SSSAJ Grow Your Career with SSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Echeverría, M. E.
Right arrow Articles by Hendrick, R. L.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Echeverría, M. E.
Right arrow Articles by Hendrick, R. L.
Agricola
Right arrow Articles by Echeverría, M. E.
Right arrow Articles by Hendrick, R. L.
Related Collections
Right arrow Soil Organic Matter
Right arrow Carbon Sequestration
Right arrow Forest Soils
Published in Soil Sci. Soc. Am. J. 68:950-958 (2004).
© 2004 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA

DIVISION S-7—FOREST & RANGE SOILS

Soil Organic Matter Fractions under Managed Pine Plantations of the Southeastern USA

Marietta E. Echeverría, Daniel Markewitz*, Lawrence A. Morris and Ronald L. Hendrick

Daniel B. Warnell School of Forest Resources, Univ. of Georgia, D.W. Brooks Dr., Athens, GA 30602-2153

* Corresponding author (dmarke{at}forestry.uga.edu).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Herbicide application in plantation forests may affect soil productive capacity through its effect on the quantity and characteristics of soil C. We examined changes in surface soil (<10 cm) soil organic matter (SOM) fractions in 12- to 18-yr-old pine stands from the Lower Coastal Plain and Piedmont of the southeastern USA that were receiving complete competition control treatments (n = 13 paired plots). Light (LF) (<1.6 g cm–3), medium (MF) (1.6–2.0 g cm–3), and heavy (HF) fraction (2.0 g cm–3) SOM were isolated by density separation and the HF was hydrolyzed isolating a hydrolyzable (H-HF) and residual fraction (R-HF). Herbicide treated surface soils had lower whole soil C (12.8 g kg–1) and N (0.51 g kg–1) compared with untreated controls, 16.1 g C kg–1 and 0.63 g N kg–1. Across all sites, the greatest decreases in soil C and N occurred in the LF and MF fractions. The majority of C and N in Lower Coastal Plain surface soils (~90% sand) is partitioned in the LF + MF, while in the Piedmont soils (~60% sand) it is in the H-HF. Decreases in these SOM fractions were only slightly greater than decreases in whole soil SOM. Additionally, there was a significant decrease of 27, 41, and 31% in net N mineralization (7-d anaerobic incubation) due to treatment in a Piedmont site for whole soil (WS), WS + L/MF, and WS + HF, respectively. For the Lower Coastal Plain site, there was an 18% decrease in the WS + L/MF mixture only. Complete competition control in both the Piedmont and Coastal Plain clearly decreased SOM quantity. Decreases in SOM quality as indicated by decreased net N mineralization potential were also evident. The results suggest herbicide treatments may decrease the productive capacity of surface soils.

Abbreviations: HF, heavy fraction • H-HF, hydrolyzable heavy fraction • LF, light fraction, MF, medium fraction • R-HF, residual heavy fraction • RPM, revolutions per minute • SOM, soil organic matter • SPT, sodium polytungstate • WS, whole soil


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
SOIL ORGANIC MATTER influences soil chemical and physical properties that control nutrient cycling and consequently have a major effect on forest productivity. Due to the important role of SOM in nutrient cycling, there has long been an interest in understanding how forest soil management affects soil C pools (Switzer and Nelson, 1972; Johnson and Curtis, 2001). Currently, there has also been an added interest in the role of SOM as a potential sink for atmospheric CO2 (Schlesinger, 1990; Post and Kwon, 2000).

Intensive pine plantation management, such as practiced in the southeastern USA and elsewhere, often incorporates herbicide treatment to greatly reduce plant competition in newly established stands. A number of studies suggest that competition control through herbicide treatments reduce soil C and N content in forest ecosystems (Aust and Lea, 1991; Polglase et al., 1992; Carlyle, 1993; Munson et al., 1993; Busse et al., 1996). For example, Aust and Lea (1991) reported significant decreases (29%) in SOM content after 2 yr of postharvest competition control in a study examining ecosystem recovery in a water tupelo (Nyssa aquatica L.)–bald cypress [Taxodium distichum (L.) Rich. var. distichum] wetland in southwest Alabama. Similarly, in a long-term study examining the role of understory vegetation in ponderosa pine (Pinus ponderosa Dougl. ex Laws.) stands, Busse et al. (1996) reported a 33 and 30% decrease in surface soil (0–4 cm) C and N content, respectively, due to competition control. Possible explanations for these observed decreases include enhanced decomposition due to elevated soil temperature and/or increased soil moisture resulting from the absence of vegetative ground cover during early stand development (Aust and Lea, 1991; Munson et al., 1993), reduced organic matter inputs from above- and/or belowground litter (Busse et al., 1996), and altered SOM quality resulting from changes in organic matter inputs (Piatek and Allen, 2001).

