SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 12 March 2007
Published in Soil Sci Soc Am J 71:457-468 (2007)
DOI: 10.2136/sssaj2006.0179
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
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 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 Google Scholar
Google Scholar
Right arrow Articles by Schroth, A. W.
Right arrow Articles by Bostick, B. C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Schroth, A. W.
Right arrow Articles by Bostick, B. C.
Agricola
Right arrow Articles by Schroth, A. W.
Right arrow Articles by Bostick, B. C.
Related Collections
Right arrow Plant and Soil Interactions
Right arrow Biogeochemical Processes

FOREST, RANGE & WILDLAND SOILS

Macronutrient Depletion and Redistribution in Soils under Conifer and Northern Hardwood Forests

Andrew W. Schrotha,b,*, Andrew J. Friedlanda and Benjamin C. Bostickb

a Environmental Studies Program, Dartmouth College, 6182 Steele Hall, Hanover, NH, 03755
b Dep. of Earth Sciences, Dartmouth College, 6105 Fairchild Hall, Hanover, NH 03755

* Corresponding author (Andrew.Schroth{at}Dartmouth.edu).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
While potentially important in the context of global biogeochemical change, the influence of different forest communities on chemical weathering rates in soils is poorly understood. We investigated the influence of four forest types (northern hardwood vs. three conifer plantations) on base cation depletion and redistribution in soils at Marsh–Billings–Rockefeller National Historical Park (MBRNHP) on 100-yr forest development time scales. This site was ideal for the examination of forest-type influence on the chemical denudation of the landscape during soil development due to a unique forest management history. Soils at MBRNHP are mildly acidic and developed on a silicate-rich parent material with trace carbonates. Soil composition beneath different forest types indicates significant depth-dependent differences in cation depletion. Conifer forest surface soils were more acidic and depleted in mineral-phase base cations than those under northern hardwood forests. This was presumably due to aggressive weathering agents produced by fine-root exudation and organic matter decomposition in conifer forest surface soils. At depth, there was a more acidic and sometimes cation-depleted soil profile under northern hardwood relative to conifer forests, which could be associated with deeper root networks of northern hardwood species and related high nutrient demands, proton release, and mechanical weathering of deep soil. Depth-dependent trends were more evident with Ca and Mg depletion profiles than Na and K, suggesting that vegetative enhancement of the weathering environment was most effective in dissolving more pedogenically reactive divalent cation mineral phases. This model of forest-type enhancement of chemical weathering has important ramifications in the context of global biogeochemical change and forest management.

Abbreviations: ICP-OES, inductively coupled plasma optical emission spectroscopy • MBRNHP, Marsh-Billings-Rockefeller National Park • OM, organic matter • PM, parent material • XRD, x-ray diffraction


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Chemical weathering and the rate by which such reactions occur in soils with time play an important role in a variety of global biogeochemical cycles, particularly in the context of global biogeochemical change (Schlesinger, 1997). Silicate weathering is known to be the primary sink of atmospheric CO2 on geologic time scales (Berner et al., 1983; Berner, 1997). Cation release during the weathering of soil minerals is becoming an increasingly vital source of macronutrients such as Ca and Mg (Huntington et al., 2000; Blum et al., 2002; Hamburg et al., 2003), as their atmospheric concentrations decline and "bioavailable" fractions are depleted in surface soils by atmospheric acid deposition. Such weathering reactions are also the primary buffer of catchments and soils to atmospheric anthropogenic inputs of SO4–2/NOx and resultant acidification (April et al., 1986; Kirkwood and Nesbitt, 1991; Palmer et al., 2004). In soil systems, chemical weathering rates are influenced by a variety of parameters that include temperature, precipitation, soil texture, mineral solubility, and the composition of soil solutions. To effectively determine the role of chemical weathering in dynamic large-scale biogeochemical systems, fundamental relationships between such parameters and weathering rates must first be defined.

It has long been understood that vegetation modifies pedogenesis (Jenny, 1941) and specifically the weathering environment by alteration of input solution compositions, cycling and mining of nutrients from mineral reserves, physical disturbance of mineral material, and the generation of aggressive aqueous weathering agents through decomposition (Jenny, 1941; Binkley and Giardina, 1998). To comprehend the role of chemical weathering in global biogeochemical cycles, considerable research has sought to quantify the effect of vegetation on present-day weathering rates at a variety of spatial scales, which have mostly indicated significant enhancement of weathering regimes by vegetation in their respective systems (Bormann et al., 1998; Moulton and Berner, 1998; Raulund-Rasmussen et al., 1998; Moulton et al., 2000). To date, only a few well-controlled studies have sought to determine the effect of different types of forests on mineral weathering, with mixed results (Augusto et al., 2002). Laboratory experiments suggest that soil solution pH and dissolved organic carbon (DOC) can control mineral weathering rates in soil environments, but such an effect may be limited by primary mineral solubility (Drever, 1994b; Stillings et al., 1996; Raulund-Rasmussen et al., 1998; van Hees et al., 2002; Hamer et al., 2003). Soil solution pH in undisturbed soil systems is primarily controlled by the partial pressure of carbon dioxide (pCO2) in the forest floor and related H2CO3 equilibria (Binkley and Giardina, 1998). Elevated soil pCO2 is generated by organic decompositional processes, the magnitude of which will vary by species and soil morphology (Binkley and Giardina, 1998). Soil DOC is also produced from organic decomposition reactions and fine root exudation through the production of weak organic (humic and fulvic) acids, the concentration and structure of which can vary by tree species (Strobel et al., 2001). While these acids lower soil solution pH, they enhance mineral dissolution rates primarily by interacting with mineral surface cations as chelating ligands (Raulund-Rasmussen et al., 1998). The extent to which such organic acids promote soil mineral dissolution is controlled by solution pH that favors complexation (Raulund-Rasmussen et al., 1998). Given that vegetation type can influence the structure and concentration of these weathering agents (e.g., Strobel et al., 2001), the extent and distribution of weathering in soils could be significantly influenced by different species of overstory vegetation. Typically, in spodic profiles, a horizon where such DOC has been removed from solution is observed (Bh) in the mineral soil, suggesting a depth after which such agents may have less influence on soil chemistry on removal from solution (McDowell and Wood, 1984). Such trends were observed over a 50-yr time scale of soil development under different overstories where soil chemistry differences in exchangeable cation concentrations and soil pH were confined to upper soil horizons, indicating that the influence of products of decomposition and fine root exudation might be confined to these upper soil horizons on forest development time scales (Binkley and Valentine, 1991).

