Carbon, nitrogen, and phosphorus mineralization in northern wetlands.
Wetlands occupy [approximately]15% of the boreal and subarctic landscapes (Post et al. 1982, Gorham 1991), and these wetlands are predominantly peatlands (Matthews and Fung 1987, Aselmann and Crutzen 1989). Large biotic feedbacks to climate change may occur in northern wetlands (Harriss and Frolking 1992, Bridgham et al. 1995a). These are of particular concern because temperature changes from the "greenhouse effect" are predicted to be greater at higher latitudes (Houghton et al. 1992). Northern wetlands contain up to one-third of the world's soil C pool (Gorham 1991), and these wetlands are a significant atmospheric C[H.sub.4] source (Bartlett and Harriss 1993).
Despite the importance of northern wetlands in the global C cycle, they have often been viewed as homogeneous ecosystems in budget estimates (Post et al. 1982, Gorham 1991), or have been broken down into only two or three community types (Matthews and Fung 1987, Aselmann and Crutzen 1989, Bartlett and Harriss 1993). Yet, these wetlands are actually a diverse set of ecosystems that are structured by a variety of factors (Bridgham et al. 1996).
Within broad climatic limits, northern wetlands appear to be largely structured by pH/alkalinity and hydrology (Glaser 1992, Bridgham et al. 1996). Minerotrophic fens have inputs of ground water or runoff from the surrounding uplands, resulting in higher alkalinity and pH. They are typically vegetated with graminoids and may have a tree canopy composed of such species as Thuja occidentalis (northern white cedar), Larix laricina (tamarack), and Fraxinus nigra (black ash). Treed fens are often termed swamp forests. Ombrotrophic bogs receive only atmospheric inputs of water, basic cations, and nutrients, resulting in low pH ([less than or equal to]4) and alkalinity. Their vegetation is typified by Sphagnum mosses, low, ericaceous shrubs, and Picea mariana (black spruce).
The ombrotrophic-minerotrophic gradient that is described by differences in pH and alkalinity is frequently assumed to be coincident with a eutrophic - oligotrophic gradient of nutrient availability (Bridgham et al. 1996). There are several reasons to expect low nutrient availability in ombrotrophic peatlands. Because they receive only atmospheric nutrient inputs, ombrotrophic peatlands have inherently low nutrient input rates compared to more hydrologically open (minerotrophic) sites. N fixation rates are quite low in peatlands (Bowden 1987, Johnston 1991) and decrease along an ombrotrophic-minerotrophic gradient (Waughman and Bellamy 1980, Koerselman et al. 1989). The net result is that nutrient availability in ombrotrophic peatlands should largely depend on internal nutrient cycling within the ecosystem (Hemond 1983, Bowden 1987). Plants in bogs generally exhibit traits adapted for extreme nutrient deficiency, such as slow growth, evergreenness, sclerophylly and high concentrations of defensive compounds to reduce grazing losses, high nutrient retranslocation before leaf abscission, high efficiency in nutrient use, high root: shoot ratio, and low optimal concentrations of nutrients (Veerkamp et al. 1980, Shaver and Melillo 1984, Walbridge 1991, Bridgham et al. 1995b, 1996).
In apparent contradiction to these plant responses, in situ nutrient availability often has been found to be higher in bogs than in fens (Waughman 1980, Verhoeven et al. 1990, Koerselman et al. 1993). Fertilization studies have given conflicting results concerning the relative nutrient limitation along this gradient (Bridgham et al. 1996). Therefore, northern wetlands present an apparent paradox: geomorphic position and nutrient-conserving traits of resident plant species suggest that nutrient availability should be lower in bogs than in fens, but direct measurements of nutrient availability and fertilizer response trials suggest otherwise.
The resolution of this paradox may benefit from a comparison of mineralization kinetics of C, N, and P in sites along the ombrotrophic-minerotrophic gradient under similar environmental conditions. Accordingly, the objectives of this study were (1) to describe and compare the kinetics of C, N, and P mineralization under aerobic and anaerobic conditions along this gradient, and (2) to determine the relationship between C, N, and P cycling rates and soil quality.
Study sites and field sampling
Most sites were in areas of extensive pentland development in the lowlands of glacial lakes Agassiz, Upham, and Aitken in northwestern and northeastern Minnesota, United States [ILLUSTRATION FOR FIGURE 1 OMITTED]. We also had sites in more isolated wetlands at Isabella in the North Shore Highlands along Lake Superior and in morainal areas in Marcell Experimental Station and Voyagers National Park, Minnesota. All sites (N = 16) were located between 46 [degrees] and 49 [degrees] N.
The bog sites (N = 5) were vegetated with black spruce (Picea mariana), Sphagnum spp., and ericaceous shrubs, and were located at Pine Island, Red Lake, Toivola, Arlberg, and Ash River near Voyagers National Park. Fen sites were dominated by sedges (Carex spp.) and were located at Red Lake, McGregor, Alborn, and Marcel. The fens were separated into acidic fens ("poor" fens, N = 2) with a mean soil pH of 4.0-4.2, and intermediate fens with a mean soil pH of 4.9-5.6 (N = 2). Vegetation differences supported this distinction between acidic and intermediate fens. Forested swamp sites included cedar swamps near Meadowlands, Isabella, and Ash River (N = 3), and tamarack swamps near Meadowlands and Ash River (N = 2). Two beaver meadows were also sampled in Voyagers National Park, with dominant vegetation of Canada blue joint (Calamagrostis canadensis), wool grass (Scirpus cyperinus), and Carex spp.
We took five 0-25 cm deep cores from hollows in each site during September and October 1991. The soil surface was set at the bottom of the green moss layer when present, 3-5 cm below the capitula. In the beaver meadows, we sampled the surface histic epipedon (i.e., an organic soil horizon saturated with water for some period of the year), 8-21 cm deep, over the mineral substratum. A known volume of soil was carefully removed with a bulb planter and a serrated knife, sealed in polyethylene bags, and kept at 4 [degrees] C prior to processing. As is typical for midcontinental wetlands at this time of year, the bogs were relatively dry, with a water table level below [approximately] 15 cm depth. Cores from the beaver meadows were taken across transects that varied considerably in wetness. The hollows in the fens and swamp forests were mostly flooded to the surface.
Without a priori knowledge of the most successful predictor of mineralization rates, we used a variety of approaches to characterize "soil quality" (Table 1). Proximate C fractions are determined by a sequential extraction with nonpolar, polar, and acidic solvents (Allen 1989, Ryan et al. 1990). Soil C fractions (humin, humic acid, and fulvic acid) are defined on the basis of sequential dissolution in acidic and alkaline extracts (Schnitzer 1982). The degree of decomposition of peats is generally determined by fiber content and the pyrophosphate color index (Day et al. 1979), which is the basis of histosol classification in the United States (Soil Survey Staff 1990) and much of the rest of the world. Other chemical and physical variables, such as pH, alkalinity, and ash content, are also very important soil parameters defining different wetland soils (Clymo 1983).
Samples were dried at 70 [degrees] C and then ashed for 4 h at 500 [degrees] C to quantify total ash content, while total C and N were determined by combustion on a Leco CHN analyzer (Leco Corporation, St. Joseph, Michigan, USA). A [H.sub.2]S[O.sub.4]-[H.sub.2][O.sub.2] digestion was used for total P estimation (Lowther 1980). pH was measured at each leaching interval (see Methods: Mineralization potentials), but we used the 2-wk pH for the soil characterization study.
Proximate C fractions were estimated with a sequential extraction of dried, ground sample with (1) methylene chloride in a sonicating bath, (2) boiling water, and (3) 72% [H.sub.2]S[O.sub.4] at 30 [degrees] C for 1 h, followed by 3% [H.sub.2]S[O.sub.4] in an autoclave at 250 [degrees] C for 1 h (Ryan et al. 1990). After each step, the sample was dried at 70 [degrees] C and weighed to determine mass loss in the extractant. Gravimetric loss in methylene chloride represents the nonpolar-extractable fraction (fats, waxes, and oils), whereas the acid-soluble fraction represents holocellulose. Additionally, polyphenol and carbohydrate concentrations were determined in the hot water and [H.sub.2]S[O.sub.4]-soluble fractions by colorimetric analysis. Polyphenols were analyzed with the Folin-Denis assay with tannic acid as a standard, and sugars were analyzed with the phenol-sulfuric acid assay with D-glucose as a standard (Allen 1989). The residue in the [H.sub.2]S[O.sub.4] was designated "lignin," although it was recognized as representing a poorly defined, chemically resistant fraction likely to be composed of humic substances. All proximate C fractions were corrected for ash content and are presented on an ash-free basis. Samples were re-ashed after the [H.sub.2]S[O.sub.4] extraction step to account for ash that dissolved in the acidic solvent.