None of these previous studies has assessed the impacts of herbicide treatments on soil quality; rather, they focused on bulk measures of SOM. Soil organic matter, however, is recognized to consist of various fractions varying in degree of decomposition, recalcitrance, and turnover rate (Schimel et al., 1985) and management practices may affect these fractions differently. The use of physical fractionation of SOM by density separation has been demonstrated to relate to turnover rate (Trumbore and Zheng, 1996) and is thought to relate directly to the structure and function of SOM in situ (Christensen, 1992). Density fractionation generally separates mineral-free, partly decomposed plant debris (LF) from organic matter adsorbed on mineral surfaces and within aggregates (HF) (Trumbore et al., 1989; Golchin et al., 1994). Light fraction SOM is characterized by elevated C and N concentration and wider C/N ratios relative to whole soil and in a few cases has been shown to be more sensitive to management practices than total SOM (Strickland and Sollins, 1987; Cambardella and Elliot, 1992; Janzen et al., 1992; Hassink, 1995; Barrios et al., 1996). Utilizing various organic matter fractions for analysis may be useful in identifying the long-term impacts of competition control on SOM pools, nutrient cycling, and site productivity.

The specific objectives of this study were to: (i) quantify the impacts of competition control on various soil C and N pools including WS, LF, MF, H-HF, and R-HF; (ii) evaluate soil C and N partitioning across SOM fractions for soils of varying texture; (iii) evaluate effects of herbicide on net N mineralization potential, and (iv) determine the contribution of light and heavy SOM fractions to mineralizable N.

We hypothesized that the lighter more labile SOM fractions would be more sensitive indicators than WS SOM to ecosystem changes resulting from competition control. Additionally, we hypothesized that fine-textured soils in the Georgia Piedmont would tend to have more protected SOM in residual fractions and would therefore be somewhat buffered to the impacts of competition control on SOM compared with sandier Lower Coastal Plain soils.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Study Areas and Sampling
We utilized long-term loblolly pine (Pinus taeda L.) study sites established on representative soils in the Piedmont and Lower Coastal Plain of Georgia as part of the Consortium for Accelerated Pine Productivity Studies (CAPPS) project (Borders and Bailey, 2001). Site locations consisted of the Dixon State Forest near Waycross (31° 15' N, 82° 24' W), the B.F. Grant Forest near Eatonton (33° 20' N, 83° 23' W), and Whitehall Forest in Athens (33° 57' N, 83° 19' W). Soils in the Waycross sites consist mainly of Paleudults and Haplohumods, and the soils in the Eatonton and Athens sites consist mainly of Hapludults and Kanhapludults (Table 1). Additionally, we utilized slash pine plots established in the northern Florida Lower Coastal Plain as part of the Plantation Management Research Cooperative (PMRC) site preparation study (Shiver et al., 1990). The three sites at the northern Florida location (approximately 30° 38' N, 81° 36' W) are situated on somewhat poorly drained Haplorthods in Callahan, Lee, and Yulee, FL (Table 1). Each site has replicated sets of treatments that include a control (C), herbicide for complete competition control (H), fertilization with NP or NPK two or more times (F), and herbicide plus fertilization (HF). For this research we utilized only soils from the C and H treatments. Herbicide treatments at the PMRC sites consisted of the use of glyphosphate before site preparation and subsequent application of glyphosphate, triclopyr, and, occasionally, diesel fuel to eliminate all competing vegetation (Shiver et al., 1990). At the CAPPS sites, herbicide treatments consisted of a broadcast of sulfomethuron methyl during early spring of the first three growing seasons following planting and subsequent application of glyphosphate as needed in mid-summer each year thereafter (Borders et al., 2002). These treatments virtually eliminated competing species within the treated plots.


View this table:
[in this window]
[in a new window]
  		Table 1. Regions, location, soil types, soil texture, pine species, and stand age at time of sampling for all sites.

 
Surficial mineral soil samples (0–10 cm or genetic A horizons in northern Florida sites) were collected from plots under control and herbicide treatments. A consistent effort was made to clear away all O-horizon material to expose bare mineral soil before sampling. Composited samples of no <10 individual points per plot were collected simultaneously for all plots within a location but locations were sampled at different times. North Florida plots were sampled in 1998 at Age 18 by Shan et al. (2001), Waycross was sampled in Dec. 1998/Jan. 1999 (Age 12), Eatonton in Dec. 1999/Jan. 2000 (Age 12), and Athens in 2001 (Age 12). Soil bulk density was estimated in each location using a 7.5-cm diameter ring sampler following Blake (1965). North Florida was sampled in 1998 using eight cores per plot, Waycross and Eatonton were sample in 2001 using replicate samples in each plot, and Athens was sampled in 2003 using six samples per plot. At the time of sampling, all stands were of mid-rotation, vigorously growing, and had a closed canopy.