While most studies concerning species-specific enhancement of weathering regimes have focused on weathering rate acceleration by conifer species associated with aggressive decomposition and root exudation products, data from recent work suggest that nutrient demand and uptake could influence soil depletion rates, particularly in the case of hardwood species like sugar maple (Acer saccharum Marshall subsp. saccharum) that have high annual nutrient demands and extensive deep rooting systems (Whittaker et al., 1974; Bockheim and Crowley, 2002; Dijkstra and Smits, 2002; Belanger et al., 2004; Fujinuma et al., 2005). The nutrient demands of northern hardwood forest trees far exceed those of conifers, with deciduous trees often containing 100% more Ca, Mg, and K in their aboveground biomass (Whittaker et al., 1974; Bockheim, 1997) and soil organic matter (OM; Finzi et al., 1998a; Dijkstra, 2003). Northern hardwoods have higher annual nutrient demand associated with annual leaf turnover; however, in mature forests much of this demand could be tightly cycled and associated with cation mineralization tied to OM decay (Tice et al., 1996). Such nutrient demands could have substantial effects on soil chemical weathering rates by enhancing soil solution disequilibria and associated mineral dissolution, particularly within deeper sections of the soil profile where much of the aggressive weathering regime associated with decomposition and fine root exudation has been neutralized (Bh), but extensive rooting networks exist (McDowell and Wood, 1984; Dijkstra and Smits, 2002; Jandl et al., 2004).

While the mechanisms of species-specific controls on soil depletion are conceptually straightforward, observation and quantification of such a species-specific effect on cation depletion, particularly on long-term time scales associated with forest management and global change, have proven difficult due to complications associated with experimental design (Augusto et al., 2002). Soil development on a landscape scale is influenced by a variety of soil-forming factors, one of which is vegetation (Jenny, 1941). Nezat et al. (2004) demonstrated the variability of soil weathering rates on a watershed scale at Hubbard Brook Experimental Forest, noting that in addition to forest type, landscape position and elevation have significant effects on soil depletion rates. Studies have shown that sugar maple and hemlock [Tsuga canadensis (L.) Carrière] tend to develop on soil with different parent material (PM) compositions, with sugar maple developing on more weatherable, Ca-rich deposits to support high nutrient demands, but hemlock may outcompete deciduous species in areas where the PM is more depleted (van Breemen et al., 1997; Dijkstra and Smits, 2002). White pine (Pinus strobus L.) has also been shown to outcompete hardwood species on more depleted and less weatherable deposits (Johnson and Siccama, 1979). In a recent review of vegetative effects on soil chemistry, Augusto et al. (2002) were critical of many studies that sought to compare overstory species effects on soil biogeochemistry due to intrinsic variables associated with the compared soils (land-use history, soil morphology and history, parent material heterogeneity, and aspect and topographic variance). Correlations between other soil-forming factors and vegetative cover make it difficult to assign forest-type effects to soil chemistry and illustrate the necessity for control of other variables to examine the role of forest communities in soil development.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Objectives and Hypotheses
In this study, we used a unique field site, Marsh–Billings–Rockefeller National Historical Park (MBRNHP), located in Woodstock, VT, to isolate vegetative influences on soil chemistry by constraining other soil-forming processes. Specifically, we sought to determine if different overstory species influence mineral base-cation depletion in the soil profile on ~100-yr soil development time scales. We hypothesized that such differences should be associated with species-specific nutrient demands, modifications of the physical soil environment, or the forest's ability to generate aggressive organic weathering agents if sufficient time under different forest cover has elapsed to manifest such processes in soil horizon chemistry.

Site Description
At MBRNHP, Fredrick Billings created plantation-style stands of red pine (Pinus resinosa Aiton), white pine, and Norway spruce [Picea abies (L.) H. Karst.], and allowed adjacent stands of successionary northern hardwood forest (sugar maple, beech [Fagus grandifolia Ehrh.], yellow birch [Betula alleghaniensis Britton]) to develop at the turn of the 20th century in areas that were used for sheep farming from the late 18th to the late 19th century (Wiggin, 1993; Wilcke et al., 2000). Geographically, the site lies within the area known as the Vermont Piedmont, with mean annual precipitation of around 100 cm and a mean annual temperature of 5 to 7°C. The park lies on the Devonian Waits River Formation (black garnetiferous phyllite with sparsely interbedded impure metamorphosed limestone). Soils have developed on ~12 000-yr-old sandy glacial till that varies in thickness from 0.3 to 1.1 m across the study site, and are of the Dummerston series. The sites selected at MBRNHP have a well-documented land-use history and stand management on the same soil type (Lautzenheizer, 2002). By comparing soils under the conifer plantations with those under adjacent hardwood stands with similar slope, aspect, elevation, and PM, we isolated forest-type effects on soil chemistry over ~100-yr time scales.