Unrubbed fiber content was determined gravimetrically by dispersion with Calgon detergent and washing through a 100-mesh screen (Day et al. 1979). Rubbed fiber content was determined by gently rubbing the samples between the fingers during washing. The pyrophosphate-soluble organic matter index was determined by extraction of soil samples with 0.025 mol/L sodium pyrophosphate and comparison of the color on chromatographic paper against a 10YR color chart (Day et al. 1979). The von Post scale is a common qualitative method of describing the degree of peat decomposition in the field, based upon visual inspection of soil structure, the color of expressed water, the amount of peat lost by squeezing, and the consistency and color of peat retained in the hand (Clymo 1983). This method has 10 decomposition classes, ranging from completely undecomposed peat (H 1) to completely humified peat (H10).
Humic acid, fulvic acids, and insoluble humin were estimated gravimetrically using sequential extracts of dried, ground soil (Day et al. 1979, Schnitzer 1982). The samples were pretreated with a dilute HCl-HF solution prior to extraction with 0.1 mol/L NaOH under an [N.sub.2] atmosphere, and then the pH of the solution was adjusted to pH 2 with 1 mol/L HCl. The alkali-insoluble fraction represents humin, the acid-insoluble fraction represents humic acids, and the acid-soluble fraction represents fulvic acids. The sum of the three fractions was significantly [greater than] 100% of the original organic matter for some samples, despite our pretreatment with dilute HCl-HF solution. This appeared to be due to dissolution of mineral matter in the pH 2 treatment, making the apparent fulvic acid fraction too high.
Extractable Ca, K, Mg, and Na were measured by [TABULAR DATA FOR TABLE 1 OMITTED] atomic absorption analysis of 1 mol/L N[H.sub.4]OAC extracts of fresh soil (Thomas 1982). Total exchangeable bases (EM) were calculated as the sum of the concentrations of Ca, K, Mg, and Na. Exchangeable acidity (EA) was determined with the barium chloride-triethanolamine method (Thomas 1982). Cation exchange capacity was estimated as the sum of total exchangeable bases and exchangeable acidity (Chapman 1965). Percent base saturation was calculated as (EM/(EM + EA)) x 100. Fe and Al were extracted with ammonium oxalate and were determined by atomic absorption (Jackson et al. 1986). Subsamples of soil were extracted in dilute NH4F-HCl at the beginning and end of the incubation period, and the extracted P was analyzed colorimetrically using the ammonium molybdate-stannous chloride method (Olsen and Sommers 1982). Additionally, the initial (week 0) 0.01 mol/L Ca[Cl.sub.2]-extractable N (N[[H.sub.4].sup.+] + N[[O.sub.3].sup.-]) and P in the mineralization study were included as soil quality parameters.
To obtain reliable bulk density estimates, we returned to each site during the autumn of 1993 and obtained five intact 10 x 25 cm soil cores using PVC pipe sections. These cores were later dried, weighed, and used to estimate bulk density.
Separate subsamples of soil were incubated aerobically and anaerobically at 30 [degrees] C for 59 wk in the laboratory to determine the kinetics of C, N, and P mineralization. This is an optimal temperature for many microbial processes, and it was used to maximize mineralization rates so that the kinetic models could be fit to the data by the end of the incubation. We outline the method here, but it is presented in full detail in Updegraff et al. (1994, 1995).
Samples were hand-mixed, removing large root masses and woody fragments, in a glovebox under a continuous flow of [N.sub.2]. Field-moist soil samples were mixed with 20-50 g acid-washed Ottawa sand to facilitate drainage during leaching, and were placed in 150-mL polypropylene Falcon filter units (model 7102, Becton, Dickinson and Company, Cockeysville, Maryland, USA).
In the aerobic incubations, samples were exposed to ambient air except during gas flux measurement. The samples were leached at 2, 4, 8, 12, 16, 22, 31, 40, 50, and 59 wk with 0.01 mol/L Ca[Cl.sub.2], and the leachate was analyzed for N[[O.sub.3].sup.-], N[[H.sub.4].sup.+], and P[[O.sub.4].sup.-3] concentrations to determine N and P mineralization rates over each time interval. The protocol followed Updegraff et al. (1995), except that no additional nutrients were added to the soils.
A similar protocol was follow for anaerobic incubations, except that samples in the Falcon filter units were immersed in water-filled 500-mL Mason jars. Nutrient concentrations were determined in both the leachate and in the incubated water in the Mason jar at each leaching period. N concentrations were determined on a Lachat Autoanalyzer using cadmium reduction for N[[O.sub.3].sup.-] and a salicylate-nitroprusside method for N[[H.sub.4].sup.+] (Lachat Instruments, Mequon, Wisconsin, USA). P concentrations were determined manually with the absorbic acid colorimetric method (Olsen and Sommers et al. 1982).
Results were expressed both as milligrams of C, N, and P mineralized per gram of total soil C, N, and P, respectively, and as milligrams of C, N, and P mineralized per cubic centimeter of soil. The first value (e.g., milligrams of N mineralized per gram of N per day) represents a turnover rate of the nutrient or C pool and, thus, is a measure of the lability of that pool. In contrast, the second value (e.g., milligrams of N mineralized per cubic centimeter per day) represents mineralization per volume of soil ([N.sub.min], [P.sub.min], or [C.sub.min]); because organisms exploit a volume rather than a mass of soil, it is a measure of soil nutrient availability (Barko and Smart 1986). Conversion between turnover and mineralization rates depends on soil bulk density and total soil nutrient content.
Gas flux measurements
[C.sub.min] was estimated by the flux of C[O.sub.2] for aerobic samples and C[O.sub.2] plus C[H.sub.4] for anaerobic samples (Updegraff et al. 1995). In the aerobic incubations, the headspace was purged with C[O.sub.2]-free air at the midpoint between consecutive leachings, sealed, and sampled with a syringe 24 h later. The C[O.sub.2] flux rate was extrapolated to the entire period between consecutive leachings. In anaerobic incubations, dissolved gases in the water equilibrated with a 100-[cm.sup.3] headspace at the end of each leaching period. The accumulated C[O.sub.2] and C[H.sub.4] were divided by the elapsed time since the previous leaching (2-11 wk) to obtain a mean daily flux rate.
C[O.sub.2] and C[H.sub.4] were determined by gas chromatography on a 2 m X 3 mm Porapak-Q column with a thermal conductivity detector (TCD) and flame ionization detector (FID), respectively. Dissolved C[H.sub.4] and C[O.sub.2] were calculated with Henry's Law, adjusting for solubility, temperature, and pH (Stumm and Morgan 1981).
Mineralization kinetics were fit with a two-compartment model (Updegraff et al. 1995):
[X.sub.t] = [X.sub.0](1 - [e.sup.[-k.sub.x]t]) + (TX - [X.sub.0])(1 - [e.sup.[-h.sub.x]t] (1)
where X represents C, N, or P; [X.sub.t] is the cumulative amount of C or nutrient released at time t; and [X.sub.0] is the amount of potentially mineralizable C or nutrient with an instantaneous release rate [k.sub.x]. [X.sub.o] and [k.sub.x] refer to the pool size and mineralization rate, respectively, of a labile organic fraction. TX is total soil C, N, or P, and (TX - [X.sub.0]) is the amount in a recalcitrant pool with instantaneous release rate [h.sub.x]. Units are milligrams of C, N, or P mineralized per gram of TC, TN, or TP for [X.sub.t] and [X.sub.0], and [week.sup.-1] for the instantaneous release rates [k.sub.x] and [h.sub.x]. The model was fit to the data using a nonlinear estimation procedure in Systat (Wilkinson et al. 1992), as described in detail in Updegraff et al. (1995). We also calculated the time for 50% of the each pool to mineralize ([t.sub.1/2]) in an exponential decay model as:
[t.sub.1/2] = 1/k or h ln 2. (2)
Many of the soil descriptor variables (N = 39, Table 1) were highly autocorrelated, had non-normal distributions, and had unequal variances among community types. Additionally, the large data set made it necessary to define overall relationships between groups of soil quality variables and mineralization rates. Therefore, relationships between soil descriptor variables were summarized by a principal components analysis (PCA) with varimax rotation, using the means from each site (Wilkinson et al. 1992). All data were standardized to a Z distribution and were then multiplied by the factor coefficients to produce factor scores for each sample. The PCA factor scores of the component axes were then used as explanatory independent variables in multiple regressions, with mineralization rates as the dependent variable (Gorsuch 1983).