Within the CAPPS project, the Eatonton sampling location consisted of four complete blocks yielding four replicated plot pairs (a pair consisting of a control and herbicide treated stand), the Athens sampling location consisted of two complete blocks, and the Waycross location had four complete blocks. The North Florida location consisted of three sites, Callahan, Yulee, and Leon, which were treated as blocks yielding three plot pairs. Overall, thirteen paired plots from control and herbicide stands were sampled, seven from the Lower Coastal Plain and six from the Piedmont (Table 1).

Samples were air-dried and sieved through a 2-mm screen before any analysis and a subsample was ball-mill ground to a fine powder in a 8000D Mixer/Mill (SPEX CertiPrep, Metuchen, NJ) for C and N determination by dry combustion with a NC Soil 2100 autoanalyzer (CE Elantech, Newark, NJ).

Soil Organic Matter Fractionation
Soil organic matter from control and herbicide plots was fractionated following the methods of Trumbore et al. (1989) and Golchin et al. (1994). Soil samples were separated into three density fractions using sodium polytungstate (SPT) without prior sonification. A test trial with sonification demonstrated only small increases in the recovery of LF C for Piedmont (13.5 ± 8.5%, mean ± 1SD) and Coastal Plain (11.0 ± 1.7%) soils and thus was abandoned (D. Meason and D. Markewitz, unpublished data, 2003). Density fractions included a LF (<1.6 g cm–3), a MF (1.6–2.0 g cm–3), and a HF fraction (>2.0 g cm–3). The HF was further fractionated by acid hydrolysis into a H-HF and a R-HF.

For the fractionation, approximately 5 g of soil was placed in a 50-mL centrifuge tube with approximately 25 mL SPT ({rho} = 1.6 g cm–3). The centrifuge tubes containing soil-SPT mixture were shaken on a reciprocating shaker at 300 revolutions per minute (RPM) for 15 min. Particles that adhered to the tube walls were washed back into suspension with SPT. The suspensions were allowed to stand for 30 min before centrifugation to prevent mechanical occlusion of LF particles. Suspensions were then centrifuged at 2000 x g for 60 min. The supernatant with floating particles was decanted onto a Millipore AP20 glass-fiber filter (Millipore, Billerica, MA) and filtered under vacuum. Particles that had adhered to the walls of the tubes were scraped off with a spatula and washed onto filter paper. All material collected on filter paper was washed with at least 500 mL of deionized H2O. The material was then washed off the filter paper and the LF was dried for 18 h at 60°C and then weighed.

Centrifuge tubes containing residual soil were brought to 25 mL with SPT ({rho} = 2.0 g cm–3) and shaken on a reciprocating shaker at 300 RPM for 15 min. Particles that adhered to tube walls were washed back into suspension with SPT. Samples were allowed to stand for 30 min before centrifugation for 60 min at 2000 x g. The MF was collected, dried and weighed in the same manner as the LF.

The residual soil was washed at least three times with deionized H2O. Ten milliliters of 6 mol L–1 HCl was then added to the centrifuge tubes. The tubes containing soil-acid mixture were heated in a water bath to 95°C for 18 h. Samples were then centrifuged at 2000 x g for 60 min and the supernatant, H-HF, was decanted and analyzed for total organic C (TOC) using a Total Organic Carbon Analyzer–TOC-5050A (Shimadzu Corp., Columbia, MD) and TKN after digestion using a flow analyzer (Clesceri et al., 1998). The remaining R-HF was washed four times with 25 mL of deionized H2O with the washes being analyzed and included as part of the H-HF. The R-HF was dried at 60°C for 48 h and weighed. The bulk soil samples, LF, MF, and R-HF were analyzed for C and N concentration by dry combustion with the NC Soil 2100 analyzer.

Net Nitrogen Mineralization Potential
A net anaerobic N mineralization experiment modified after Whalen et al. (2000) was conducted on fresh samples from control and herbicide plots from the Eatonton sites of the Piedmont in June and December 2001 and from the Waycross sites of the Lower Coastal Plain in September and November 2001. Net anaerobic N mineralization potential was assessed on 5 g (oven-dry basis, sieved <2 mm) of WS, on 4.95 g of soil plus 0.05 g of oven dry L/MF, <2.0 g cm–3, (WS + L/MF) and on 4.50 g of soil plus 0.50 g of oven dry HF, >2.0 g cm–3 (WS + HF). The amounts of L/MF, and HF were chosen to provide similar quantities of C and N to WS and to minimize the effect of SPT on rates of N mineralization (Whalen et al., 2000).