Comparative Site Selection
Sites for quantitative soil pit excavation (Johnson et al., 1991) were selected based on criteria to minimize the influence of soil-forming factors, aside from overstory vegetation type, on soil chemistry (Fig. 1 ). Soil sampling sites were selected based on the following criteria: (i) a site contained successionary stands of northern hardwood forest adjacent to plantations of conifer species of the same age (Table 1); (ii) soils were developed on similar slope, aspect, topography, PM (defined as <10% standard deviation of elements used for depletion calculations in deep till), bedrock, and elevation (within 10 m), and have similar land-use history. The oldest stands (~100 yr old) that met these criteria were chosen for quantitative pit excavation to maximize the time for soil development under different forest communities (Table 1).


Figure 1
View larger version (35K):
[in this window]
[in a new window]

  		Fig. 1. Site relative location map within Marsh–Billings–Rockefeller National Historical Park, Woodstock, VT.

 

View this table:
[in this window]
[in a new window]

  		Table 1. Characteristics of sample sites chosen for this study.

 
Quantitative Pit Protocol
Triplicate quantitative soil pits of a 0.5-m2 in area were excavated following the method of Johnson et al. (1991) under a conifer plantation and a comparative northern hardwood stand. The pit face nearest the tree was <1 m from the base of the tree and <1.5 m from the base of another tree of this type and age. Soil pits were excavated to bedrock. Forest floor and A horizon samples were identified and sampled through field-based estimations of OM content. Two subsequent 10-cm intervals of soil samples were then extracted from B horizons, with additional samples collected at 20-cm intervals until bedrock was encountered. Each horizon was separated with a 12-mm sieve and field masses were obtained from each fraction for bulk density determination. Soil that passed through the sieve was homogenized in the field and an ~3.6-L representative sample was collected. Soil samples were stored in polyethylene bags in field-moist condition until deposited in the laboratory.

Soil Mineralogy
Whole-rock soil mineralogy was determined to identify likely minerals that were present in the PM of the soil and then predict which of these would be most active in the weathering profile. To determine predominant mineralogy, bulk powdered PM samples were analyzed by powder x-ray diffraction (XRD). In addition, trace carbonate was identified by effervescence in 7% HCl, as these phases are too low in concentration in these silicate-rich soils to be identified by XRD.

Soil Chemistry
Subsamples of each horizon were oven dried in the laboratory at 105°C to calculate water content. Soil pH was determined for a 1:1 mixture of air-dried soil to deionized water after 10 min of reaction time. Samples of the <2-mm soil fraction of each horizon were isolated by sieving air-dried soil, and were then used for elemental analysis. Major element concentrations in the OM were determined for the forest floor, A horizon, and 0- to 10-cm-interval soil samples using standard 1 M HNO3 extractions (Friedland et al., 1984). Surficial soil horizons were also extracted with 0.1 M BaCl2 to determine exchangeable or "bioavailable" cation concentrations (Hendershot et al., 1993). To minimize the potential of removing cations associated with mineral material, exchangeable extractions were used for the 0- to 10-cm profiles because there was less OM present in these mineral soil samples. All soil samples underwent a LiBO2 fusion at 1000°C to determine bulk soil composition. Each solution resulting from extraction and digestion procedures was analyzed for Al, Ca, K, Mg, Na, Si, Ti, and Zr by inductively coupled plasma optical emission spectroscopy (ICP-OES). Quality control of extractions and digestions was assured by running standard reference materials (National Institute of Standards and Technology reference materials 2711 and 1572) of soil and vegetation samples. Our values for relevant elements are within 5% of known compositions of these reference materials of similar matrix to our mineral and organic soil samples.

Long-Term Chemical Weathering Rates and Depletion Factors
Long-term weathering rate and depletion factor calculations are based on concentrations of labile (Ca, Mg, K, and Na) and immobile (usually Ti or Zr) elements. The principal assumptions associated with this method are that (i) Ti or Zr is immobile, (ii) the >2-mm fraction in the soil profile contributes minimally to chemical weathering, and (iii) the soil's original PM composition and age are accurately constrained. Long-term element depletion rates and factors were calculated from soil elemental concentrations and bulk density measurements using the following equations (Brimhall and Dietrich, 1987; Egli and Fitze, 2000):

Formula 1[1]

Formula 2[2]

Formula 3[3]

Formula 4[4]
where {rho}s and {rho}p are field-based, water-content-corrected density measurements for soil and parent material, respectively, CTi,p and Cl,p are Ti and labile base cation concentrations (mol kg–1) in the PM, CTi,s and Cl,s are Ti and labile base cations in soil horizons, isinTi,s is the strain or volumetric change of a soil horizon (of {Delta}z field-measured soil horizon thickness) during pedogenesis, and {tau}l,s is the mass transport function of the select labile element, subscript l.

Depletion factors (D) for labile base cations were also calculated to examine chemical loss or enrichment of each cation in soil horizons relative to the original composition of glacial PM. A depletion factor for a labile cation in the soil is:

Formula 5[5]

Depletion factors <1.0 indicate the fraction of the labile cation that has been lost relative to the original PM composition. A depletion factor = 1.0 indicates that none of the labile element has been lost from that section of soil, while a depletion factor of near 0 would indicate close to complete removal of the labile element from the soil section during pedon development. A depletion factor >1.0 indicates relative enrichment in the labile element in the soil profile.