For ANOVAs, sites (N = 16) were grouped according to dominant vegetation type and, for the fens, by pH (see Study sites), with the five cores per site nested within each community type to account for intrasite variation. Data were log-transformed, if necessary, to normalize the distribution. Two-way ANOVAs were run to determine the significance of community type and aeration status (aerobic or anaerobic) on cumulative mineralization. To determine relative differences among community types, one-way ANOVAs were run separately for the aerobic and anaerobic incubations, and Tukey's test was used to determine pairwise differences among community types (Wilkinson et al. 1992).
The PCA of soil characteristics supported our grouping of sites into community types based upon dominant plant composition, except that cedar and tamarack swamps clustered together [ILLUSTRATION FOR FIGURE 2 OMITTED]. The first, second, and third axes explained 30%, 28%, and 17% of the total variation in soil characteristics, respectively. Axis 1 separated the bogs from the other wetlands, reflecting their lower extractable Fe and Al, total N, and total P content, and lesser degree of humification and decomposition (i.e., higher content of structural C and easily soluble compounds). The second axis separated the bogs and fens from the other sites because of their lower alkalinity, percentage of lignin, and extractable N content, and their higher percentage of cellulose. The beaver meadows were strongly separated from the other sites on axis 3. They had a shallow histic epipedon, 821 cm deep, with high ash content that would not classify them as peatlands (cf. Table 1). Their histic epipedon had higher total P, fulvic acid, and humic acid content, and lower percentage of C, percentage of humin, and C:P ratio.
Total [C.sub.min]. (i.e., C[O.sub.2] plus C[H.sub.4]), expressed either as a turnover rate or per cubic centimeter of soil, differed among community type, aeration status, and their interaction term (P [less than] 0.001). Despite the significant interaction, the aerobic: anaerobic ratio varied only from 4.5 in the tamarack swamps to 7.7 in the bogs [ILLUSTRATION FOR FIGURE 3 OMITTED]. Aerobically, the beaver meadow and bog sites had the highest C turnover; anaerobically, the beaver meadows alone had the highest C turnover. When [C.sub.min] was expressed per cubic centimeter of soil, aerobic mineralization was greatest in the beaver meadows and cedar swamps and lowest in the bogs and acidic fens [ILLUSTRATION FOR FIGURE 3B OMITTED]. Anaerobic results for [C.sub.min] per cubic centimeter of soil were similar, with the beaver meadows highest and the bogs and acidic fens lowest.
C[H.sub.4] production, expressed either as per gram of C or per cubic centimeter of soil, differed significantly among sites (P [less than] 0.001), and was greatest in the beaver meadows and intermediate fens and lowest in the acidic fens and bogs [ILLUSTRATION FOR FIGURE 4 OMITTED]. The proportion of anaerobic [C.sub.min] as C[H.sub.4] varied extensively among sites (bogs 0.5%; acidic fens 2%; intermediate fens 10%; cedar swamps 5%; tamarack swamps 5%; meadows 12%).
Cumulative N turnover (milligrams of N mineralized per gram of TN) differed among community types, aeration status (aerobic vs. anaerobic), and their interaction term (all P [less than] 0.001). Aerobic N turnover was highest in the bogs and acidic fens and lowest in the cedar swamps and intermediate fens [ILLUSTRATION FOR FIGURE 5A OMITTED]. In contrast, anaerobic N turnover was highest in the beaver meadows and lowest in the intermediate fens. The significant interaction between community type and aeration status is readily evident in Fig. 5A. Aerobic [N.sub.min] was 2.6 times greater than anaerobic [N.sub.min] in the bog, but only 1.1 times greater in the beaver meadow.
Cumulative [N.sub.min] per cubic centimeter also differed among community types and aeration status (P [less than] 0.001), but the interaction term was not significant. [N.sub.min] per cubic centimeter was greatest in the beaver meadows and lowest in the bogs and acidic fens [ILLUSTRATION FOR FIGURE 5B OMITTED]. The importance of considering mineralization both as a turnover rate and per unit soil volume is apparent from the contrast of the relative magnitudes of [N.sub.min] in Fig. 5A, B.
There were dramatic differences among sites in the importance and temporal dynamics of nitrification under aerobic conditions. Nitrate was [less than]5% of total inorganic N in the initial 0.01 mol/L Ca[Cl.sub.2] extract for all sites, except for the beaver meadows, where it was 54% of initial extractable inorganic N. The percentage 98%). Nitrification as a percentage of total aerobic [N.sub.min] increased in importance in all soils over time, accounting for [greater than]70% of [N.sub.min] by the end of the incubation for all but the bogs [ILLUSTRATION FOR FIGURE 6 OMITTED]. Additionally, more minerotrophic soils began to nitrify much sooner than more ombrotrophic soils. In contrast, rates of aerobic [[NH.sub.4].sup.+] mineralization decreased rapidly with time, so absolute rates of nitrification were either constant or increased slowly with time (data not shown).
Cumulative net [P.sub.min], both as a turnover rate and per cubic centimeter of soil, differed among communities, aeration status, and their interaction term (P [less than] 0.001). In contrast to [N.sub.min] and [C.sub.min], [P.sub.min] was generally higher under anaerobic than aerobic conditions [ILLUSTRATION FOR FIGURE 7 OMITTED]. The aerobic:anaerobic ratio varied from 0.28 in the tamarack swamps to 1.12 in the acidic fens. Under both aerobic and anaerobic conditions, the highest P turnover rates were in the bogs and acidic fens, with the lowest rates in the intermediate fens and beaver meadows [ILLUSTRATION FOR FIGURE 7A OMITTED]. Similarly, aerobic [P.sub.min]. per cubic centimeter was highest in the bogs and lowest in the intermediate fens [ILLUSTRATION FOR FIGURE 7B OMITTED]. Under anaerobic conditions, [P.sub.min] per cubic centimeter was greatest in the tamarack swamps and lowest in the acidic and intermediate fens.
We summarized mineralization differences among sites in a PCA, separately considering turnover [ILLUSTRATION FOR FIGURE 8A OMITTED] and mineralization rates per cubic centimeter [ILLUSTRATION FOR FIGURE 8B OMITTED]. The first and second PCA axes explained 55% and [TABULAR DATA FOR TABLE 2 OMITTED] 31% of the variance in the turnover rates, respectively [ILLUSTRATION FOR FIGURE 8A OMITTED]. The first axis separated out the bogs and acidic fens from the more minerotrophic sites, with the bogs and acidic fens having higher turnover rates of aerobic and anaerobic P, aerobic N, and aerobic C. The second axis separated out the beaver meadows from the other sites, with the beaver meadows having higher turnover of anaerobic N, anaerobic C, and aerobic C.
Although the bog and acidic fen soils had low total N and P content and C:N and C:P ratios (Table 1), the P that was there turned over relatively rapidly under both aerobic and anaerobic conditions. Similarly, under aerobic conditions, the N and C pools turned over relatively rapidly in the bogs and acidic fens. In contrast, under anaerobic conditions, the beaver meadows with their low soil C:N ratios were distinguished by having higher anaerobic N and C turnover.
The first and second PCA axes explained 63% and 24% of the variance, respectively, in C and nutrient mineralization per cubic centimeter [ILLUSTRATION FOR FIGURE 8B OMITTED]. The first axis was described by higher aerobic and anaerobic [C.sub.min] and [N.sub.min], whereas the second axis was described by higher aerobic and anaerobic [P.sub.min]. Thus, in terms of nutrient mineralization per cubic centimeter, P appeared to have an independent trajectory relative to N and C. The first axis followed the ombrotrophic-minerotrophic gradient among the sites. Along the second axis, bogs and some of the cedar and tamarack swamp [TABULAR DATA FOR TABLE 3 OMITTED] forests had higher [P.sub.min], with low values in the acidic and intermediate fens and in the beaver meadows.