The different soil mixtures were placed in test tubes and 12.5 mL of deionized H2O was added. The tubes were inverted a couple of times to submerge the sample and then lightly tapped against the lab bench to remove air bubbles. The samples were incubated anaerobically for 7 d at 40°C (Keeney, 1982). The incubated samples were extracted with 12.5 mL of 4 mol L–1 KCl and analyzed for NH+4–N with an Alpkem AutoAnalyzer (OI Analytical, College Station, TX).

Statistical Analysis
Data for the whole soil and fractionation analysis were analyzed by ANOVA using a general linear model (GLM). Before analysis the fractionation data was normalized to 100% C and N recovery. This was done to avoid bias from variation resulting from the uncertainty of measuring the mass of a small amount of material, particularly in the LF and MF, which could weigh as little as 30 mg, and then multiplying by high C and N concentrations to obtain contents. Additionally, a square root transformation was performed on the data due to significant differences (P = 0.05) in variance within locations based on Levene's test for heteroscadisity (Dean and Voss, 1999). The statistical design was a nested block design with region (Piedmont or Coastal Plain), location (Eatonton or Athens and Waycross or northern Florida) nested within region, and replicate blocks nested within location within region. Region, location, herbicide, and their interactions were tested with the appropriate error term. For the analysis of the N mineralization data, the two sampling locations, Eatonton in the Piedmont and Waycross in the Lower Coastal Plain, were analyzed separately. The factorial combination of laboratory treatment (WS, WS + L/MF, WS + HF) and herbicide were evaluated using a blocked design.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Soil Carbon and Nitrogen
Average bulk soil C and N concentrations ranged from 9.9 to 18.7 g C kg–1 and 0.40 to 0.74 g N kg–1 across the study sites and there were no observable differences due to geographic region (Table 2). Across all locations herbicide treated soils had significantly lower soil C concentration (P < 0.05), 12.8 g kg–1, compared with untreated control soils, 16.1 g kg–1. This decrease was consistent for sites from all locations and ranged in magnitude from 17% at the Waycross sites to 25% at the Athens sites (Table 2). Associated decreases in soil N concentration, 0.51 g kg–1 compared with 0.63 g kg–1, due to herbicide were also observed (P < 0.10). The N decrease was consistent for all the sites ranging in magnitude from 12 to 32% (Table 2). A slight increase in the bulk density of herbicide plots was associated with the observed C decrease, although the percentage change (1–16%) was less than changes in C or N concentrations (Table 2).


View this table:
[in this window]
[in a new window]
  		Table 2. Carbon and N concentration and C/N ratio (mean values and SE of mean in parentheses) of surficial (0–10 cm or genetic A horizon) mineral soils under pine plantations from various sites in the Piedmont and Lower Coastal Plain regions of Georgia and North Florida. Soils were sampled between 1998 and 2001 when pine stands ranged in age from 12 to 18 yr.**

 
The percentage of total mass recovered in the sum of the fractions (LF, MF, H-HF, and R-HF) ranged from 99.2 to 100.1% and averaged 99.8% (data not shown). Carbon and N recovered in the sum of the fractions was more variable and ranged from 71 to 123% and averaged 94% for C and 92% for N (data not shown). The C and N recoveries were normalized to 100% before analysis by correcting the C and N content in each fraction proportionally to its weight (Table 3 and 4).


View this table:
[in this window]
[in a new window]
  		Table 3. Mass of carbon in light, medium, hydrolyzable and residual soil fractions (mean values and SE of mean in parentheses) isolated from surficial (0–10 cm or genetic A horizon) mineral soils under pine plantations from various sites in the Piedmont and Lower Coastal Plain regions of Georgia and North Florida. Soils were sampled between 1998 and 2001 when pine stands ranged in age from 12 to 18 yr. Fractionation data is normalized to 100% soil C recovery.

 

View this table:
[in this window]
[in a new window]
  		Table 4. Mass of N in light, medium, hydrolyzable, and residual soil fractions (mean values and SE of mean in parentheses) isolated from surficial (0–10 cm or genetic A horizon) mineral soils under pine plantations from various sites in the Piedmont and Lower Coastal Plain regions of Georgia and North Florida. Soils were sampled between 1998 and 2001 when pine stands ranged in age from 12 to 18 yr. Fractionation data is normalized to 100% soil N recovery.