To include data from the organic-rich surface horizons in this calculation, the labile element fraction associated with OM must be removed before calculating a weathering rate for such a horizon. Inclusion of base cations associated with OM could drastically underestimate weathering rates in organic-rich surface soils. To our knowledge, the first and only study to attempt to include OM-rich surface soils in depletion factor calculation was Nezat et al. (2004), who used the composition of the litter layer (Oi horizon) as a proxy for the major element composition of organic matter in the Oa horizon. Although not relevant to their study's results, this approach could lead to inaccuracy of surface soil weathering due to known changes of OM major element composition with depth and age (Yanai et al., 1999). The composition of the Oi horizon is probably significantly different from that of the Oa or A horizons, which contain older, more mature, partially mineralized organic matter (Yanai et al., 1999). To avoid inaccuracy in quantifying surface soil weathering, which was vital to achieve the goals of our study, we used two selective extractions (Friedland et al., 1984; Hendershot et al., 1993) to determine the amount of labile elements associated with OM in the forest floor and surface soils (horizons ≤10 cm below the forest floor for this study). To remove the fraction of labile cations associated with OM from depletion factor calculations, Eq. [5] requires modification to incorporate selective extraction data. For Oa or A horizons, depletion factors were calculated by

Formula 6[6]
where f and n subscripts indicate concentrations (mol kg–1) obtained from LiBO2 fusion data and concentrations (mol kg–1) for the same sample obtained from 1 M HNO3 extractions, respectively. For the 0- to 10-cm sections of each soil pit, an exchangeable (0.1 M BaCl2) extraction (Hendershot et al., 1993) was used to remove OM-associated cations, and depletion factors were calculated by

Formula 7[7]
where subscript e indicates concentrations (mol kg–1) from exchangeable extractions. These selective extractions allowed us to characterize more accurately chemical depletion and weathering in OM-bearing surface soils.

Parent Material Normalization
The parent material of MBRNHP is dominated by sandy till overlying the Devonian Waits River Formation. While much of this material could be sourced in the underlying formation, a component of its composition is probably sourced in other regional formations that were transported from adjacent areas by the receding glaciers. We assumed parent material was unaltered at 1-m depths and collected at least five such samples at each site. The composition of these "deep till" samples was then averaged within each site and assumed to represent original PM composition for weathering rate calculations. The normalization of weathering rates to a deep till composition was required to account for soil pits with a shallow till to bedrock contact, where the relatively shallow base section of the pit was probably altered by the weathering front and therefore could not be assumed to be a representative composition of unaltered PM for overlying soil horizons.

Statistical Analyses
To identify significant differences in soil composition, it was important to rigorously examine analytical and sampling variability in our measurements. Standard errors for each element are associated with both the variance of triplicate fusions (n = 3 for each horizon of each pit) and that of three soil samples collected from the same depth interval extracted from soils underneath the same forest type at the same site (Fig. 1 and Table 1). Standard errors for A horizon and 0- to 10-cm B horizon data also included the standard error associated with extraction data (n = 3). Error bars associated with each depletion factor data point were then calculated through the propagation of the standard error for all concentrations involved in calculating a depletion factor for an individual element in a soil profile (Eq. [5]–[7]). Error associated with chemical weathering rates also included the standard error of field-measured soil density (n = 3). The standard error of each labile cation weathering rate in each soil horizon was then propagated to calculate the bulk soil profile weathering rate error.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Parent Material Properties
Parent material mineralogy, as identified by XRD, was predominantly quartz and muscovite, with trace amounts of plagioclase, biotite, and andradite. Calcite was not detectable by XRD but was identified by effervescence when PM samples were exposed to dilute (7%) HCl. Of these minerals, plagioclase, biotite, and calcite are considered relatively reactive in silicate-rich parent materials, and their dissolution probably regulates weathering reactions and nutrient supply from mineral sources to forests (April et al., 1986; Blum et al., 2002; Oliva et al., 2004). To confirm PM compositional homogeneity within sites, we examined the variability of the elemental composition (Na, K, Mg, Ca, Ti, Zr) relevant for weathering rate calculations of five deep (~1-m) till samples collected randomly within each site. The standard deviation of major elements and Ti in parent material samples selected within individual sites was <10% in all cases, while Zr concentrations had somewhat higher variations (Table 2). Higher Zr variability can be explained by the extremely low concentration (near the detection limit of our ICP–OES) of Zr in these surficial deposits. Due to the variability and low concentration of Zr within the PM and soil samples, we chose Ti as the immobile index element to use for weathering rate and depletion factor calculations. The major and trace element data for the PM indicate that soils at Sites 1, 2, and 4 have developed on similar PM on a site-by-site basis and can be considered suitable for comparative purposes within our sample plan, since PM does not vary by forest type within any of the sites (Table 2). Data from Site 3 on the eastern side of the park was discarded due to parent material heterogeneity associated with a localized fluvial deposit at the northeast corner of the park.


View this table:
[in this window]
[in a new window]

  		Table 2. Mineral phase composition of soil profiles at Marsh–Billings–Rockefeller National Historical Park. The A horizon and 0- to 10-cm samples have had organic-associated cations removed from their composition based on selective extraction data. Most of the 60- to 80-cm intervals do not have standard error values because only one of the three pits at that site contained soil at that depth. In that case, error bars are based exclusively on parent material standard error and are probably underestimates of error for those measurements.