The two-pool kinetic model (Eq. 1) was fit to the data for N, P, and C turnover under aerobic (Table 2) and anaerobic (Table 3) conditions. Aerobically, the labile C pool ([C.sub.0]) was about twice as great in the bogs as in the other sites (Table 2). [C.sub.0] accounted for 4.4% to 11.4% of total soil C, with time to 50% mineralization ([t.sub.1/2]) ranging from 4.3 to 7.7 wk. Anaerobically, Co was greatest in the beaver meadows and intermediate fens, but it accounted for [less than]3% of total soil C (Table 3) and had a [t.sub.1/2] ranging from 5.3 to 8.7 wk. The aerobic: anaerobic ratio for [C.sub.0] ranged from 10.2 in the bogs to 2.1 in the intermediate fens, and generally decreased from ombrotrophic to minerotrophic sites. Under aerobic conditions, [t.sub.1/2] for recalcitrant soil C ([h.sub.c]) ranged from 5.6 to 9.5 yr, whereas under anaerobic conditions it was [approximately]67 yr.
Aerobically, [N.sub.0] was substantially larger in the bogs and acidic fens than in the more minerotrophic sites. For [N.sub.0], [t.sub.1/2] was 1.3 wk in the tamarack swamps, 11.6 wk in the cedar swamps, and 5-6 wk in the other sites. Recalcitrant N dominated the total soil N pool ([greater than]91%) in all sites, and it mineralized at very slow rates, with a [t.sub.1/2] of [approximately]21 yr.
Anaerobically, [N.sub.0] was highest in the beaver meadow [TABULAR DATA FOR TABLE 4 OMITTED] soils, but otherwise it did not follow any obvious trend across the ombrotrophic-minerotrophic gradient (Table 3). The aerobic:anaerobic ratio for [N.sub.0] in most cases was [approximately]2, ranging from 2.57 in the acidic fens to 0.24 in the tamarack swamps. [N.sub.0] mineralized faster in the more ombrotrophic sites, with [t.sub.1/2] increasing from [approximately]1.5 wk in the bogs and fens, to 2.4 wk in the swamps, to 5.0 wk in the beaver meadows. Thus, although [N.sub.0] was largest in the beaver meadows, it mineralized at a slower rate. For recalcitrant N, [t.sub.1/2] was [approximately]44 yr for all soils, twice as long as under aerobic conditions.
[P.sub.0] was much greater in the bogs and acidic fens than in minerotrophic sites (Tables 2 and 3), with maximum differences under aerobic conditions. The contribution of [P.sub.0] to total P ranged from 33% in anaerobic bog soils to [less than] 1% in the aerobic swamp and beaver meadow soils. The aerobic: anaerobic [P.sub.0] ratio was lower in more minerotrophic sites (0.61-0.89 in bogs and fens, 0.060.26 in beaver meadows and swamps). Aerobically, [t.sub.1/2] for [P.sub.0] ranged from 1.3 wk in the beaver meadows to 8.7 wk in the bogs; anaerobically, this trend was reversed, being 1.1 wk in the bogs and 8.7 wk in the beaver meadows. For recalcitrant soil P, [t.sub.1/2] ranged from 7 yr in bogs under aerobic conditions to [greater than]1300 yr in more minerotrophic sites.
Acid-fluoride extractable P
We determined acid-F extractable P (as milligrams of P per gram of total P) at the beginning and end of the 59-wk incubation (Table 4), with acid-F removing easily acid-soluble (or "labile") mineral P, primarily Ca-, Al-, and Fe-phosphates (Olsen and Sommers 1982). We then compared initial acid-F P to readily mineralizable P ([P.sub.0]) calculated from the kinetic equation. Aerobically, acid-F P increased in importance in more minerotrophic sites, constituting [less than]4% of [P.sub.0] in the bogs, 25-30% in the fens, and 65-130% in the swamps and beaver meadows. The ratio of final: initial acid-F P was [greater than] 1, except in the acidic fens, indicating that the labile mineral fraction was a net sink for P over the incubation. Anaerobically, acid-F P accounted for only 2-22% of [P.sub.0]. The ratio of final: initial acid-F P was [less than] 1, indicating a net release of P from the labile mineral fraction, in all but the beaver meadows.
Relationships between substrate quality and mineralization
We used PCAs to explore the relationships between substrate quality and mineralization (Gorsuch 1983). The PCA that best separated the sites was not necessarily the best set of variables to predict mineralization. Additionally, few studies will have our detailed set of soil characteristics, and a simpler set of variables would be useful to predict mineralization in other sites. Hence, we examined the correlations of PCA factor scores from various subsets of soil characteristics relative to mineralization, and present those that had the highest [R.sup.2] values or that were of particular ecological interest (Tables 5 and 6). These five sets of principal components were: (1) the complete data set of soil characteristics, (2) "nutrient variables" (percent ash, pH, and percent total C, N, and P), (3) "field variables" (percent ash, percent rubbed fiber, Von Post scale, and pH), (4) "C quality variables" (percentages of nonpolar extractables, water-soluble compounds, acid-soluble compounds, lignin, soluble phenolics, water-soluble carbohydrates, and acid-soluble carbohydrates; and the lignin: cellulose index), and (5) a set of "optimal variables" (percentages of total N, nonpolar extractables, water-soluble compounds, ash, and rubbed fiber; Von Post scale; and pH). The "field variables" can be measured easily with minimum expense and time and, hence, are included in many databases on northern wetlands. The set of "optimal variables" was generally found to yield the highest [R.sup.2] through trial and error.
The PCA scores explained a large amount of the variation in aerobic N and P turnover, anaerobic P turnover, and C[H.sub.4] production (Table 5). [R.sup.2] values were often [greater than]0.75. Aerobic C turnover, anaerobic N turnover, and anaerobic C[O.sub.2] turnover, in general, were poorly predicted by any group of PCA axes, except that anaerobic C[O.sub.2] production was reasonably well correlated with the field variables ([R.sup.2] = 0.69).
Mineralization variables (aerobic C, anaerobic N, anaerobic C) that were poorly predicted as a turnover rate (Table 5) by the PCA axes were predicted much better as mineralization per cubic centimeter (Table 6). In contrast, [P.sub.min] per cubic centimeter was poorly predicted [TABULAR DATA FOR TABLE 5 OMITTED] by the PCAs, except for the full suite of variables for anaerobic [P.sub.min]. Differences in bulk density among the soils predicted mineralization per cubic centimeter well ([r.sup.2] = 0.75 to 0.89) for all but [P.sub.min]. Alkalinity variables (pH, exchangeable bases, cation exchange capacity, percent of base saturation, exchangeable Na, Mg, and Ca) generally had a high positive correlation to aerobic C, anaerobic P, and anaerobic C[O.sub.2] mineralization per cubic centimeter.
Mineralization potentials have been employed extensively as a relative measure of nutrient availability and soil C turnover (Keeney and Nelson 1982, Binkley and Hart 1989). Our potential mineralization rates are not necessarily equal to in situ mineralization rates, because of disturbance of soil structure, high incubation [TABULAR DATA FOR TABLE 6 OMITTED] temperatures, the long incubation period, and constant aeration status. However, standard incubation conditions are essential in comparisons of inherent nutrient availability and carbon lability in soils from diverse sites with varied environmental conditions, the first objective of this study. Standard incubation conditions are also necessary to compare soil quality characteristics with mineralization rates, the second objective of this study.
Mineralization potentials are well correlated with plant uptake of nutrients, and repeated leaching techniques in long-term incubations are a standard, highly successful method of determining soil nutrient availability in agriculture (Keeney and Nelson 1982). These methods have also been used successfully in comparisons of nutrient mineralization and availability among diverse natural ecosystems (Nadelhoffer et al. 1991, MacDonald et al. 1995). The readily mineralizable nutrient and carbon pools ([N.sub.0], [C.sub.0], and [P.sub.0]) derived from the kinetic model (Eq. 1) are thought to be particularly important ecosystem parameters and can only be derived from long-term incubations with repeated leachings (Stanford and Smith 1972, Keeney and Nelson 1982).
In order to attain stable solutions to the two-pool kinetic model, it was necessary to extend the incubations to 59 wk, during which time the microbial community would be expected to change dramatically. Nevertheless, relative differences in mineralization rates among sites were generally consistent over the incubation, with good correlations between 2-wk and 59-wk cumulative mineralization (r [greater than] 0.71, P [less than] 0.002). The exception was for aerobic [P.sub.min] (r = 0.40, P = 0.13), and the poor correlation is more likely to be due to the temporal dynamics of geochemical P sorption than to changes in the microbial community.