 
Across all locations competition control has resulted in reduced soil C in the LF and MF fractions (Table 3). The soil C decreases in the LF accounted for 42% of the whole soil C decreases while the MF accounted for an additional 36% of soil C decreases (Table 3). Although the majority of soil C decreases occurred in the lightest fractions, LF and MF C were only slightly more sensitive to herbicide treatment than WS C. There was a 24% decrease in LF C and a 23% decrease in MF C due to competition control compared with a 20% decrease in whole soil C (Table 2 and 3). Decrease in LF and MF C might well explain the slight increases in surface soil bulk density (Table 2). For N, there were also substantial decreases in the LF and MF, which account for 31 and 39%, respectively, of the decreases in WS N (Table 4). No significant differences were apparent in C and N contents in H-HF or R-HF when tested across all sites. There were also few significant effects of herbicide treatments on the C/N ratios of whole soils or the other fractions. The C/N ratio of the H-HF was an exception being significantly higher in herbicide treated soils, 9.6:1, compared with untreated control soils, 8.5:1 (Table 5).


View this table:
[in this window]
[in a new window]
  		Table 5. Ratio of C/N in light, medium, hydrolyzable and residual soil fractions (mean values and SE of mean in parentheses) isolated from surficial (0–10 cm or genetic A horizon) mineral soils under pine plantations from various sites in the Piedmont and Lower Coastal Plain regions of Georgia and North Florida. Soils were sampled between 1998 and 2001 when pine stands ranged in age from 12 to 18 yr.

 
Within a given region or location within a region, statistical significance was less apparent, although measured decreases in WS or soil fractions for C and N were highly consistent. For example, in the Coastal Plain not only were decreases in LF and MF C significant (Table 3), but decreases of C in each fraction were 24 and 23%, respectively, compared with a 20% decrease in WS C for the region (Table 3). These results for the Coastal Plain were equal to those estimated across all locations. Results for the Piedmont were not statistically significant but demonstrated similar percentage losses (e.g., 25% in LF, 21% in MF, and 20% in WS for the region). Results for N were similar. In the Lower Coastal Plain, LF, and MF N decreased by 19, and 26%, respectively (Table 4), compared with a 19% decrease in WS N (Table 2) and in the Piedmont, LF and MF N decreased by 30, and 24%, respectively (Table 4), compared with a 19% decrease in WS N (Table 2).

Soil Carbon and Nitrogen Partitioning among Fractions
Partitioning of SOM into the defined fractions differed across soils of varying textural classes. Surface soils with sand or loamy sand textures, prevalent in the Lower Coastal Plain, tended to have the majority of C and N occurring in the lighter more labile fractions (Fig. 1) . Approximately 79% of the C and 75% of the N in these soils was accounted for in the LF and MF. In the finer-textured Piedmont surface soils, only 48% of the C and 27% of the N was accounted for in the LF and MF. The remaining 52% of the C and 73% of the N in the Piedmont soils was accounted for in the H-HF and R-HF.



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 1. Proportion of total (a) soil C and (b) soil N in light, medium, hydrolyzable and residual fractions from upper surface soils under pine plantations across Lower Coastal Plain and Piedmont sites in Georgia and North Florida. Soils were sampled between 1998 and 2001 when pine stands ranged in age from 12 to 18 yr.

 
Net Nitrogen Mineralization Potential
No significant differences in net N mineralization potential occurred between the two sampling dates at either location (Table 6). Net N mineralization tended to be higher at the Eatonton sites compared with the Waycross sites. For example, WS net N mineralization averaged across all sampling dates and treatments was 40.3 mg N g N–1 7 d–1 at the Eatonton sites compared with 18.6 mg N g N–1 7 d–1 at the Waycross sites. For the Eatonton sites, there was a significant (P < 0.05) decrease of 27, 41, and 31% in net N mineralization due to herbicide for WS, WS + L/MF, and WS + HF, respectively (Table 6). For the Waycross sites, however, there was only a decrease (P < 0.10) in net N mineralization due to herbicide for the WS + L/MF treatment (18%); herbicide-treated soils did not have decreased mineralization for the WS and WS + HF combinations (Table 6). Net N mineralization was significantly (P < 0.05) lower in the WS + L/MF combination compared to WS or WS + HF, approximately 31% lower in both cases, at the Eatonton sites (Table 6). This trend was also observed (P < 0.10) at the Waycross sites with an approximate decrease of 19% in the WS + L/MF combination compared with the others (Table 6).