 
Soil Properties
The soils at these sample sites are fine sandy loams of the Dummerston series, developed on 8 to 10° slopes, which are loamy, mixed, active, frigid, Typic Dystrudepts. Based on field observations, these soils were poorly developed with little horizonation under all tree species, and differences in soil appearance under each forest type were not observed. The chemical composition of these soils is listed in Table 2. Soils were generally moderately acidic in the upper horizons, with decreasing acidity with depth (Fig. 2 ). Norway spruce had the most acidic upper soil horizons, while northern hardwood soils were slightly more acidic at depth than those under conifers (Fig. 2). Soil pH values ranged from 4.8 in the upper horizons of soils under Norway spruce to 6.7 in the deepest horizons under conifers (Fig. 2). In general, the deeper horizons of hardwood trees were slightly more acidic than those under conifers, while the upper horizons under conifers were slightly more or equally acidic than comparable soils under northern hardwood forests (Fig. 2).


Figure 2
View larger version (22K):
[in this window]
[in a new window]

  		Fig. 2. Soil pH data from soils at Marsh–Billings–Rockefeller National Historical Park under northern hardwoods and conifers.

 
Interestingly, as a pedon, these soils were more alkaline in pH than soils developed on similar glacial deposits in surrounding areas, suggesting that the glacial deposits on which MBRNHP soils have developed contain minerals with higher acid buffering capacity than the regional till bodies that underlie much of New Hampshire and Vermont (Jersak et al., 1995; Nezat et al., 2004). Classic spodic development typical of soils formed on silicate-rich glacial deposits in this area of Vermont, particularly under conifer species, was noticeably absent at MBRNHP. We attribute this difference to the trace carbonate phases observed in the glacial deposit, possibly sourced in the underlying calcitic Waits River Formation, which provide an additional source of acid buffering in this soil (Doll, 1969). Its presence, even as a trace component of the soil, appears to be an important component of its weathering system and related buffering capacity. This is not surprising, as it is well known that relatively soluble minerals, even at low concentrations, can exert a strong influence on soil weathering profiles, acid buffering capability, and forest nutrient supply (April et al., 1986; Blum et al., 2002; Oliva et al., 2004).

Surface Soil and Forest Floor Organic-Matter-Associated Cation Composition
Through examination of labile major element nutrient concentrations in surface soils at MBRNHP by forest type, certain species-specific trends in the data are evident. Labile Ca concentrations were significantly higher in northern hardwood forest floors than comparative conifer forest floors (Table 3). This is not surprising, as high Ca concentrations due to high Ca nutrient demands of hardwood species relative to conifer stands is well documented (Whittaker et al., 1974; Tice et al., 1996; Finzi et al., 1998b; Dijkstra and Smits, 2002; Dijkstra, 2003; Fujinuma et al., 2005). Forest floor concentrations of Mg and K did not differ significantly with the exception of Norway spruce having a more Mg-rich forest floor than its comparable northern hardwood forest. This suggests that either the annual nutrient demand for Mg and K does not differ by species or that these nutrients are more tightly cycled by the deciduous forest than Ca. The literature supports the latter contention, as higher demand for all of these elements by northern hardwood forests is well documented and tight cycling of these nutrients by hardwood species relative to conifers was also observed by Tice et al. (1996) during 40 yr of forest and soil development at San Dimas Experimental Forest.


View this table:
[in this window]
[in a new window]

  		Table 3. Organic and exchangeable extraction data for forest floor (FF), A horizon (A), and 0- to 10-cm mineral soil from Marsh–Billings–Rockefeller National Historical Park.

 
Depletion Factors and Long-Term Chemical Weathering Rates
Through the normalization of weathering rate calculations to a soil parent material composition of deep till sampled at each site (Table 2), removal of organic fractions from surface soils (Table 3), and the control of other surficial variables (Table 1), we can examine long-term chemical weathering rates and soil depletion profiles for forest-type effects on the chemical denudation of MBRNHP soils. Depletion profiles at the sites selected for this study show systematic differences by overlying vegetation type, which can be explained based on the mechanisms of vegetative enhancement of chemical weathering discussed above.

Base Cation Depletion in Surface Soils
Calcium, Mg, Na, and K are more depleted in the upper soil horizons underneath conifer stands than under comparable northern hardwood stands (Fig. 3 ), which corresponds to lower pH and Ca/Al ratios observed in conifer soils (Table 3, Fig. 2). These profiles indicate that in the upper section of the soil profile, high concentrations of aggressive weathering agents derived from conifer OM decomposition and root exudation (i.e., low-molecular-weight organic acids) were more aggressively weathering the soil profile than those produced by hardwood trees. Our data are consistent with other studies that indicate Norway spruce has the potential to deplete the soil profile more than hardwood species, particularly in surface soils where decomposition products are an important component of the weathering system (Augusto et al., 2002). Previous studies are, however, less conclusive concerning the influence of pine species on chemical weathering rates (Augusto et al., 2002). While our pH data did not indicate a difference in acidity in surface soils when comparing soils under pine and hardwood forests, depletion profiles under red and white pine species were slightly more depleted in Ca, Mg, Na, and sometimes K than those under comparative northern hardwood stands in surficial horizons (Fig. 3). These data suggest that while decomposition products of these species did not differ significantly in their effect on soil pH, solutions draining pine species appear to have more aggressive chelation properties that influence mineral cation depletion in surface soils.


