Ehleringer et al., 2000; Accoe et al., 2002; Swap et al., 2004) have been successful in applying d13C to understand plant–soil SOM dynamics, making stable isotopic analysis a useful technique. When one type of vegetation is replaced with another, d13C values can be used to identify SOM derived from residues in the native vegetation and the new vegetation based on discrimination between C3 and C4 plants. The reported d13C values range from À9 to À19% for C4 plants and À20 to À35% for C3 plants (Nyberg et al., 2000; Biedenbender et al., 2004; Staddon, 2004). When a C4 plant is introduced to a system that had previously been under a C3 plant (or vice versa), the relative contribution of new vs. old soil organic C can be quantified using the mass balance of stable isotope contents based on the change in 13C signature of SOM (Dawson et al., 2002; Del Galdo et al., 2003). In a combined tree 1 grass land-use system, C3 inputs are dominated by either woody shrubs or trees and C4 inputs are dominated by grass (McClaran & McPherson, 1995). In pine-based silvopastoral systems where C3 and C4 plants are grown simultaneously, such studies that use the natural abundance of d13C to understand the C dynamics are rare. The d13C isotope technique requires comparison between a site where the photosynthetic pathway of the dominant vegetation (C3 or C4) has been changed and a reference site where photosynthetic pathway of the vegetation remains unchanged. The plant community in silvopasture systems of the southeastern USA is composed of slash pine (Pinus elliottii) (a C3 plant; d13C  À29.5%) and C4 plants dominated by bahiagrass (P. notatum) (d13C  À13.3%). The d13C value ranges of C3 and C4 plants do not overlap. Therefore, differences in isotope ratio can be used to quantify the relative contribution of plants of each photosynthetic pathway to SOM (Balesdent et al., 1988). Thus, combining SOM fractionation techniques with the 13C natural abundance technique offers a compelling approach to investigating small shifts in soil C stores that would be significant in the long term, which might not be detected by conventional methods. A change in land use from open pasture (OP) to silvopasture presents a unique opportunity to use the stable C isotope methodology to study SOC dynamics following the integration of trees into OP. The present study was undertaken in this context, with the objectives of assessing the impact of integrating trees in pastures systems on the SOC content and SOM fraction size compared with OP systems, and quantifying the relative SOC contribution of trees and warm-season grasses of silvopastoral systems using their natural C isotopic differences.

Materials and methods

Study area
Soil samples were collected from four sites, located in Alachua (29145 0 N, 82133 0 W), Osceola (2819 0 N, 81110 0 W), Hardee (27113 0 N, 8218 0 W), and Suwannee (30124 0 N, 8310 0 W) counties in Florida, USA. Hereafter, the sites are designated by county names where the farms are located. The sites represented two major soil orders [Spodosols (Hardee and Osceola counties) and Ultisols
r 2009 Blackwell Publishing Ltd, Global Change Biology, 16, 427–438

CONTRIBUTION OF TREES TO CARBON STORAGE IN SOILS
(Alachua and Suwannee counties)] of Florida. The Spodosol sites are on Ona (sandy siliceous, hyperthermic Typic Alaquods) or Immokalee fine sands (sandy siliceous, hyperthermic Arenic Alaquods), whereas the Ultisol sites are on Kendrick (loamy, siliceous, semiactive, hyperthermic Arenic Paleudults) or Blanton sands (loamy, siliceous, semiactive, thermic Grossarenic Paleudults). Detailed description of the study sites can be found in Haile et al. (2008). At each site, soils were sampled in a silvopasture with slash pine (P. elliottii ´ ) and adEnglem) and bahiagrass (P. notatum Fluegge jacent open (treeless) pasture with bahiagrass. The soil particle-size distribution in these sites ranged from 87% to 96% sand, 2% to 4% silt, and 3% to 9% clay (Haile et al., 2008). At each site, silvopasture and adjacent treeless pasture plots were selected from which soil samples were taken. Slope, aspect, and soil series were uniform across plots within a site, thus ensuring that the land-use system (pasture vs. silvopasture) was the primary factor influencing the soil C content in plots. The OPs on Suwannee, Osceola, Hardee, and Alachua farms, were 40, 22, 48, and 55 years old, respectively, since pasture establishment. The respective ages (years) of silvopastures since establishment at the four sites were: Suwannee (40), Osceola (14), Hardee (12), and Alachua (8).

429

slash pine (C3) plant parts was prepared to determine the d13C values. A similar procedure was adopted for bahiagrass (C4 plant) by collecting leaves (shoots) and roots of grass from the OPs.

Size fractionation
Soils were physically fractionated by wet-sieving following a procedure using the disruptive forces of slaking and wet-sieving through a series of two sieve sizes (250 and 53 mm) to obtain three soil fraction size classes, as described by Elliott (1986) and Six et al. (1998). Briefly, all field-moist composite samples were air dried and passed through a 2 mm sieve. A subsample of 100 g of the composite sample was submerged in deionized water with disruptive forces of slaking for approximately 5 min before placing it on top of 250 mm sieve. The sieving was done manually by moving the sieve up and down approximately 50 times in 2 min. The fraction remaining on the top of a 250 mm sieve was collected in a hard plastic pan and allowed to dry in oven at 65 1C and weighed. Water plus soil o250 mm was poured through a 53 mm sieve and the same sieving procedure was repeated. The overall wet sieving procedure yielded a water-stable macro-sized fraction, 250– 2000 mm; a micro-sized fraction, 53–250 mm, and a silt 1 clay-sized fraction, o53 mm. The recovery of mass soil fractions after overall wet sieving procedure ranged from 96% to 99% of the initial soil mass. Takimoto et al. (2008) reported a recovery range from 97% to 99% mass soil fractions using a similar fractionation procedure.