View this table:
[in this window]
[in a new window]
  		Table 6. Rate of net N mineralization potential during 7-d anaerobic incubation (mean values and SE of mean in parentheses) for control and herbicide treated 0-10 cm mineral soils from the Eatonton sites in the Piedmont, sampled in June and December 2001, and Waycross sites in the Lower Coastal Plain, sampled in September and November 2001. Laboratory treatments include WS = whole soil; WS + L/MF = whole soil with additional light and medium fraction material, <2.0 g cm–3; and WS + HF = whole soil with additional heavy fraction material, >2.0 g cm–3.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
After 12 to 18 yr of intensive understory competition control through herbicide treatments there has been significant decreases of surface soil C and N in these Southeastern pine stands. On average, soil C and N concentrations were 20 and 19% lower, respectively, in herbicide treated soils than untreated controls (Table 2). These soil C and N decreases are of similar magnitude as those reported in the literature that range from 18 to 33% for soil C and 10 to 30% for soil N (Aust and Lea, 1991; Polglase et al., 1992; Carlyle, 1993; Munson et al., 1993; Busse et al., 1996). In the work of Busse et al. (1996), after 35 yr of understory vegetation removal in ponderosa pine stands in central Oregon, the corresponding loss in surface soil (0–12 cm) contents of C and N were 4.2 Mg C ha–1 and 131 kg N ha–1. In the current study, C and N concentration decreases translate to surface soil content decreases of 3.3 to 3.9 Mg C ha–1 and 93 to 220 kg N ha–1 (Fig. 2) . The content estimates in Fig. 2 use a constant mass basis (Davidson and Ackerman, 1993) and thus assumes surface soils have not eroded but have compacted due to loss of LF and MF C. Much of the C in these surface soils was likely lost from the soil as CO2 while N may either have leached from the ecosystem, been retained in deeper depths, or been captured by vegetation during early stand development.



View larger version (27K):
[in this window]
[in a new window]
  		Fig. 2. Contents of C and N in surface soils of control and herbicide plots. Eatonton (n = 4 at Age 12), Athens (n = 2 at Age 12), and Waycross (n = 4 at Age 12) were all sampled from 0- to 10-cm depth while N. Florida (n = 3 at Age 18) was sample through the genetic A horizon. Herbicide treatments were used to achieve complete competition control in these experimental plots.

 
The observed decreases in surface soil C are not trivial but represent only a small percentage of expected total ecosystem C in these mid-rotation pine plantations (Richter et al., 1999; Shan et al., 2001). In fact, on an ecosystem basis, observed increased accumulations of forest floor C (i.e., soil O horizons) in herbicide treated plots compared to controls of 4.4, 4.3, and 1.1 Mg-C ha–1 in North Florida, Eatonton, and Waycross, respectively (Shan et al., 2001; Borders et al., 2004), could counter balance the observed surface soil decreases. The long-term fate of these two C pools (i.e., forest floor or mineral soil) might differ, however, under continuing management such as prescribe burning. Furthermore, the functional role of these different C pools in relation to soil moisture storage or N cycling likely differs and may have important consequences for soil productive capacity.

One possible mechanism for the observed decreases in mineral surface soil C and N is increased organic matter decomposition. Rates of decomposition may increase due to elevated soil temperatures and/or more favorable soil moisture resulting from the absence of vegetative cover in the herbicide plots, particularly early in stand development. Vegetative cover can affect soil temperature by changing the albedo, intercepting radiation, and through latent heat transfer during evapotranspiration (Hillel, 1982). Aust and Lea (1991) reported a 29% decrease in surface soil (0–10 cm) organic matter content 2-yr postharvest due to herbicide treatments that was associated with an average temperature increase of 2.4°C. Similarly, Munson et al. (1993) attributed significant decreases in total C and N concentration in forest humus and surface mineral soil following 4 yr of vegetation control to increases in soil temperature as well as soil moisture. Soil temperature and moisture were not measured in the current study but model simulations for both a Piedmont and Coastal Plain location suggest that temperature increases of as much as 10°C during early stand development would not be sufficient to account for all the observed soil C decrease (Echeverría and Markewitz, unpublished data, 2004).

A second mechanism for the observed decreases in SOM is reduced organic matter input from above- and belowground litter. Competing vegetation can significantly influence soil C levels, particularly in the early stages of stand development when C inputs from the developing plantation are minimal (Woods et al., 1992). Busse et al. (1996) reported 21% less C in aboveground litter in the absence of understory vegetation corresponding to a 25% decrease in upper mineral soil C content. Additionally, competition control is expected to reduce the input of root litter into the SOM pool, a major source of C in forest soils (Tate et al., 1993). According to Ewel and Gholz (1991), root turnover is a major input to soil C in Florida pine plantations, and the contribution from understory vegetation is nearly as great as that from pine. In a comprehensive ecosystem study of the North Florida sites used in this study, Shan et al. (2001) found at Age 17 that fine root turnover was reduced by 0.4 Mg C ha–1 yr–1 in the absence of competition. Assuming a constant reduction in fine root turnover over the 18 yr of stand growth, this lost input could account for a decrease of 7.8 Mg C ha–1 in the soil C pools. Similarly, the recent work of Crocker (2002) at the Waycross sites found an average 0.93 Mg C ha–1 yr–1 decrease in fine root productivity in herbicide treated plots compared with controls at Age 13. This nearly 2/3 decrease in fine root C inputs at the Waycross sites could account for much of the 17% decrease observed in soil C. Unfortunately both these studies investigate stands at mid-rotation making it difficult to ascertain if similar effects existed early in stand development. In model simulations Echeverría and Markewitz (unpublished data, 2004) found that a decrease in root C inputs of 10% throughout stand development under herbicide treatment could generate soil C depletions that are consistent with observed results.