Figure 3
View larger version (35K):
[in this window]
[in a new window]

  		Fig. 3. Base cation depletion factors for soil horizons at three sample sites with conifer forests and comparative northern hardwood forests.

 
It is also important to note that, in general, species-specific differences in depletion profiles were more evident and statistically significant in Ca profiles than in other labile base cation profiles (Fig. 3). It has been demonstrated by laboratory studies that in simulated pedogenic systems, organic weathering agents (H2CO3, chelating ligands) will have a more substantial influence on the dissolution of the more reactive minerals in the soil profile (Drever and Stillings, 1997; Raulund-Rasmussen et al., 1998; Dijkstra et al., 2003). Our data indicate that Ca-bearing phases (i.e., calcite and plagioclase) were more susceptible to organic decompositional products than less soluble K-bearing phases (i.e., muscovite and orthoclase). These data also suggest that perhaps many studies that have attributed differing soil pH values to differences in pine and hardwood OM decomposition products alternatively might be observing differences in PM buffering capacity to atmospheric and organic acids (c.f., Augusto et al., 2002). Elevated concentrations of organic acids would not necessarily have a strong influence on soil pH values, as the negative logarithm of the acid dissociation constant (pKa) values for many such organic acids is high relative to the soil pH, and pH differences by species could be masked by the influence of elevated soil pCO2 and strong atmospherically derived acids on soil pH values, particularly on soils with low buffering capacities (Strobel et al., 2001). These field-based data highlight the importance of the chelation process rather than acidity with respect to tree species' influence on chemical weathering reactions, particularly in OM-rich surface soils. Furthermore, we propose a scale of Norway spruce > pine > northern hardwood forests in enhancement of chemical weathering of these OM-rich surface soils, probably due to differences in chelation properties of soil solutions of each forest type.

Cation Depletion in Deep Soil Intervals
In most cases, deeper soil horizons (>10–20-cm interval) under northern hardwood forests were more depleted in Ca, Mg, and, to some extent (differences by cation explained by the previously mentioned rationale), K and Na than those under conifer stands, although rarely are differences between species more than one standard error (Fig. 3). The enhanced cation depletion of deeper soils beneath hardwoods relative to conifers is opposite to trends observed in surface soils. These data can be explained by a shift in the mechanism by which vegetation enhances weathering at depth in these soils. Northern hardwood trees in general have higher nutrient demands than conifer species for base cations in temperate forests, particularly Ca, and more extensive deep rooting systems to access mineral nutrient sources (Dijkstra and Smits 2002; Jandl et al., 2004). By extracting macronutrients from deep soil solutions, vegetation can enhance chemical weathering according to Le Châtelier's principle (Zumdahl, 1997), through increasing soil solution disequilibria with respect to adjacent nutrient-bearing soil minerals. In addition, nutrient uptake causes proton release from root surfaces, which further enhances adjacent mineral solubility by lowering solution pH. The physical disturbance associated with hardwood species' deep root networks should also enhance chemical depletion in this zone of the soil profile by actively exposing fresh mineral surfaces to the active weathering front. A more aggressive deep weathering front under hardwood species at MBRNHP is supported by the elevated acidity of these soils at depth under hardwood species relative to conifers. While it is often assumed that weathering under conifers is greater than under deciduous trees due to evidence from surface soils and associated solutions (c.f., Augusto et al., 2002), recent work by Dijkstra and Smits (2002) suggested that mature sugar maples weather deep mineral soil Ca pools to a greater extent than hemlock trees based on deep soil solution compositional data. Bockheim and Crowley (2002) also proposed possible greater weathering of soil minerals under mature sugar maple trees than hemlock based on deep soil solution data from Wisconsin forests. Our data from 100 yr of soil development support the results of these studies, but we used a different method and time scale. Furthermore, we found that the forest-type effect on divalent base cation depletion in the deeper section of the soil profile follows a scale of northern hardwood > Norway spruce ~ pine. In general, these trends remain within one standard error of depletion factors and are less evident for Na and K because of the association of these cations with more resistant silicate phases and less demand or tighter cycling by biota. We predict that such forest-type-specific trends could become well defined with increased time of solum development under these forest types.

Alternative Explanations for Observed Depletion Trends
Considering that these soils have been developing for thousands of years since PM deposition, and that we attribute some trends in elemental depletion to 100 yr of soil development under different forest types, alternative physical or chemical processes during soil development history that might produce the observed trends are herein explored. A possible scenario that could produce different depletion profiles and must be considered in soil development on glacial unconsolidated deposits is that heterogeneity within the PM is responsible for the observed differences in cation depletion. By examining the variability of PM composition within each site, we have demonstrated that the deep till is laterally homogeneous within each sample site, at least within 10%, and systematic differences in PM composition by forest type do not exist (Table 2). Alternatively, these trends in depletion profiles may be explained by systematic vertical differences in the soils due to variations in the depositional or erosional history of the PM deposit. Considering the experimental design of this study (Fig. 1, Table 1), such a scenario seems unlikely. Given the similar geologic setting and history of the area, it does not seem plausible that a paleoerosional, depositional, or weathering surface existed across hydrologic boundaries in the park that is not manifest in surface topography or PM composition, but produces systematic depletion profile differences that are consistent with observed vegetative cover at different areas of multiple watersheds, elevations, and aspects (Fig. 1, Tables 1 and 2). Recent land use such as agriculture (plowing, soil amendments) or livestock (sheep grazing) could influence soil composition at this site, but in examination of soil horizonation and review of well-documented land use at the park, we found no evidence of such. The plantations were not sites of intensive agriculture; there is no documented use of soil amendments at the plantations and a plow horizon was never observed in these soils (National Park Service, 1994; Wilcke et al., 2000). The role of sheep grazing on soil development at this site is unknown, but it probably would not influence "deep soil" depletion profiles and is unlikely to cause systematic differences between areas occupied presently by different forests across the park. While we cannot with absolute certainty discard every possible alternative scenario to our interpretation, we do believe that we have presented by far the most likely explanation for our observed trends based on our experimental design, previous studies concerning vegetative modification of pedogenesis, and the land-use history of MBRNHP.