Sampling
Soil samples were collected from three different sampling plot sets: two locations in the silvopasture sites including between trees in a row at about 50 cm from a tree trunk (SP-T) and at the center of an alley at 2–3 m from tree trunk (SP-A), and one set of samples from OP. Each of these sample sets had stratified grid sampling points made by three parallel rows with four sampling points in a row. The distance between samples was on average 50 m. From each sampling point, soils were collected using an auger, from six soil depths: 0–5, 5–15, 15–30, 30–50, 50–75, and 75–125 cm. While in the field, a composite sample was created for each depth interval. Equal amounts of soil from corresponding depth intervals of four sampling points in a row were mixed. Thus, at each site, there were three composite samples for each depth interval. The composite samples [each soil depth interval in a row (grid line in OP)] constituted the experimental unit of analysis and each unit had three replications (rows), giving a total of 216 samples [six depths  three replications  three ‘locations’ (SP-A, SPT, and OP)  four sites/counties] (Haile et al., 2008). Needles, roots, wood (branch), bark, and cones were collected from slash pine trees of silvopastures. The samples were oven-dried at 60 1C for 72 h, and ground to fine powder using a mill. A composite sample of the
r 2009 Blackwell Publishing Ltd, Global Change Biology, 16, 427–438

Chemical analysis
For chemical analysis, whole and fractionated soils were oven-dried at 60 1C for 72 h, and ground to fine powder using a ball mill (Cianflone Scientific Instruments, Pittsburgh, PA, USA). Total SOC content was determined for whole and fractionated soil samples by dry combustion on an automated FLASH EA 1112 N C elemental analyzer (Thermo Fisher Scientific Inc., Waltham, MA, USA).

Stable C isotope analysis
Soil samples were analyzed for C concentrations and for d13C values using a Carlo Erba EA-1108 (CE Elantech, Lakewood, NJ, USA) interfaced with a Delta Plus (Thermo Finnigan, San Jose, CA, USA) isotope ratio mass spectrometer operating in continuous flow mode. C isotope ratios are presented in d-notation: d13 C ¼ ð½RSAMPLE À RSTD Š=RSTD Þ Â 103 ; ð1Þ

430 S . G . H A I L E et al.
where RSAMPLE is the 13C/12C ratio of the sample, and RSTD is the 13C/12C ratio of the Vienna Pee Dee Belemnite (VPDB) standard (Coplen, 1996). Precision of duplicate measurements was 0.1%. None of the samples contained CaCO3 or other forms of inorganic C. A relative proportion of SOC derived from the bahiagrass, a C4 plant, or from the slash pine, a C3 plant, was estimated based on the following equations (Balesdent & Mariotti, 1996): % C4-derived SOC ¼ ðd À dT Þ=ðdG À dT Þ Â 100; % C3-derived SOC ¼ 100 À %C4-derived SOC;
13

difference between SP-A and SP-T locations. In general, whole soil d13C values in silvopasture were strongly C3dominated signals with value ranging from À17.5% to À25.1% at all sites (Table 1). The d13C values of whole soil had a maximum value of À15% at the surface (0– 5 cm) in OP, and a minimum value of À25.1% in the deeper soil profile of the silvopasture. Fractionated samples. The d13C values of the macro-sized fraction (Fig. 1) showed a similar trend as the whole soil sample with high values at the surface of the OP and decreasing with depth in all four sites. At the upper 5 cm, the d13C values in the OP were higher by 38% in Alachua (Fig. 1a, Po0.001), 33% in Suwannee (Fig. 1b, Po0.001), 27% in Hardee (Fig. 1c, Po0.001), and 25% in the Osceola sites (Fig. 1d, Po0.01) than the average value at the silvopasture (SP-A and SP-T). Except at the depth interval 30À50 cm, the OP d13C values in macrosized fraction were significantly higher than values on an adjacent silvopasture across Alachua, Suwannee, and Osceola sites (Fig. 1a, b, and d). The d13C values of the OP system in the micro-sized fraction (53–250 mm) were significantly higher than those at corresponding sampling locations in the silvopasture in the Ultisol sites at any given depth (Fig. 1e and f). Average values were À19.3% and À20.5% for OP in Alachua and Suwannee, respectively, compared with average values À24.6% and À24.7%, SP-T and SP-A, respectively, in Alachua and À25.5% and À25.6% for SP-A and SP-T, respectively, in Suwannee. Similarly, the d13C values of OP system at surface in the Spodosol sites were higher than those of locations in silvopasture. In the silt 1 clay-sized fraction, a clear difference between d13C values of OP and silvopasture was observed in Ultisol sites as opposed to Spodosol sites that showed no difference across depths (Fig. 1 g and h). In the upper 5 cm of the Alachua site, the values were À17.45% for OP as compared with À23.8% and À23.5% for SP-A and SP-T, respectively.

ð2Þ ð3Þ

where d is the d C of a given sample, dT a composite sample of the C3 plant and dG is a composite sample of pasture grass shoots and roots (C4). Based on the percentages of C3- and C4-derived SOC and the C content in each soil sample, the contributions to total SOC by C3 and C4 plants were calculated as follows: C3-derived SOC ðkg mÀ2 Þ ¼ð% C3-derived SOCÞ Â ðSOC content; kg mÀ2 Þ; ð4Þ C4-derived SOC ðkg mÀ2 Þ ¼ð% C4-derived SOCÞ Â ðSOC content; kg mÀ2 Þ: ð5Þ

Statistical analysis
Planned comparison analysis of variance (ANOVA) with Tukey’s studentized range test (HSD) was used to compare the mean differences between land-management practices and plant sources on SOC in whole soil, macro-sized, micro-sized and silt- and clay-sized fractions at four sites. Statistical analyses were carried out separately for all depth-classes. The composite sample in a row (grid line in OP) within each soil depth constituted the experimental unit of the analysis and each unit had three replications (rows). All statistical tests were performed with STATISTICAL ANALYSIS SYSTEM (SAS Institute Inc., 2004) and differences were considered significant when Po0.05.

Plant sources of SOC in whole soil sample
In all the sites, except at depth 5–15 cm in Suwannee and at depth 50–75 in Osceola, C3-derived SOC in the whole soil sample was significantly higher at the two silvopasture locations (SP-A and SP-T) than in the OP system. A reverse trend was observed for the SOC derived from C4 plants (Table 2), although the difference was not always significant for all depths. On average, the amount of SOC derived from C3 plants ranged from 9.4 kg mÀ2 in OP, 25.6 kg mÀ2 in SP-A, and 30.3 kg mÀ2 in SP-T at the surface soil, compared with corresponding values of 2.6, 4.1, and 6.4 kg mÀ2, respecr 2009 Blackwell Publishing Ltd, Global Change Biology, 16, 427–438

Results

Differences in the natural abundance of
13

13

C SOC

Whole soil sample. The average d C value of whole soil in OP for the four sites was À19.8%. This value was significantly higher (Po0.001) than in SP-A or SP-T of silvopasture with an overall average d13C value of À22.6% and À22.9%, respectively, but there was no

CONTRIBUTION OF TREES TO CARBON STORAGE IN SOILS

431

Table 1 d13C values of soil organic carbon (SOC) in whole (unfractionated) soil in different soil depths of silvopasture and open pasture sites on two Ultisols (Alachua and Suwannee counties) and two Spodosols (Hardee and Osceola counties) in Florida, USA d13C values of SOC, % Ultisol sites Depth (cm) OP Alachua À15.9a (0.9) À20.6 (1.0) À20.6a (0.3) À19.8a (0.3) À19.5a (0.6) À20.3a (0.9) SP-A À23.6b À23.0 À24.9b À25.0b À23.1b À23.3b SP-T À24.2b À24.6 À23.2b À23.2b À23.3b À23.2b Spodosol sites OP Hardee À15.2a (0.4) À19.1a (0.1) À21.0 (0.2) À22.2 (0.7) À23.7 (0.1) À23.3a (0.5) SP-A À20.9b À20.3ab À21.7 À21.1 À23.6 À24.7b SP-T À22.4c À21.7b À21.2 À21.8 À24.4 À24.7b

0–5 5–15 15–30 30–50 50–75 75–125

(0.5) (0.6) (1.7) (1.7) (0.5) (0.6)

(0.4) (1.1) (0.1) (0.3) (0.4) (0.5)

(0.0) (0.4) (0.8) (1.2) (0.7) (0.2)

(0.2) (0.4) (0.1) (0.4) (0.3) (0.1)

Suwannee 0–5 5–15 15–30 30–50 50–75 75–125 À16.9 À19.2a À20.9a À21.5a À22.0a À22.0a
a

Osceola À23.5 À23.6b À25.1b À25.0b À23.9b À24.3b
b

(0.3) (0.4) (0.3) (0.1) (0.3) (0.4)

(0.5) (0.1) (0.9) (1.0) (0.3) (0.2)

À23.9 À24.2b À24.4b À24.7b À24.7b À25.1b

b

(0.3) (0.3) (0.4) (0.4) (0.1) (0.1)

À15.0a À17.2a À19.5 À20.8 À20.9 À20.4a

(0.1) (0.5) (0.4) (0.1) (0.3) (0.7)

À17.5b À19.7b À20.5 À21.2 À21.5 À21.4b

(0.3) (0.4) (0.3) (0.2) (0.5) (0.3)

À20.7c À19.8b À20.0 À21.5 À21.6 À21.4b

(0.6) (0.2) (0.1) (0.5) (0.5) (0.8)

Lowercase letters indicate significant differences in SOC among pasture locations at a given depth and site. The values in parentheses are the standard error of means. OP, open pasture; SP-A, silvopasture, center of the alley; SP-T, silvopasture, in between trees in row.

tively, at the 75–125 cm depth. However, these C3 values accounted for 16.7% in OP, 55.5% in SP-A, and 65.1% in SP-T at the surface soil compared with the corresponding values of 58.9%, 69.4%, and 70.6%, respectively, at the 75–125 cm depth. The remaining percentages of SOC were considered to be contributions by C4 plants.

Plant sources of SOC in fraction sizes
The amounts of C3- and C4-derived SOC in the macrosized fraction were significantly different among the SPA, SP-T, and OP at the surface soil layers for Alachua and Osceola sites, but not for Suwannee and Hardee sites (Fig, 2a-d). Compared with the C4-derived SOC, the proportion of C3-derived SOC generally increased with soil depth in both land-use systems for all sites (Fig. 2). For example, at the Alachua site (Fig. 2a, e, i, m, q, and u) 11%, 68%, and 75% of SOC were C3-derived in OP, SP-A, and SP-T, respectively, on surface soil (0À5 cm). The corresponding values in the lowest profile (75–125 cm) were 63%, 93%, and 82% at the same site (Fig. 2a, e, i, m, q, and u). The C4-derived SOC in the micro-sized (53–250 mm) fraction was significantly higher in the OP surface soil
r 2009 Blackwell Publishing Ltd, Global Change Biology, 16, 427–438

(0–5) than in SP-A or SP-T for Alachua, Suwannee, and Osceola sites (Fig. 3a–d). The C3 component (slash pine) contributed o20% of the total SOC in this fraction size on surface soils (0À5 cm) at all four sites (Fig. 3a–d). The trend was similar to that of the larger fraction in that there were significant differences between OP and the silvopasture locations. The greatest difference between land uses in contribution of C3derived C was in the surface in all four sites, ranging from 13% to 19% in OP and 30% to 76% in the silvopasture (Fig. 3a–d). The SP-T generally had significantly higher C3-derived SOC than that of OP or SP-A, at the two lowest depths in all the sites except in Alachua (Fig. 3q–x). The silvopasture locations in general had accumulated higher C3- and C4-derived SOC in the silt 1 clay fractions than OP land use in the upper 15 cm depth of the Spodosol sites. Generally, with an increase in depth, proportion of C3-derived in SOC in the smallest fraction size increased compared the proportion of C4-derived SOC for all the sites. The C3-derived content in the o53 mm fraction was also relatively higher at the spodic horizon (upper level of spodic was between 15 and 50 cm) of the Spodosol sites (Fig. 4o and p).

432 S . G . H A I L E et al.

Fig. 1 Differences in d13C values of soil organic carbon (SOC) in macro-sized (250–2000 mm; a–d), micro-sized (53–250 mm; e–h), and silt 1 clay (o53 mm; i–l) fractions with mid-points of sampled depths at three pasture locations [silvopasture center of the alley (SP-A) and in-between tree rows (SP-T) and open pasture (OP)] for whole-soil of the Alachua (a, e, and i), Suwannee (b, f, and j), Hardee (c, g, and k), and Osceola (d, h, and l) sites in Florida, USA. Lower and upper case letters indicate significant differences in C3- and C4-derived SOC, respectively, among pasture locations at a given depth and site. Bars represent standard errors of the mean.

Discussion
The d13C value for composite sample of slash pine (dT) measured in this study was À28.5%. Values for individual plant parts such as needles, root, cones, bark, and branches (wood) were within Æ 2% of the dT. The values were within the range of d13C values reported in literature for a composite sample of plant parts or soil underneath slash pine. Parasolova et al. (2003) reported range between À25% and À29.5% for slash pine  Caribbean pine (Pinus caribaea) hybrids. Similarly, Mortazavi & Chanton (2002) found value range between À27.3% and À28.5% for slash pine needles. The d13C values range of terrestrial plants grown under natural conditions are between À9% to À19% for C4 plants and À20% to À35% for C3 plants (Vogel, 1993; Nyberg et al., 2000; Biedenbender et al., 2004; Staddon, 2004). The d13C value for bahiagrass measured in this study was also very close to the d13C value of À13.3% for shoots or roots of bahiagrass reported by Nakano et al. (2001). The d13C value of bahiagrass shoot and root samples as well

as their composite sample (dG) in this study were not different, with a value of À12.9%. Although the d13C values measured in the silvopasture soil in this study (on average  À24%) were within the general range of values for terrestrial C3 plants as given in literatures, they were higher than d13C value (À28.5%) for a composite slash pine sample. It is plausible that the increase in d13C values was caused by SOC input from bahiagrass (C4 plants) in the silvopasture. The main source of C input to the soil is usually root biomass, which is transformed to SOC by microorganisms. Compared with trees, grass species generally develop shallow root systems or alternatively allocate the main root biomass to the uppermost soil layers, even if single roots can reach depths of several meters. The most negative d13C values ( À25%) were found in deep soils of the silvopasture soils and the least negative values ( À15%) at the surface of soil layer of OP. The d13C values showed a generally decreasing (become more negative) trend with soil depth. The deep-rooted
r 2009 Blackwell Publishing Ltd, Global Change Biology, 16, 427–438

Table 2 C3- and C4-derived soil organic carbon (SOC) in whole (unfractionated) soil in different land-use locations at four sites across soil depth on two Ultisols (Alachua and Suwannee counties) and two Spodosols (Hardee and Osceola counties) in Florida, USA

SOC (kg mÀ2) Spodosol sites C4 SP-T OP SP-A SP-T OP SP-A SP-T C3 C4 OP SP-A SP-T

Ultisol sites

C3

r 2009 Blackwell Publishing Ltd, Global Change Biology, 16, 427–438

Depth (cm) OP

SP-A

0–5 5–15 15–30 30–50 50–75 75–125

Alachua Hardee 11.78b (1.97) 34.25a (1.26) 30.85a (0.67) 43.63A (0.29) 21.32B (0.46) 10.67C (0.37) 6.89c (0.29) 16.49b 4.42b (0.63) 12.17a (0.07) 11.96a (2.22) 7.06A (0.04) 7.80A (0.07) 2.82B (0.51) 5.54c (1.47) 44.77a 3.43c (0.10) 10.00a (0.01) 4.62b (0.05) 3.59A (0.24) 4.26A (0.15) 2.16B (0.24) 6.16c (1.59) 25.23a 1.58 (0.30) 2.32 (0.15) 1.92 (0.17) 2.33A (0.16) 1.17B (0.11) 1.01B (0.07) 11.68c (0.41) 13.4b 2.20 (0.20) 2.53 (0.01) 2.05 (0.15) 2.37A (0.03) 1.13B (0.08) 1.04B (0.04) 4.15c (0.23) 13.75a 0.91b (0.05) 1.64a (0.10) 0.75b (0.04) 0.82A (0.04) 0.73A (0.06) 0.35B (0.02) 2.98c (0.61) 12.68a (0.60) (3.18) (0.34) (0.35) (0.12) (1.18)

23.65a 24.2b 15.76b 19.37a 11.08b 7.26b

(0.01) 35.08A (0.79) 15.81B (0.57) (0.88) 8.55C (2.25) 48.34A (0.62) (0.58) 4.21C (0.28) 27.02A (0.24) (0.33) 7.96B (0.42) 9.15B (0.49) (0.71) 2.52B (0.18) 9.51A (0.83) (0.99) 1.46C (0.10) 4.12A (0.18)

17.29B 18.97B 14.21B 14.85A 3.50B 2.37B

(1.15) (1.70) (0.67) (1.56) (0.76) (0.27)

Suwannee
a A B B

Osceola

0–5 5–15 15–30 30–50 50–75 75–125

8.01 9.19a 4.38c 3.57b 3.80b 2.56b

c

(1.20) 25.73 (0.88) 32.28 (0.17) 18.83 (1.11) 13.98 (0.66) 14.14 (0.94) 10.93c (0.38) 26.13b (0.73) 35.23a (0.83) 74.23A (2.38) 57.70B (1.95) 42.79C (1.74) (0.55) 6.98b (0.34) 7.17b (0.37) 12.45A (0.29) 3.28B (0.25) 3.76B (0.39) 7.40b (0.70) 9.64ab (0.56) 11.84a (0.11) 18.98A (0.35) 13.64B (0.03) 14.47B (0.47) (0.19) 6.49b (0.00) 8.04a (0.02) 5.18A (0.04) 2.01C (0.38) 3.82B (0.32) 6.35b (0.70) 8.40b (0.68) 12.33a (0.82) 8.74AB (0.54) 7.52B (0.91) 11.33A (0.74) (0.59) 2.71b (0.14) 6.92a (0.00) 3.40A (0.28) 1.13C (0.05) 1.98B (0.10) 5.66b (0.14) 6.09b (0.02) 9.73a (0.13) 5.67B (0.21) 6.10B (0.63) 8.86A (0.66) b a B C A a b a (0.13) 2.79 (0.03) 13.08 (1.66) 2.78 (0.16) 1.23 (0.10) 4.64 (0.06) 5.33 (0.83) 1.09 (0.46) 5.09 (0.08) 6.40A (0.62) 0.96B (0.41) 5.41A (0.17) c a (0.10) 1.03 (0.08) 11.53 (0.36) 1.83B (0.12) 0.38C (0.01) 3.31A (0.07) 4.11b (0.46) 1.06c (0.29) 6.03a (0.27) 4.50A (0.47) 1.22B (0.05) 5.77A (0.82)

b

CONTRIBUTION OF TREES TO CARBON STORAGE IN SOILS 433

Lower and upper case letters indicate significant differences in C3- and C4-derived SOC, respectively, among pasture locations at a given depth and site. The values in parentheses are the standard error of means. OP, open pasture; SP-A, silvopasture, center of the alley; SP-T, silvopasture, in between trees in row.

434 S . G . H A I L E et al.

Fig. 2 Differences in C3- and C4-derived soil organic carbon (SOC) in macro-sized (250–2000 mm), fractions with sampled depth intervals at three pasture locations [silvopasture center of the alley (SP-A) and in-between tree rows (SP-T) and open pasture (OP)] for the Alachua (a, e, i, m, q, and u), Suwannee (b, f, j, n, r, and v), Hardee (c, g, k, o, s, and w), and Osceola (d, h, l, p, t, and x) sites in Florida, USA. Lower and upper case letters indicate significant differences in C3- and C4-derived SOC, respectively, among pasture locations at a given site. Bars represent standard errors of the mean.

system of the tree component is most likely to increase the C3-signature or decrease (more negative) d13C values with depth as reported by Kramer & Gleixner (2008). By contrast, the least negative d13C values in the surface soil layer (0–5 cm) of the OP suggests that bahiagrass was the main plant source of the SOC. The SOC beneath the silvopasture on the SP-T and SP-A locations, including in the surface soils, were substantially derived from the slash pine litter and fine roots (Table 2). Soil C decomposition is relatively fast in the surface soil layer (Yang et al., 2009). Consequently, the SOC residence time will be shorter at surface soils than at lower depths. Evidence from radiocarbon studies also indicate that the average age of SOC increases with depth in the profile (Scharpenseel & Neue, 1984; Balesdent et al., 1990). The soil C at the surface is likely composed of recent accumulation of SOM at a given site.

The d13C values combined with the possible variations in C turnover rate of the different soil fraction sizes provide an indication of plant community history. The d13C values of fraction sizes revealed that the macro-sized fraction (Fig. 2) and micro-sized fraction (Fig. 3) contributed most to the C derived from the C3 woody plant in the silvopasture compared with that in OP across the soil depth, but particularly so in the lower profiles (Figs 2o, p, s, w and 3m–x). Given that the soil of the study sites is high in sand (94%, Haile et al., 2008), soil C in the larger soil fraction is likely to be the C that is newly incorporated into the soil. Thus, most of this SOC, including that in the lower profile, could be considered as contributed by the current C3 plant component in the silvopasture, i.e. slash pine, which has deeper root systems than the grass component. When comparing d13C values of functionally distinct SOC fractions, the silt 1 clay fraction showed a higher
r 2009 Blackwell Publishing Ltd, Global Change Biology, 16, 427–438

CONTRIBUTION OF TREES TO CARBON STORAGE IN SOILS

435

Fig. 3 Differences in C3- and C4-derived soil organic carbon (SOC) in micro-sized (53–250 mm), fractions with sampled depth intervals at three pasture locations [silvopasture center of the alley (SP-A) and in-between tree rows (SP-T) and open pasture (OP)] for the Alachua (a, e, i, m, q, and u), Suwannee (b, f, j, n, r, and v), Hardee (c, g, k, o, s, and w), and Osceola (d, h, l, p, t, and x) sites in Florida, USA. Lower and upper case letters indicate significant differences in C3- and C4-derived SOC, respectively, among pasture locations at a given site. Bars represent standard errors of the mean.

proportion of C3-derived SOC than C4-derived SOC at the lower soil profile (Fig. 4n–x). However, in the silt 1 clay fraction, no difference was observed between silvopasture and OP in the lower depths for most sites (Fig. 4v–x), except the Alachua site where the land was under agriculture (corn, Zea mays) before the current OP (Fig. 4u). The silt 1 clay fraction was less enriched with 13 C than the whole soil and appeared to contain SOC largely derived from C3 (o75%). In fact, the C associated with the silt 1 clay fraction tends to be ‘older’ C (Anderson & Paul, 1984), indicating that prior land-use history of the site has a substantial effect on the current status of C storage in soil fraction sizes. Forty-five years ago, the current study sites were under forest vegetation, popularly called Florida Flatwoods and it is logical to surmise that the current C status of the site is a reflection of that land-use history. At the Spodosol sites, at and below the spodic horizon (Bh), which occurred between 15 and 50 cm depth,
r 2009 Blackwell Publishing Ltd, Global Change Biology, 16, 427–438

the C3-derived C in silt- and clay-sized fraction was relatively high for both the pasture and the silvopasture (Fig. 4k, l, o, t). The spodic horizons differ from other soil profile materials due to the prevalence of organically associated aluminum (Al) and have very high surface area for the retention of SOC (Eswaran et al., 2003). Percival et al. (2000) found that the pyrophosphate-extractable Al was correlated strongly with both soil C content and soil C concentration across different types of soils. Schwartz et al. (1986) showed from an analysis of the depth distribution of d13C in podzols (equivalent name for Spodosols, according to the FAO system of soil classification) developed under C4 savanna in the Congo that the Bh horizons exhibited typical C3 values because the podzolic morphology was inherited from past forest phases 430 000 years old. Using long-term incubation experiments and natural 13C-labelling of C3- and C4-derived SOC collected from across major environmental gradients in Australia,

436 S . G . H A I L E et al.

Fig. 4 Differences in C3- and C4-derived soil organic carbon (SOC) in silt + clay-sized (o53 mm), fractions with sampled depth intervals at three pasture locations [silvopasture center of the alley (SP-A) and in-between tree rows (SP-T) and open pasture (OP)] for the Alachua (a, e, i, m, q, and u), Suwannee (b, f, j, n, r, and v), Hardee (c, g, k, o, s, and w), and Osceola (d, h, l, p, t, and x) sites in Florida, USA. Lower and upper case letters indicate significant differences in C3- and C4-derived SOC, respectively, among pasture locations at a given site. Bars represent standard errors of the mean.

Wynn & Bird (2007) found that the active pool of SOC derived from C4 plants decomposed at more than twice the rate of total pool of active SOC, and concluded that C4-derived SOC decomposed faster than C3-derived SOC in mixed C3/C4 soils. This would suggest that the higher amount of C3-derived SOC at lower depths in silvopastoral systems could also be a result of lower decomposition rate of C3-derived SOC and not only of its higher input. The silvopasture system in the present study, which represents a change from a C4-dominated herbaceous community (bahiagrass) to a mixed plant community of C3 woody slash pine and bahiagrass, provides an ideal situation for the use of the natural 13C-labeling technique (Nitschelm et al., 1997). The quantification of the progressive incorporation of new C into soil organic fractions provides a powerful means to elucidating the pathways of C transformations and stabilizations. However, the 13C method is unable to distinguish between residual ‘primary forest’ C and new C derived from tree

component in the silvopasture (Gillson et al., 2004). The use of 14C and bomb C models (Trumbore, 2000) is relevant to ‘date’ soil fractions and to distinguish newly incorporated C derived from tree component in the silvopasture from old residual C retained from primary forest. Such methods could also reduce uncertainties in turnover rates.

Conclusions
The study showed a consistent trend of increase in C3signature dominated the d13C values with soil depth. Evidently, this has translated to a slight increase in C3derived SOC pools following tree integration into pastures in this study. The fractionated soil revealed that a substantial proportion of C3-derived SOC was found in macro- and micro-sized fractions at the surface and deeper soil profiles. At all sites, we found accumulations of larger proportion of C3-derived than C4-derived SOC associated with silt 1 clay, particularly in the
r 2009 Blackwell Publishing Ltd, Global Change Biology, 16, 427–438

CONTRIBUTION OF TREES TO CARBON STORAGE IN SOILS
deeper soil layer. These results suggest that, in the long term, silvopasture may help sequester more SOC in the soil. Obviously, such long-term sequestration of C in the soil has promising environmental implications. However, more information on several key issues is needed before the suggested potential of silvopasture can be convincingly established.

437

Acknowledgements
This research was supported in part by a grant from the USDAInitiative for Future Agriculture and Food Systems (IFAFS) through the Center for Subtropical Agroforestry. We thank all the individuals who participated in this research project especially, Fred Clark (Alachua), Bruce Goff (Suwannee), and Harris Hill (Osceola) who allowed us to collect soil samples and other field data from their farms. We also thank Sam Allen, Alain Michel and Subhrajit Saha (SFRC, UF) for their help during soil sampling.

References
Accoe F, Boeckx P, Cleemput OV, Hofman G, Xu H, Huang B, Chen G (2002) Characterization of soil organic matter fractions from grassland and cultivated soils via C content and d13C signature. Rapid Communication in Mass Spectrometry, 16, 2157– 2164. Anderson DW, Paul EA (1984) Organo-mineral complexes and their study by radiocarbon dating. Soil Science Society of America Journal, 48, 298–301. Archer S, Boutton TW, Hibbard KA (2001) Trees in grasslands: biogeochemical consequences of woody plant expansion. In: Global Biogeochemical Cycles in the Climate System (eds Schulze E-D, Heimann M, Harrison S, Holland E, Lloyd J, Prentice I, Schimel D), pp. 115–137. Academic Press, San Diego, CA, USA. Balesdent J, Mariotti A (1996) Measurement of soil organic matter turnover using 13C natural abundance. In: Mass Spectrometry of Soils (eds Boutton TW, Yamasaki SI), pp. 83–111. Marcel Dekker, New York, NY, USA. Balesdent J, Mariotti A, Biosgontier D (1990) Effect of tillage on soil organic carbon mineralization estimated from 13C abundance in maize fields. Journal of Soil Science, 41, 587–596. Balesdent J, Wagner GH, Mariotti A (1988) Soil organic matter turnover in long-term field experiments as revealed by carbon13 natural abundance. Soil Science Society of America Journal, 52, 118–124. Bernoux M, Cerri CC, Neill C, De Moraes JFL (1998) The use of stable carbon isotopes for estimating soil organic matter turnover rates. Geoderma, 82, 43–58. Biedenbender SH, McClaran MP, Quade J, Weltz MA (2004) Landscape patterns of vegetation change indicated by soil carbon isotope composition. Geoderma, 119, 69–83. Carter MR (1996) Analysis of soil organic matter in agroecosystems. In: Structure and Organic Matter Storage in Agricultural

Soil (eds Carter MR, Stewart BA), pp. 3–11. CRC press, Boca Raton, FL, USA. Christensen BT (1992) Physical fractionation of soil and organic matter in primary particle size and density separates. Advanced Soil Science, 20, 1–90. Connin SL, Virginia RA, Chamberlain CP (1997) Carbon isotopes reveal soil organic matter dynamics following arid land shrub expansion. Oecologia, 110, 374–386. Coplen TB (1996) More uncertainty than necessary. Paleoceanography, 11, 369–370. Dawson TE, Mambelli S, Plamboeck AH, Templer PH, Tu KP (2002) Stable isotopes in plant ecology. Annual Review Ecology System, 33, 507–559. Del Galdo I, Six J, Peressotti A, Cotrufo MF (2003) Assessing the impact of land-use change on soil C sequestration in agricultural soils by means of organic matter fraction and stable C isotopes. Global Change Biology, 9, 1204–1213. Ehleringer JR, Buchmann N, Flanagan LB (2000) Carbon isotope ratios in belowground carbon cycle processes. Ecological Applications, 10, 412–422. Elliott ET (1986) Aggregate structure and carbon, nitrogen and phosphorus in native and cultivated soils. Soil Science Society of America Journal, 50, 627–633. Eswaran H, Ahrens R, Rice TJ, Stewart BA (2003) Soil Classification: A Global Desk Reference. CRC Press, Boca Raton, FL, USA. Garrett HE, Rietveld WJ, Fisher RF), eds (2000) North American Agroforestry: An Integrated Science and Practice. American Society of Agronomy, Madison, WI. Gill RA, Burke IC (1999) Ecosystem consequences of plant life form changes at three sites in the semiarid United States. Oecologia, 121, 551–563. Gillson L, Waldron S, Willis KJ (2004) Interpretation of soil d13C as an indicator of vegetation change in African savannas. Journal of Vegetation Science, 15, 339–350. Haile SG, Nair PKR, Nair VD (2008) Carbon storage of different soil-size fractions in Florida silvopastoral systems. Journal of Environmental Quality, 37, 1789–1797. IPCC (2001) Third Assessment Report: Climate Change 2001. Cambridge University Press, Cambridge. Jackson RB, Schenk HJ, Jobbagy EG et al. (2000) Belowground consequences of vegetation change and their treatment in models. Ecological Applications, 10, 470–483. ´ gy EG, Jackson RB (2000) The vertical distribution of soil Jobba organic carbon and its relation to climate and vegetation. Ecological Applications, 10, 423–436. Kramer C, Gleixner G (2008) Soil organic matter in soil depth profiles: distinct carbon preferences of microbial groups during carbon transformation. Soil Biology and Biochemistry, 40, 425–433. Martens DA (2000) Plant residue biochemistry regulates soil carbon cycling and carbon sequestration. Soil Biology and Biochemistry, 32, 361–369. Matson PA, Parton WJ, Power AG, Swift MJ (1997) Agricultural intensification and ecosystem properties. Science, 277, 504–509. McClaran MP, McPherson GR (1995) Can soil organic carbon isotopes be used to describe grass–tree dynamics at a savanna-

r 2009 Blackwell Publishing Ltd, Global Change Biology, 16, 427–438

438 S . G . H A I L E et al.
grassland ecotone and within the savanna? Journal of Vegetation Science, 6, 857–862. Montagnini F, Nair PKR (2004) Carbon sequestration: an underexploited environmental benefit of agroforestry systems. Agroforestry Systems, 61 & 62, 281–298. Mortazavi B, Chanton J (2002) Carbon isotopic discrimination and control of nighttime canopy d18O–CO2 in a pine forest in the Southeastern US. Global Biogeochemical Cycles, 16, 1–13. Nair PKR (2005) Agroforestry: trees in support of sustainable agriculture. In: Encyclopedia of Soils in the Environment, Vo1. 1 (eds Hillel H, Hatfield JL, Powlson DS, Rosenzweig C, Scow KM, Singer MJ, Sparks DL), pp. 35–44. Elsevier, London, UK. Nair PKR, Bannister ME, Nair VD, Alavalapati JRR, Ellis EA, Jose S, Long AJ (2005) Silvopasture in southeastern United States: more than just a new name for an old practice. In: Silvopastoralism and Sustainable Land Management (eds Mos´guez A), pp. quera-Losada MR, McAdam J, Rigueiro-Rodrı 72–82. CABI, Wallingford, UK. Nair PKR, Gordon AM, Mosquera-Losada MR (2008) Agroforestry. In: Encyclopedia of Ecology, Vol. 1 (eds Jorgensen SE, Fath BD), pp. 101–110. Elsevier, Oxford, UK. Nakano A, Takahashi K, Kimura M (2001) Effect of host shoot clipping on carbon and nitrogen sources for arbuscular mycorrhizal fungi. Mycorrhiza, 10, 287–293. Nitschelm JJ, Scher AL, Hartwig UA, Kessel CV (1997) Using stable isotopes to determine soil carbon input differences under ambient and elevated atmospheric CO2 conditions. Global Change Biology, 3, 411–416. ˇ gber P (2000) Respiration Nyberg G, Ekblad A, Buresh RJ, Ho from C3 plant green manure to a C4 plant carbon dominated soil. Plant and Soil, 218, 83–89. Parasolova NV, Xu ZH, Lundkvist K, Farquar GD, Dieters MJ, Walker S, Saffigna PG (2003) Genetic variation in foliar carbon isotope composition in relation to tree growth and foliar nitrogen concentration in clones of the F1 hybrid between slash pine and Caribbean pine. Forest Ecology and Management, 172, 145–160. Paul EA, Paustian K, Elliott ET, Cole CV (1997) Soil Organic matter in Temperate Ecosystems. CRC Press, New York. Percival HJ, Parfitt RL, Scott NA (2000) Factors controlling soil carbon levels in New Zealand grasslands: is clay content important? Soil Science Society of America Journal, 64, 1623–1630. Saha SK, Nair PKR, Nair VD, Kumar BM (2009) Soil carbon stock in relation to plant diversity of homegardens in Kerala, India. Agroforestry Systems, 76, 53–65. Scharpenseel HW, Neue HU (1984). Use of isotopes in studying the dynamics of soil organic matter in soils. In: Papers from the International Conference on Organic Matter and Rice, (eds ˜ os, Stephen Banta, Corazon V. Mendoza) held at IRRI, Los Ban Laguna, Philippines, 27 September to 1 October 1982, pp. 273– 310. International Rice Research Institute, IRRI publ., Manila, Philippines. Scholes RJ, Hall DO (1996) The carbon budget of tropical savannas, woodlands and grasslands. In: Global change: Effects on Coniferous Forests and Grasslands (eds Breymeyer AI, Hall ˚ gren GI), pp. 69–100. John Wiley & Sons Ltd., DO, Melillo JR, A Chichester, UK. Schwartz D, Mariotti A, Lanfranchi R, Guillet B (1986) 13C/12C ratios of soil organic matter as indicators of vegetation changes in the Congo. Geoderma, 39, 97–103. Six J, Elliott ET, Paustian K, Doran JW (1998) Aggregation and soil organic matter accumulation in cultivated and native grassland soils. Soil Science Society of America Journal, 62, 1367–1377. Staddon PL (2004) Carbon isotopes in functional soil ecology. Trends in ecology and evolution, 19, 148–153. Steinbeiss S, Beler H, Engels C et al. (2008) Plant diversity positively affects short-term soil carbon storage in experimental grasslands. Global Change Biology, 14, 2937–2949. Stout JD, Goh KM, Rafter TA (1981) Chemistry and turnover of naturally occurring resistant organic compounds in soil. In: Soil Biochemistry, Vol. 5 (eds Paul EA, Ladd JN), pp. 1–73. Marcel Dekker, New York, NY. Swap RJ, Aranibar JN, Dowty PR, Gilhooly WP III, Macko SA (2004) Natural abundance of 13C and 15N in C3 and C4 vegetation of southern Africa: patterns and implications. Global Change Biology, 10, 350–358. Takimoto A, Nair VD, Nair PKR (2008) Contribution of trees to soil carbon sequestration under agroforestry systems in the West African Sahel. Agroforestry Systems, 76, 11–25. Tiessen H, Stewart JWB (1983) Particle-size fractions and their use in studies of soil organic Matter: II. Cultivation effects on organic matter composition in size fractions. Soil Science Society of America Journal, 47, 509–514. Trumbore S (2000) Age of soil organic matter and soil respiration: radiocarbon constraints on belowground C dynamics. Ecological Applications, 10, 399–411. Vogel JC (1993) Variability of carbon isotope fractionation during photosynthesis. In: Stable Isotopes and Plant Carbon–Water Relations (eds Ehleringer JR, Hall AE, Farquhar GD, Saugier B), pp. 29–46. Academic Press Inc., San Diego, CA, USA. Workman SW, Bannister ME, Nair PKR (2003) Agroforestry potential in the southeastern United States: perceptions of landowners and extension professionals. Agroforestry Systems, 59, 73–83. Wynn JG, Bird MI (2007) C4-derived soil organic carbon decomposes faster than its C3 counterpart in mixed C3/C4 soils. Global Change Biology, 13, 2206–2217. Yang Y, Fang J, Smith P et al. (2009) Changes in topsoil carbon stock in the Tibetan grasslands between the 1980s and 2004. Global Change Biology, doi: 10.1111/j.1365-2486.2009.01924.x.

r 2009 Blackwell Publishing Ltd, Global Change Biology, 16, 427–438

Sign up to vote on this title
UsefulNot useful