Density fractionation revealed that the majority of soil C and N decreases associated with herbicide treatments occurred in the LF and MF (Table 3 and 4). Combined, these two most labile fractions account for 78 and 70% of the C and N decreases, respectively. In the Lower Coastal Plain sites these same fractions account for 97% of the C and 92% of the N decreases while in the Piedmont LF and MF account for 61% of the C and 48% of the N decreases. Light fraction SOM has been shown to be more sensitive to management practices than total SOM (Strickland and Sollins, 1987; Cambardella and Elliot, 1992; Janzen et al., 1992; Hassink, 1995; Barrios et al., 1996). For example, Barrios et al. (1996) studying maize and legume cropping systems in a field experiment in Kenya found that macro-organic matter (150–3000 µm) LF (<1.13 g cm–3) was the most sensitive fraction, followed by LF (<1.7 g cm–3) isolated from WS. Although the majority of SOM losses have occurred in the lightest fractions in this study, the LFs were only slightly more sensitive to herbicide treatment than whole soil SOM. This is understandable in the Coastal Plain sites where LF and MF C and N constitute 79 and 75% of the total soil C and N, respectively (Fig. 1) suggesting that changes in WS SOM would be driven by, and analogous to, changes in LF SOM. However, in the Piedmont sites where LF and MF C and N constitute only 48 and 27% of the total soil C and N, respectively, it might be expected that the effects of herbicide on SOM would be larger in the lighter fractions compared with WS. This does not appear to be true.

Fractionation of SOM demonstrated striking differences in soil C and N partitioning across the soils of varying textural classes. The coarser-textured soils prevalent in the Lower Coastal Plain tended to have the majority of C and N in the lighter more labile fractions compared with the finer-textured soils of the Piedmont (Fig. 1). This difference was most impressive for N, with the Piedmont soils having only 27% of the total soil N in the lighter fractions compared with 75% in the Lower Coastal Plain soils. Although, in the current case, soil texture is confounded with differences in physiographic region and land use history, this result is consistent with previous work that has shown that the proportion of SOM in light fractions decreases as soil clay and silt content increases (Greenland and Ford, 1964; Richter et al., 1975; Hassink, 1995). For example, Hassink (1995) reported a significant negative correlation between silt + clay content and the proportions of soil organic N in light (<1.13 g cm–3) and intermediate (1.13–1.37 g cm–3) macro-organic matter (>150 µm) fractions. Such soil texture effects have been explained by differences in the specific surface area, which increases from sand to clay, and by the physical protection of SOM in fine pores where it is inaccessible to decomposers (Schulten and Leinweber, 2000). These results suggest that physical protection of SOM by clay surfaces is of greater importance in the Piedmont surface soils, characterized by finer textures, than the surface soils in the Lower Coastal Plain (Schulten and Leinweber, 2000).

The combined effect of decreases in SOM fractions and differential partitioning of SOM on nutrient cycling in coarse- versus fine-textured soil was assessed through the potential net N mineralization assays (Table 6). In general, net N mineralization potential was substantially higher in the finer-textured Piedmont soils than in the Lower Coastal Plain soils. All things being equal, we would expect that decomposition rates of organic residues in fine-textured soils would be lower than in coarse textured soils due to physical and chemical protection of SOM by clays (Sörensen, 1975; Jenkinson, 1977; Ladd et al., 1985; Hassink, 1995). However, disturbance caused by soil sampling, which would disrupt clay aggregates, and differences in SOM quality between the Lower Coastal Plain and Piedmont soils may also contribute to observed differences in mineralization (Hassink, 1994). Average WS C/N ratio in the Lower Coastal Plain soils used in the mineralization experiment was 33:1 compared with 21:1 for the Piedmont soils. The Lower Coastal Plain soils also contain a larger proportion of LF/MF SOM compared with the Piedmont soils, which may indicate greater microbial immobilization resulting in less net N mineralization.