Long-Term Weathering Rates
Although we have discarded "alternative scenarios" that could produce our results, it remains necessary to address the perplexing conclusion that the observed differences in bulk soil composition have been attributed to differences in overstory vegetation during ~1% of the soil development period. While we cannot conclusively explain this discrepancy, we can provide what we feel is its most likely explanation, which is supported by integrated soil profile long-term weathering rate data, field-based observations, and documented land-use history (Table 4).


View this table:
[in this window]
[in a new window]

  		Table 4. Bulk soil profile long-term weathering rates of Marsh–Billings–Rockefeller National Historical Park soils.

 
Taylor and Blum (1995) found an empirical relationship between granitic parent material age and long-term integrated soil weathering rates. Long-term weathering rates became lower with increasing granitic parent material age as fresh surfaces or more reactive minerals in the deposit became depleted from the parent material, slowing the annual rate of soil chemical denudation (Taylor and Blum, 1995). Granitic material weathering rates are generally dominated by plagioclase, hornblende, and biotite dissolution (the relative importance of each depends on the modal mineral abundance of each in the deposit and phase reactivity [Garrels and Mackenzie, 1967]). All of these minerals are orders of magnitude less soluble than calcite (Drever, 1994a), which should be a mineral strongly influencing weathering rates and nutrient supplies at MBRNHP. Thus, one would expect to observe significantly higher bulk weathering rates at MBRNHP than those calculated for soils developed on similar-age deposits that are in purely granitic terraines. Interestingly, our data are within one standard error or less (Table 4) of an estimated long-term weathering rate for 12 000-yr-old soil developed on granitic material of 3.68 cmolc m–2 yr–1 as predicted by Taylor and Blum (1995). Considering the mineralogy of the PM at this site, this suggests that we have underestimated long-term weathering rates in MBRNHP, which we can explain by two observation-based mechanisms. First, before revegetation of the landscape at the turn of the century, substantial surficial erosion due to sheep farming probably occurred, removing much of the highly weathered surficial soil and causing our weathering rates to be lower than anticipated due to the removal of these highly weathered horizons. As a result, differences in soil composition since revegetation of the site would be magnified relative to the pre-sheep-farming soil development, particularly considering that a deeper, less weathered section of the soil would be exposed to the most active region of the weathering front during reforestation of the landscape. Second, in the sandy, permeable soils of MBRNHP, the PM that we have sampled, particularly its calcite composition, could have been altered since deposition, which would also lower integrated weathering rates and magnify relative differences associated with recent modifications of pedochemical or physical conditions by different forest types during the past 100 yr. Given weathering by groundwater or soil solutions at depth to some extent during the 12 000 yr that they were in place, calcite would be depleted relative to the original PM. Considering the solubility of calcite, the permeability of such sandy soils, and that chemical alteration is often observed at the glacial deposit–bedrock interface of till-mantled landscapes due to the hydraulic conductivity contrast between relatively permeable glacial deposits and im- or semipermeable bedrock, we find this to also be a plausible explanation (April et al., 1986, Kirkwood and Nesbitt, 1991). While we cannot conclusively determine the relative or absolute contribution of these factors to our results, both are observation based and would result in integrated weathering rates similar to granitic profiles and magnify relative differences in bulk chemistry associated with soil development under different forest types during the past 100 yr (Taylor and Blum, 1995). Thus, it is important to note that it is probable that the observed quantitative effects of forest type on mineral cation depletion are limited in application to the unique physical and chemical nature of MBRNHP parent material along with the site's land-use history, but the qualitative effects of different forest types found here should translate to other temperate forest soils.

Upon examination of integrated soil profile weathering rate data by forest type, it is not surprising to find that integrated rates do not differ by one standard error for any of the comparative sites. Because the standard error associated with bulk profile weathering rates incorporates those of each cation depletion factor in each horizon of the soil profile, and that of field-based density measurements, the relative standard error of long-term weathering rates is much larger than those associated with depletion factors of individual horizon labile base cation depletion factors (Eq. [1]). As a result, all integrated long-term weathering rates in conifer soils are within one standard error of their comparative northern hardwood soil (Table 4). With that said, it is interesting to note that long-term weathering rates at MBRNHP are consistently higher underneath northern hardwood forests relative to comparable conifer forests.

Descriptive Model of Chemical Depletion by Tree Species
The descriptive model presented here for species-specific effects on chemical weathering in forest soils is outlined in Fig. 4 . Our data from soil pH, depletion profiles and bulk long-term weathering rates in MBRNHP soils suggest several interesting trends concerning the mechanisms by which different tree species alter the weathering profile and deplete mineral macronutrients (Fig. 4). In surface soils (above 10 cm for this study), vegetation's primary influence on chemical weathering appears to be associated with chelating agents (or the relative ligand effect as used by Raulund-Rasmussen et al. [1998]) derived from OM decomposition and fine root exudation, which accelerate primary mineral dissolution in this area of the soil profile. Conifer species, particularly Norway spruce, deplete the upper soil profile to a greater extent than northern hardwood forests of the same age, particularly with regard to cations associated with soluble mineral phases. At depth, there appears to be a fundamental shift in the mechanism by which vegetation modifies the weathering regime, possibly due to the consumption of chelating ligands in weathering reactions in the upper component of the soil profile and the deep root networks associated with northern hardwood forests. In this section of the soil profile, root systems and associated nutrient uptake, proton release, and mechanical weathering appear to influence divalent base cation depletion by enhancing soil solution disequilibria with surrounding minerals and exposing fresh minerals to weathering agents. In the deeper soils, northern hardwood species appear to deplete relatively reactive mineral-phase-associated base cations in the soil more rapidly due to higher nutrient demands and more extensive deep rooting systems. Manifestation of deep soil processes in soil composition is less clear than that in surface horizons and could require more time to develop than surficial profile differences. Bulk ~12 000-yr-old soil profiles do not differ within one standard error in degree of total chemical denudation due to 100 yr of different forest-type development. Perhaps after more time or in more soluble PM, a more significant divergence of alteration would be observed. It is often assumed that conifer species more aggressively weather the soil profile than northern hardwood trees (c.f., Augusto et al., 2002), but our data suggest that no such difference exists on forest development time scales, and perhaps given more time, hardwoods could deplete the soil to a greater extent when other variables are controlled. It is also important to note that these effects appear to be primarily associated with the more reactive primary mineral phases in the soil system and become less evident when dealing with depletion profiles associated with resistant primary silicates like muscovite and orthoclase, as observed in the K profiles of MBRNHP (Fig. 3). These data reflect relatively long-term pedogenic processes in a managed ecosystem with uniform soil properties that allow us to examine the influence of forest type on chemical weathering on a scale that has not been examined before.


Figure 4
View larger version (39K):
[in this window]
[in a new window]

  		Fig. 4. Descriptive model for forest-type enhancement of chemical weathering in soil based on the soils and vegetation of Marsh–Billings–Rockefeller National Historical Park.

 

   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
We found that at MBRNHP, soils underlain by conifer and northern hardwood forests differ fundamentally in their distribution of base cation depletion, particularly with regard to divalent base cation mineral phases. Our observations indicate that in surface soils, aggressive weathering agents such as chelating ligands and H2CO3 generated from OM decomposition and fine root exudation actively deplete soils to a greater extent under conifer than northern hardwood forests. We observe, however, that at depth in the soil profile, northern hardwood forests deplete the soil to a slightly greater extent than conifer forests, probably through nutrient uptake by deep root networks and associated high annual nutrient demand, physical disturbance of deep soils by root networks, and proton release by roots during nutrient uptake. We also found that even when other soil-forming factors are controlled, on forest development time scales, bulk soil profiles do not differ significantly in degree of weathering, even though cation depletion distribution changes at this time scale. It should also be noted that, while a trace component of the PM and one that has little effect on bulk soil weathering rates, calcite dissolution must be an important weathering reaction in this system as a nutrient source and pH buffer. The quantitative differences observed here are probably only applicable to the specific chemical and physical nature of this PM and the land-use history of MBRNHP, but the descriptive model of cation depletion enhancement by forest type proposed here should be broadly applicable to soil systems under these forest types and at these time scales. This provides a substantial improvement in our understanding of forests' influence on the chemical denudation of the landscape and base cation redistribution within soil profiles. These conclusions indicate that in reforestation efforts, northern hardwood forests can be expected to deplete deep soil nutrient reserves to a greater extent than conifer forests if other variables are similar. Such effects on nutrient reserves should be considered when assessing land-use history and planning forest management. From a climate change perspective, if northern hardwood forests encroach on areas currently occupied by spruce or fir forests at higher elevations or more northern latitudes due to increasing temperatures, a change in CO2 consumption and sequestration by chemical weathering in soil is unlikely, although a change in C storage in OM is still possible due to a shift in litter quality. Our data indicate that such a change in forest type would, however, redistribute macronutrients from previously inaccessible deep mineral sources to surficial labile forms associated with northern hardwood OM.


   ACKNOWLEDGMENTS
 
We thank the National Park Service for permission to conduct work at MBRNHP. Funding from The Geological Society of America, The Vermont Geological Society, and Dartmouth College Earth Sciences Department and Environmental Studies Program to AWS is gratefully acknowledged. Insightful discussions with J. Kaste, C. Marts, R. April, C. Nezat, X. Feng, E. Miller, D. DeSimone, E. Postmentier, C Renshaw, and B. Dade improved this study. Laboratory and field assistance by N. Salant, S. Uhl, V. Solbert, T. Sullivan, and P. Zietz were also much appreciated.

Received for publication May 4, 2006.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Mineral MagHome page
M. Y. Andrews, J. J. Ague, and R. A. Berner
Weathering of soil minerals by angiosperm and gymnosperm trees
Mineralogical Magazine, February 1, 2008; 72(1): 11 - 14.
[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 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 Google Scholar
Google Scholar
Right arrow Articles by Schroth, A. W.
Right arrow Articles by Bostick, B. C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Schroth, A. W.
Right arrow Articles by Bostick, B. C.
Agricola
Right arrow Articles by Schroth, A. W.
Right arrow Articles by Bostick, B. C.
Related Collections
Right arrow Plant and Soil Interactions
Right arrow Biogeochemical Processes

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