Net [P.sub.min] is rarely measured in soils because of difficulties in interpreting the results in terms of geochemical P sorption, although net [P.sub.min] has proven useful in understanding P availability in peatlands (Verhoeven et al. 1990, Walbridge 1991). These difficulties may be compounded by using a weak salt leachate, such as the 0.01 mol/L Ca[Cl.sub.2] used in this study, although a stronger leachate would have inhibited microbial activity. We believe that our net [P.sub.min] values are a valid relative comparison among sites because (1) plant roots must naturally compete with geochemical sorption sites for P, and (2) we have interpreted our results in terms of a strong P extractant (acid-F P) at the beginning and end of the incubation to account for the increase or decrease in geochemically adsorbed P.
Community type, aeration status, and their interaction were all significant in explaining the variation in C turnover and mineralization per cubic centimeter among wetland soils. Nevertheless, differences in total C turnover among community types were relatively small (1.4 times for both aerobic and anaerobic conditions). Differences in [C.sub.min] per cubic centimeter among community types were larger (7.8 times for aerobic and 9.5 times for anaerobic conditions), and generally increased along the ombrotrophic-minerotrophic gradient. Bulk density alone explained from 75% to 83% of the variation in [C.sub.min] per cubic centimeter, suggesting that substrate density is a major control over mineralization in organic soils.
Litter bag studies have shown very low rates of decomposition for Sphagnum (Clymo and Hayward 1982), the dominant moss genus of ombrotrophic peatlands and the primary constituent of bog peats. Cellulose decay rates are also lower in bogs than in fens (Verhoeven et al. 1990), suggesting inhibition of decomposition by soil conditions, e.g., low pH, in bogs. Yet, we found relatively high nutrient and C turnover rates in our bog sites. Verhoeven et al. (1990) also found higher mineralization of N and P in bogs than in fens. They hypothesized that this discrepancy between litter bag studies and their mineralization data was due to the rapid breakdown of the nutrient-rich protoplasm of Sphagnum, yielding high nutrient and [C.sub.min] rates over short incubation periods. The majority of the plant is composed of recalcitrant cell wall material (Clymo and Hayward 1982), and it would presumably decay slowly.
We offer an alternative hypothesis, in that litter bag studies are typically done with fresh plant material, whereas our study and that of Verhoeven et al. (1990) used surface soil organic matter. Because Sphagnum litter decomposes slowly, it is left in a relatively undecomposed state in the surface soil, yielding low bulk densities and high fiber content (Table 1). Organic matter derived from more labile litter would be expected to become more decomposed and humified over time, as we found with soils from our minerotrophic sites. The net result is that soil organic matter attains similar rates of C turnover with time [ILLUSTRATION FOR FIGURE 3 OMITTED], no matter what type of starting plant litter is involved. In contrast to relatively similar C turnover rates among wetland types, aerobic nutrient turnover is much higher in the less decomposed bog and acidic fen peats [ILLUSTRATION FOR FIGURES 5 AND 7 OMITTED]).
The relative proportion of anaerobic [C.sub.min] as C[O.sub.2] and C[H.sub.4] varied significantly among community types, whether expressed as a turnover rate or per cubic centimeter. C[H.sub.4] production was extremely low in the bogs and acidic fens, and increased dramatically under more minerotrophic conditions [ILLUSTRATION FOR FIGURE 4 OMITTED]. Despite continuous anaerobic conditions for 59 wk at 30 [degrees] C, C[H.sub.4] accounted for only 0.5% to 12.1% of total anaerobic [C.sub.min], with this percentage increasing along the ombrotrophic-minerotrophic gradient. Similar results have been found previously in peatlands (Moore and Knowles 1989, Bridgham and Richardson 1992, Valentine et al. 1994, Updegraff et al. 1995). Thus, although nutrient and C turnover as C[O.sub.2] are relatively high in ombrotrophic peats under both aerobic and anaerobic conditions, C[H.sub.4] production is extremely low. Nevertheless, the production of C[H.sub.4]-C per gram of TC was significantly correlated with anaerobic C[O.sub.2]-C per gram of TC ([r.sup.2] = 0.46, P = 0.004), but not with aerobic C[O.sub.2]-C per gram of TC ([r.sup.2] = 0.00, P [greater than] 0.05).
We suggest that variable C quality among soils has different effects on methanogens and the microbial populations that produce C[O.sub.2]. In nonsaline environments, C[H.sub.4] is produced almost solely from acetate and [H.sub.2]-C[O.sub.2], with both acetate and [H.sub.2] being produced as end products of the fermentative pathway (Conrad 1989), whereas most heterotrophic microbial reactions produce C[O.sub.2]. Factors such as substrate quality and temperature (Bridgham et al. 1995a, Updegraff et al. 1995) will affect the flow of C through the fermentation pathway with its syntrophic microbial consortia. In turn, this will determine the ratio of C[O.sub.2] to C[H.sub.4] produced anaerobically. The absence of any relationship between aerobic C[O.sub.2] production and C[H.sub.4] production suggests that different C pools ultimately serve as substrates for the two groups of bacteria, with the effect of C quality on methanogens modulated through the fermentative microbial consortium. Additionally, our results support the hypothesis that, in studies in which high C[H.sub.4] fluxes have been found in bogs (Dise et al. 1993, Moore 1994), the organic C driving the reaction was probably derived from a labile pool, such as root exudates, and not from bulk peat. Detailed experiments with radio-labeled organic compounds would be necessary to substantiate these hypotheses.
N and P mineralization
As with C mineralization, community type, aeration status, and their interaction were all highly significant in explaining the variation in N and P turnover and mineralization per cubic centimeter in the broad range of wetland soils tested in this study. Thus, soil nutrient mineralization rates vary among northern wetland communities in a complex interaction with site hydrology and the resultant oxidation-reduction status of the soil.
Our mineralization results set up an interesting test of the long-held assumption that more ombrotrophic wetlands are also more nutrient deficient, i.e., ombrotrophic is synonymous with oligotrophic (Moore and Bellamy 1974, Maimer 1993, cf. Bridgham et al. 1996). Although ombrotrophic and minerotrophic wetlands are strictly defined based on hydrology, the "trophic" status of wetlands is often inferred from soil or water chemistry and characteristic plant communities (Bridgham et al. 1996). Detailed hydrologic measurements of all sites were beyond the scope of this study, but the soil pH, conductivity, basic cation concentrations, percent base saturation, fiber content, and plant communities (Table 1) suggest that our sites occur across an ombrotrophic-minerotrophic gradient. In relative order of increasing minerotrophy are the bogs, acidic fens, intermediate fens, swamp forests (cedar and tamarack), and beaver meadows. The beaver meadows have a histic epipedon, but the soils would not be classified as Histosols (i.e., peats) because of their high ash content and limited organic soil depth (Soil Survey Staff 1990).
Our data do not support the commonly held supposition that nutrient turnover and availability are inherently lower in more ombrotrophic sites. Additionally, it appears that N and P have quite different dynamics along this gradient. The small total nutrient pools in ombrotrophic sites were relatively labile and turned over quickly compared to minerotrophic sites, particularly under aerobic conditions for N [ILLUSTRATION FOR FIGURE 5A OMITTED] and under both aerobic and anaerobic conditions for P [ILLUSTRATION FOR FIGURE 7A OMITTED]. Mineralization rates expressed per unit soil volume are a more relevant predictor of plant nutrient availability. Our results indicate that N availability does increase substantially in more minerotrophic sites [ILLUSTRATION FOR FIGURE 5B OMITTED]. In contrast, [P.sub.min] per cubic centimeter was highest in the bogs under aerobic conditions and moderate in the bogs under anaerobic conditions [ILLUSTRATION FOR FIGURE 7B OMITTED].
There is a fundamental distinction between nutrient turnover rates and availability. To convert from a turnover rate to a mineralization rate per volume of soil, one multiplies the turnover rate by the total nutrient content on a dry mass basis and by bulk density. Consequently, soils with a high turnover rate but a low total nutrient content and a low bulk density, such as bogs and acidic fens (Table 1), would have a low mineralization rate per unit soil volume. To the extent that a mineralization rate per cubic centimeter is an appropriate indicator of nutrient availability, bogs and acidic fens would also be more deficient in nutrients for plant growth.
Organic soils are distinguished from mineral soils in their low and highly variable bulk densities. For example, gravimetric percent total N and percent total P in soil varied by 2.2 to 2.4 times among communities, whereas bulk density varied by 11.8 times. N turnover varied only moderately among communities (3.5 times under aerobic conditions, 2.6 times under anaerobic conditions), whereas P turnover varied greatly among communities (29.9 times for aerobic and 15.5 times for anaerobic conditions). Similarly to C mineralization, bulk density alone explained 80%-83% of the variation in aerobic and anaerobic [N.sub.min] per cubic centimeter. In contrast, bulk density was not significantly correlated with aerobic or anaerobic [P.sub.min] per cubic centimeter.
Barko and Smart (1986) found that sediment density was a dominant control over nutrient availability for two submersed macrophytes. They attributed the inhibitory effect of low sediment density on plant growth to long diffusion distances for nutrient uptake, as diffusion of nutrients through the soil to roots is often the rate-limiting step for nutrient uptake of plants (Chapin 1980). Bulk density also controls the nutrient pool size within a soil volume that microbes or plant roots can access.
Our data illustrate the importance of how nutrient mineralization and availability data are expressed. Most investigators report their results per mass of dry soil. Relative differences in [N.sub.min], [P.sub.min], and [C.sub.min] among our sites were very similar when expressed as a turnover rate or as mineralization per gram dry mass of soil (data not shown). This may be a minor problem in comparing mineralization rates in mineral soils, with their relatively similar bulk densities, but this can cause dramatic differences in the conclusions reached in site comparisons with organic soils.
The few studies that have examined relative differences in nutrient mineralization and availability across an ombrotrophic-minerotrophic gradient support our conclusions. Waughman (1980) examined 50 southern German peatlands and found that extractable N[[H.sub.4].sup.+] and P[[O.sub.4].sup.-3] were lower in fens than in bogs, although peat and vegetation in bogs had lower total P levels. Verhoeven et al. (1990) compared 10 bogs and fens in the Netherlands and found that the inorganic N pool was unrelated to peat type, the labile P pool was larger in bogs, and both N and P mineralization were faster in bogs.
We compared our data to nutrient mineralization potentials reported in other ecosystems, examining results over similar incubation periods. [N.sub.min] in our sites was [approximately]1-4 times higher than in upland forested ecosystems in Michigan, whereas [C.sub.min] was comparable (Zak et al. 1989, MacDonald et al. 1995). [N.sub.min] in our sites was 38-119% of that found in Minnesota old-field sites (Wedin and Pastor 1993). Three tundra tussock soils incubated at 37 [degrees] C (Marion and Black 1987) had similar [N.sub.min] rates to those in our sites. Much lower rates of [N.sub.min] and [P.sub.min], but similar rates of [C.sub.min], were found in arctic ecosystems when they were incubated at more natural summer soil temperatures between 3 [degrees] C and 15 [degrees] C (Nadelhoffer et al. 1991). Thus, mineralization potentials of N and P in a variety of Minnesota wetlands are similar to those in upland and arctic ecosystems, when incubated at similar temperatures and under similar conditions.
In general, nutrient mineralization rates are initially high and decrease rapidly through time as labile organic pools are depleted, leaving progressively more recalcitrant organic matter to be mineralized. This scenario was generally observed for N[[H.sub.4].sup.+] and P mineralization (data not shown). In contrast, the percentage of net [N.sub.min] as N[[O.sub.3].sup.-] increased in all sites over time, with a varying lag period that increased in duration with degree of ombrotrophy [ILLUSTRATION FOR FIGURE 6 OMITTED]. Additionally, the cumulative amount of N[[O.sub.3].sup.-] produced over the incubation increased significantly with higher initial soil pH ([ILLUSTRATION FOR FIGURE 9 OMITTED], P [less than] 0.001). Nitrification is inhibited generally in acidic soils (Focht and Verstraete 1977, Williams and Wheatley 1988), although significant nitrification rates occur in some acidic soils (Taylor et al. 1982, De Boer et al. 1992). Our results indicate a substantial capacity for nitrification in most wetland soils, but there is an important temporal component that is associated with the degree of ombrotrophy of the site. It is difficult to ascribe cause and effect for these temporal trends, because "quality" of both the soil and the microbial communities was changing through the incubations.
The apparent labile pools of N ([N.sub.0]), P ([P.sub.0]), and C ([C.sub.0]) were generally [less than]10% of their respective total organic pools (Tables 2 and 3). Similar results were found for [N.sub.0] in old-field soils in Minnesota (Wedin and Pastor 1993). This small labile pool was mineralized quickly, with time to 50% turnover ([t.sub.1/2]) ranging from [approximately]1 wk to 12 wk. The much larger recalcitrant pool had a [t.sub.1/2] ranging from [approximately]6 yr to [greater than]1300 yr. Thus, although there is a huge amount of organic matter in these wetlands, most of it cycles very slowly, and a small, labile pool is very important for determining short-term mineralization dynamics.
P in the bogs was unusual, in that 33% of it was labile anaerobically and 18% aerobically. About 12% of the total P was also labile in the acidic fens. The relative lability of total P in the more ombrotrophic sites can be attributed to our measurement of net P mineralization, because P[[O.sub.4].sup.-3] is strongly adsorbed by mineral components such as Al and Fe (Richardson 1985). The bog and acidic fen soils had very low ash and extractable Fe and Al contents (Table 1), and thus most of the mineralized P would be in an available form. Aerobic and anaerobic P turnover were negatively correlated with extractable Fe and Al content ([r.sup.2] = 0.43-0.58) and percent ash ([r.sup.2] = 0.31-0.34).
The acid-F data (Table 4) suggest that, under aerobic conditions, the labile mineral P pool in minerotrophic wetlands is quite large relative to the readily mineralizeable P pool ([P.sub.0]). In addition, we found that, aerobically, this labile mineral P pool generally increased over time, and thus was actually a net sink for mineralized P. The size of this mineral sink can be ecologically significant, accounting for [approximately]50% of net P mineralization in the minerotrophic wetlands (Table 4). This indicates that aerobic net P mineralization substantially underestimated gross mineralization in the minerotrophic sites.
The labile mineral P pool was much smaller relative to anaerobic [P.sub.0] than to aerobic [P.sub.0], particularly in the minerotrophic soils. Additionally, this labile mineral P pool generally decreased in size in anaerobic incubations, so it contributed to net mineralization. We estimate that desorption of P from the mineral fraction contributed [approximately]0-16% of anaerobic net mineralization in the peatland sites, whereas the mineral fraction was a net sink in the beaver meadows (Table 4).
Our results indicate that high mineral soil content in minerotrophic soils reduces P availability through the binding of P mineralized from the organic pool, particularly under aerobic conditions. Richardson (1985) found that aerobic P sorption in a broad variety of soils [TABULAR DATA FOR TABLE 7 OMITTED] was greater in more minerotrophic soils and was closely correlated with extractable Al content. Total P and ash content were positively correlated in our study ([r.sup.2] = 0.58). Thus, ombrotrophic peatlands have higher P availability because of the rapid turnover of a labile organic P pool, and the mineralized P is made biologically available rather than being bound to soil mineral matter. In contrast, more minerotrophic wetlands have a higher total soil P content, but the net mineralization rate of P is lower because of binding of phosphate from gross mineralization and, possibly, lower overall gross P mineralization. With our techniques, we could not separate gross and net P mineralization. Differences in the relative importance of microbial immobilization across communities may also be important in determining the relative availability of P in wetlands (Richardson and Marshall 1986, Walbridge 1991).
Anaerobic conditions depressed the size of the labile C ([C.sub.0]) and N ([N.sub.0]) pools. Generally, [k.sub.N] and [k.sub.p] were greater anaerobically, whereas [k.sub.c] values were similar under aerobic and anaerobic conditions. Similar results were found in modeling the kinetics of [C.sub.min] and [N.sub.min] in four upland, northern hardwood forests, where [k.sub.N]. and [k.sub.c] were not consistently related to temperature, but [C.sub.0] and [N.sub.0] were highly temperature dependent (MacDonald et al. 1995). In a previous experiment with soil from a beaver meadow and surface and deep peat (1 m depth) in a bog in northern Minnesota, [N.sub.0] and [C.sub.0] were more responsive than [k.sub.N] and [k.sub.c] to both temperature and aeration status (Updegraff et al. 1995) Thus, environmental conditions such as aeration status and temperature may affect the apparent labile pool size to an even greater degree than they affect the rate constants. The apparent pool size is a highly biologically dynamic variable and not a constant, as has often been assumed (cf. MacDonald et al. 1995).
Generally, the kinetic model (Eq. 1) predicted within a few percent the final cumulative amount of N, P, and C mineralized at week 59 ([N.sub.59], [P.sub.59], or [C.sub.59]). We determined the ratio of [N.sub.0], [P.sub.0], and [C.sub.0] to [N.sub.59], [P.sub.59], and [C.sub.59], respectively, to examine what proportion of mineralization could be accounted for, after a year, by the labile and recalcitrant pools (Table 7). From 10% to 87% of the mineralized nutrients and C could be accounted for by the labile pool, with an average across communities ranging from 41% for aerobic N to 77% for anaerobic P. There were no obvious trends among communities or aeration status. After 59 wk, a larger percentage of [P.sub.min] than of [C.sub.min] or [N.sub.min] was derived from the labile pool. The analysis in Table 4 suggests that [P.sub.59] is substantially underestimated in the more minerotrophic sites under aerobic conditions, because of sorption of mineralized P onto soil mineral matter. Although the recalcitrant pools had very low turnover rates, they were large and contributed to a substantial percentage of the observed mineralization over time.
Relationships between soil characteristics and mineralization
We correlated the factor scores from the PCA axes for the complete and for several reduced data sets of soil characteristics against cumulative mineralization (Tables 5 and 6). In general, mineralization, whether as a turnover rate or per unit soil volume, was well predicted by the PCAs.
Often, the group of soil variables from the proximate C analysis had the lowest correlation with nutrient and C mineralization, and in no case was it significantly better than other groups of variables that are much easier to measure. Litter bag experiments have often found that percent lignin (Meentemeyer 1978), the lignin:N ratio (Melillo et al. 1984), percent soluble phenolics (Palm and Sanchez 1990), and the lignin : holocellulose ratio (Melillo et al. 1989) can explain a large portion of both the short-term and long-term dynamics of litter decay. Surprisingly, the simple set of field variables was equally good or superior to more complicated chemical analyses in its ability to predict C and nutrient mineralization in northern wetlands. These field variables can be readily measured and are often part of previously collected databases on wetlands. Moreover, the physical degree of decomposition, as percent rubbed fiber, is used to classify and map peats. This should make predictions of C and nutrient mineralization much more tractable for large-scale efforts. Bulk density appears to be a particularly important soil parameter for predicting [C.sub.min] and [N.sub.min] per unit volume or area of soil, whereas it is of lesser importance for [P.sub.min].
This study showed large differences in mineralization rates of N, P, and C[H.sub.4] in different northern wetland communities. Attempts to model these areas or predict their response to climate change should not consider them as homogeneous entities.
Our results suggest that common perceptions of the role of nutrients in structuring these communities need to be reassessed, particularly the common assumption that more ombrotrophic sites are inherently more nutrient deficient. Although ombrotrophic sites had low total soil nutrient pools, the N and P in those pools turned over relatively rapidly. N availability was greater in more minerotrophic sites, but largely due to greater bulk density, whereas P availability remained high in ombrotrophic sites. Increasing the total mineral content of soils increased their total P content, but the net result was reduced P availability due to increased geochemical sorption of mineralized R
There were only small differences in total C mineralization among sites, but C[H.sub.4] production was extremely low in ombrotrophic sites. This suggests that the soil respiratory response of northern wetlands to climate change may be very different for these two important greenhouse gases.
We thank Anatasia Bamford for technical assistance in the laboratory. This research was funded by a grant from NASA's Terrestrial Biosphere Program, the National Science Foundation (DEB-9496305), and a Distinguished Global Change Postdoctoral Fellowship from the Department of Energy to Scott Bridgham.
Allen, S. E. 1989. Chemical analyses of ecological materials. Blackwell Scientific Publications, Oxford, UK.
Aselmann, I., and P. J. Crutzen. 1989. Global distribution of natural freshwater wetlands and rice paddies, their net primary productivity, seasonality, and possible methane emissions. Journal of Atmospheric Chemistry 8:307-359.
Barko, J. W., and R. M. Smart. 1986. Sediment-related mechanisms of growth limitation in submersed macrophytes. Ecology 67:1328-1340.
Bartlett, K. B., and R. C. Harriss. 1993. Review and assessment of methane emissions from wetlands. Chemosphere 26:261-320.
Binkley, D., and S. C. Hart. 1989. The components of nitrogen availability in forest soils. Advances in Soil Science 10:57-112.
Bowden, W. B. 1987. The biogeochemistry of nitrogen in freshwater wetlands. Biogeochemistry 4:313-348.
Bridgham, S. D., C. A. Johnston, J. Pastor, and K. Updegraff. 1995a. Potential feedbacks of northern wetlands on climate change. BioScience 45:262-274.
Bridgham, S. D., J. Pastor, J. Janssens, C. Chapin, and T. Malterer. 1996. Multiple limiting gradients in peatlands: a call for a new paradigm. Wetlands 16:45-65.
Bridgham, S. D., J. Pastor, C. A. McClaugherty, and C. J. Richardson. 1995b. Nutrient-use efficiency: a litterfall index, a model, and a test along a nutrient-availability gradient in North Carolina peatlands. American Naturalist 145:1-21.
Bridgham, S. D., and C. J. Richardson. 1992. Mechanisms controlling soil respiration (C[O.sub.2] and C[H.sub.4]) in southern peatlands. Soil Biology and Biochemistry 24:1089-1099.
Chapin, F. S., III. 1980. The mineral nutrition of wild plants. Annual Review of Ecology and Systematics 11:233-260.
Chapman, H. D. 1965. Cation exchange capacity. Pages 891-901 in C. A. Black, editor. Methods of soil analysis. First edition. American Society of Agronomy, Madison, Wisconsin, USA.
Clymo, R. S. 1983. Peat. Pages 159-224 in A. J. Gore, editor. Mires: swamp, bog, fen and moor. Ecosystems of the world, 4A. Elsevier Scientific Publishing, New York, New York, USA.
Clymo, R. S., and P. M. Hayward. 1982. The ecology of Sphagnum. Pages 229-289 in A. J. E. Smith, editor. Bryophyte ecology. Chapman and Hall, London, UK.
Conrad, R. 1989. Control of methane production in terrestrial ecosystems. Pages 39-58 in M. O. Andreae and D. S. Schimel, editors. Exchange of trace gases between terrestrial ecosystems and the atmosphere. John Wiley, Chichester, UK.
Day, J. H., P. J. Rennie, W. Stanek, and G. P. Raymond, editors. 1979. Peat testing manual. Technical memorandum number 125. Associate Committee on Geotechnical Research, National Research Council of Canada, Ottawa, Canada.
De Boer, W., A. Tietema, G. P. J. Klein, and H. J. Laanbroek. 1992. The chemolithotrophic ammonium-oxidizing community in a N-saturated acid forest soil in relation to pH-dependent nitrifying activity. Soil Biology and Biochemistry 24:229-234.
Dise, N. E., E. Gorham, and E. S. Verry. 1993. Environmental factors influencing the seasonal release of methane from peatlands in northern Minnesota. Journal of Geophysical Research 98:10 583-10- 594.
Focht, D. D., and W. Verstraete. 1977. Biochemical ecology of nitrification and denitrification. Pages 135-214 in M. Alexander, editor. Advances in microbial ecology. Plenum Press, New York, New York, USA.
Glaser, P. H. 1992. Ecological development of patterned peatlands. Pages 27-42 in H. E. Wright Jr., B. A. Coffin, and N. E. Aaseng, editors. The patterned peatlands of Minnesota. University of Minnesota Press, Minneapolis, Minnesota, USA.
Gorham, E. 1991. Northern peatlands: role in the carbon cycle and probable responses to climatic warming. Ecological Applications 1:182-195.
Gorsuch, R. L. 1983. Factor analysis. Lawrence Erlbaum Associates, Hillsdale, New Jersey, USA.
Harriss, R. C., and S. E. Frolking. 1992. The sensitivity of methane emissions from northern freshwater wetlands to global warming. Pages 48-67 in R Firth and S. G. Fisher, editors. Global climate change and freshwater ecosystems. Springer-Verlag, New York, New York, USA.
Hemond, H. F. 1983. The nitrogen budget of Thoreau's Bog. Ecology 64:99-109.
Houghton, J. T., B. A. Callander, and S. K. Varney. 1992. Climate change 1992: the supplementary report to the IPCC scientific assessment. Cambridge University Press, Cambridge, UK.
Jackson, M. L., C. H. Lim, and L. W. Zelazny. 1986. Oxides, hydroxides, and aluminosilicates. Pages 101-150 in A. Klute, editor. Methods of soil analysis. Part 1. Physical and mineralogical methods. Second edition. American Society of Agronomy and Soil Science Society of America, Madison, Wisconsin, USA.
Johnston, C. A. 1991. Sediment and nutrient retention by freshwater wetlands: effects on surface water quality. Critical Reviews in Environmental Control 21:491-565.
Keeney, D. R., and D. W. Nelson. 1982. Nitrogen: inorganic forms. Pages 643-698 in A. L. Page, R. H. Miller, and D. R. Keeney, editors. Methods of soil analysis. Part 2. Chemical and microbiological properties. Second edition. American Society of Agronomy, Madison, Wisconsin, USA.
Koerselman, W., H. De Caluwe, and W. M. Kieskamp. 1989. Denitrification and dinitrogen fixation in two quaking fens in the Vechtplassen area, The Netherlands. Biogeochemistry 8:153.
Koerselman, W., M. B. Van Kerkhoven, and J. T. A. Verhoeven. 1993. Release of inorganic N, P, and K in peat soils: effect of temperature, water chemistry, and water level. Biogeochemistry 20:63-81.
Lowther, J. R. 1980. Use of a single sulfuric acid and hydrogen peroxide digest for the analysis of Pinus radiata needles. Communications in Soil Science and Plant Analysis 11:175-188.
MacDonald, N. W., D. R. Zak, and K. S. Pregitzer. 1995. Temperature effects on kinetics of microbial respiration and net nitrogen and sulfur mineralization. Soil Science Society of America Journal 59:233-240.
Malmer, N. 1993. Mineral nutrients in vegetation and surface layers of Sphagnum-dominated peat-forming systems. Advances in Bryology 5:223-248.
Marion, G. M., and C. H. Black. 1987. The effect of time and temperature on nitrogen mineralization in arctic tundra soils. Soil Science Society of America Journal 51:1501-1508.
Matthews, E., and I. Fung. 1987. Methane emission from natural wetlands: global distribution, area, and environmental characteristics of sources. Global Biogeochemical Cycles 1:61-86.
Meentemeyer, V. 1978. Macroclimate and lignin control of litter decomposition rates. Ecology 59:465-472.
Melillo, J. M., J. D. Aber, A. E. Linkins, A. Ricca, B. Fry, and K. J. Nadelhoffer. 1989. Carbon and nitrogen dynamics along the decay continuum: plant litter to soil organic matter. Plant and Soil 115:189-198.
Melillo, J. M., R. J. Naiman, J. D. Aber, and A. E. Linkins. 1984. Factors controlling mass loss and nitrogen dynamics of plant litter decaying in northern streams. Bulletin of Marine Science 35:341-356.
Moore, P. D., and D. J. Bellamy. 1974. Peatlands. Springer-Verlag, New York, New York, USA.
Moore, T. R. 1994. Trace gases and climate change. Wetlands 14:223-228.
Moore, T. R., and R. Knowles. 1989. The influence of water table levels on methane and carbon dioxide emissions from peatland soils. Canadian Journal of Soil Science 69:33-38.
Nadelhoffer, K. J., A. E. Giblin, G. R. Shaver, and J. A. Laundre. 1991. Effects of temperature and substrate quality on element mineralization in six arctic soils. Ecology 72:242-253.
Olsen, S. R., and L. E. Sommers. 1982. Phosphorus. Pages 403-430 in A. L. Page, R. H. Miller, and D. R. Keeney, editors. Methods of soil analysis. Part 2. Chemical and microbiological properties. Second edition. American Society of Agronomy, Madison, Wisconsin, USA.
Palm, C. A., and P. A. Sanchez. 1990. Decomposition and nutrient release patterns of the leaves of three tropical legumes. Biotropica 22:330-338.
Post, W. M., W. R. Emanuel, P. J. Zinke, and A. G. Stangenberger. 1982. Soil carbon pools and world life zones. Nature 298:156-159.
Richardson, C. J. 1985. Mechanisms controlling phosphorus retention capacity in freshwater wetlands. Science 228:1424-1426.
Richardson, C. J., and P. E. Marshall. 1986. Processes controlling movement, storage, and export of phosphorus in a fen peatland. Ecological Monographs 56:279-302.
Ryan, M. G., J. M. Melillo, and A. Ricca. 1990. A comparison of methods for determining proximate carbon fractions of forest litter. Canadian Journal of Forest Research 20:166-171.
Schnitzer, M. 1982. Organic matter characterization. Pages 581-594 in A. L. Page, R. H. Miller, and D. R. Keeney, editors. Methods of soil analysis. Part 2. Chemical and microbiological properties. Second edition. American Society of Agronomy, Madison, Wisconsin, USA.
Shaver, G. R., and J. M. Melillo. 1984. Nutrient budgets of marsh plants: efficiency concepts and relation to availability. Ecology 65:1491-1510.
Soil Survey Staff. 1990. Keys to soil taxonomy. SMSS technical monograph number 6. Soil Management Support Services, Blacksburg, Virginia, USA.
Stanford, G., and S. J. Smith. 1972. Nitrogen mineralization potentials of soils. Soil Science Society of America Journal 36:465-472.
Stumm, W., and 1. J. Morgan. 1981. Aquatic chemistry. John Wiley, New York, New York, USA.
Taylor, A. A., J. De-Felice, and D.C. Havill. 1982. Seasonal variation in nitrogen availability and utilization in an acidic and calcareous soil. New Phytologist 92:141-152.
Thomas, G. W. 1982. Exchangeable cations. Pages 159-165 in A. L. Page, editor. Methods of soil analysis. Part 2. Chemical and microbiological properties. Second edition. American Society of Agronomy, Madison, Wisconsin, USA.
Updegraff, K., S. D. Bridgham, J. Pastor, and C. A. Johnston. 1994. A method to determine long-term anaerobic carbon and nutrient mineralization in soils. Pages 209-219 in J. Doran, D. Bezdicek, and D. Coleman, editors. Defining soil quality for a sustainable environment. Soil Science Society of America, Madison, Wisconsin, USA.
Updegraff, K., J. Pastor, S. D. Bridgham, and C. A. Johnston. 1995. Environmental and substrate controls over carbon and nitrogen mineralization in northern wetlands. Ecological Applications 5:151-163.
Valentine, D. W., E. A. Holland, and D. S. Schimel. 1994. Ecosystem and physiological controls over methane production in northern wetlands. Journal of Geophysical Research 99:1563-1571.
Veerkamp, M. T., W. J. Corre, B. J. Atwell, and P. J. C. Kuiper. 1980. Growth rate and phosphate utilization of some Carex species from a range of oligotrophic to eutrophic swamp habitats. Physiologia Plantarum 50:237-240.
Verhoeven, J. T. A., E. Maltby, and M. B. Schmitz. 1990. Nitrogen and phosphorus mineralization in fens and bogs. Journal of Ecology 78:713-726.
Walbridge, M. R. 1991. Phosphorus availability in acid organic soils of the Lower North Carolina Coastal Plain. Ecology 72:2083-2109.
Waughman, G. J. 1980. Chemical aspects of the ecology of some south German peatlands. Journal of Ecology 68:1025-1046.
Waughman, G. J., and D. J. Bellamy. 1980. Nitrogen fixation and the nitrogen balance in peatlands. Ecology 61:1185-1198.
Wedin, D. A., and J. Pastor. 1993. Nitrogen mineralization dynamics in grass monocultures. Oecologia 96:186-192.
Wilkinson, L., M. Hill, and E. Vang. 1992. Systat: statistics, version 5.2 edition. Systat, Evanston, Illinois, USA.
Williams, B. L., and R. E. Wheatley. 1988. Nitrogen mineralization and water-table height in oligotrophic deep peat. Biology and Fertility of Soils 6:141-147.
Wright, H. E., Jr. 1992. Introduction. Pages xv-xx in H. E. Wright Jr., B. A. Coffin, and N. E. Aaseng, editors. The patterned peatlands of Minnesota. University of Minnesota Press, Minneapolis, Minnesota, USA.
Zak, D. R., G. E. Host, and K. S. Pregitzer. 1989. Regional variability in nitrogen mineralization, nitrification, and overstory biomass in northern Michigan. Canadian Journal of Forest Research 19:1521-1526.
|Printer friendly Cite/link Email Feedback|
|Author:||Bridgham, Scott D.; Updegraff, Karen; Pastor, John|
|Date:||Jul 1, 1998|
|Previous Article:||The response of tundra plant biomass, aboveground production, nitrogen, and CO2 flux to experimental warming.|
|Next Article:||Ectomycorrhizal fungal community structure of pinyon pines growing in two environmental extremes.|