The addition of L/MF material to WS tended to reduce net N mineralization. This was most prominent in the Piedmont soils, where the WS + L/MF treatment had a 32% lower net N mineralization than WS; compared with a 19% reduction found in the Lower Coastal Plain soils. In the Lower Coastal Plain soils, the majority of SOM is in the L/MF (>75%). Therefore, the addition of L/MF material to WS did not significantly alter the C and N content or the C/N ratio. The average C/N ratio of WS, WS + L/MF and WS + HF was 33.1:1, 33.9:1, and 32.4:1, respectively, for Lower Coastal Plain soils compared with 21.4:1, 25.0:1, and 20.6:1, respectively, for Piedmont soils. This smaller range in SOM C/N ratio across the laboratory treatments may help explain the muted response of net N mineralization to the addition of L/MF material in the Lower Coastal Plain soils. In general, our results suggest that LF SOM tends to be a sink for mineral N in these Piedmont and Lower Coastal Plain soils. This is consistent with previous data on N mineralization that purport HF as the primary source of mineralizable N and LF as a potential substrate for microbial immobilization in agricultural and forest soils (Sollins et al., 1984; Boone, 1994; Whalen et al., 2000; Compton and Boone, 2002). Compton and Boone (2002) highlight, however, that the decomposition state of LF material will affect the mineralization status of the LF material much the same way that forest floor organic material can change from a sink to a source over time.

Given the observation that LF SOM tends to be a sink for mineral N in these soils, it follows that competition control, which has been shown to reduce LF SOM, may result in an increase in net N mineralization. Vitousek et al. (1992) working in 4- to 5-yr-old herbicide treated loblolly pine stands in the Piedmont of North Carolina reported increased net N mineralization associated with decreased microbial N immobilization. Conversely, our data suggest that competition control through herbicide tends to reduce net N mineralization potential (N mineralized per unit of N). This was the case for all laboratory treatments for the Piedmont soils and for the WS + L/MF treatment in the Lower Coastal Plain soils. The older age of the stands in the current study compared to the Vitousek et al. (1992) study may indicate that the losses of LF SOM (that provided a source of labile N) occur earlier during stand development. Furthermore, measurements of net N mineralization potential under controlled laboratory conditions, as used in this study, provide a means of assessing the influence of substrate quality on mineralization in the absence of microclimatic changes resulting from management practices (Matson and Vitousek, 1981). It follows, perhaps, that despite decreasing LF SOM that serves as a sink for mineral N, competition control has resulted in the alteration of SOM quality over these 12 to 18 yr of stand growth such that this seemingly anomalous effect on potential N mineralization can be explained. Further work investigating the effect of competition control on SOM quality is warranted.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
This study provides further evidence that herbicide treatment for complete competition control can decrease bulk soil C and N contents. It should be recognized, however, that in this study herbicide treatments were on experimental plots that exceeded normal forestry operational treatments. None the less, increased decomposition rates and/or decreased belowground C inputs that contribute to SOM decreases, will likely persist on operationally herbicided stands, albeit to a lesser degree. Similarly, although decreases in the surface soil C pool were significant, from an ecosystem C budget perspective these decreases were small. The functional importance of this decrease in surface soil C as relates to soil moisture storage, surface charge, or N mineralization, however, may outweigh the C mass lost.

We hypothesized that LF SOM would be most depleted by herbicide treatments. In both the Coastal Plain and Piedmont this was, in fact, observed, although in neither case did LF depletion far exceed bulk soil depletions. Conversely, we hypothesized that Piedmont sites would be somewhat buffered against SOM depletions due to a higher percentage of adsorbed SOM. Although Piedmont soils did demonstrate a higher percent of soil C and N as H-HF and R-HF, bulk soil C and N decreases were of a similar order of magnitude. Furthermore, decreases in potential net N mineralization were more readily evident in the Piedmont, although the large pool of HF N at the Eatonton site did maintain a higher rate of potential net N mineralization compared with the Coastal Plain site at Waycross. The ability of Piedmont soil to sustain a higher rate of N availability despite C decreases may indicate a higher sustainability of inherent soil productive capacity (i.e., productivity achieved without fertilizer additions or improvements in genotype or silviculture) under herbicide treatment. In both Piedmont and Lower Coastal Plain locations, however, it appears that herbicide treatments have reduced the productive capacity of these surface soils.

Received for publication July 31, 2002.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Soil Sci.Home page
D. V. Sarkhot, N. B. Comerford, E. J. Jokela, J. B. Reeves III, and W. G. Harris
Aggregation and Aggregate Carbon in a Forested Southeastern Coastal Plain Spodosol
Soil Sci. Soc. Am. J., September 28, 2007; 71(6): 1779 - 1787.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Echeverría, M. E.
Right arrow Articles by Hendrick, R. L.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Echeverría, M. E.
Right arrow Articles by Hendrick, R. L.
Agricola
Right arrow Articles by Echeverría, M. E.
Right arrow Articles by Hendrick, R. L.
Related Collections
Right arrow Soil Organic Matter
Right arrow Carbon Sequestration
Right arrow Forest Soils

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome