HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (16) Citing Articles via Google Scholar Google Scholar Articles by Venterea, R. T. Articles by Schwarz, P. A. Search for Related Content PubMed Articles by Venterea, R. T. Articles by Schwarz, P. A. GeoRef GeoRef Citation Agricola Articles by Venterea, R. T. Articles by Schwarz, P. A. Related Collections Forest Soils Nutrient Cycling Nitrogen

DIVISION S-3—SOIL BIOLOGY & BIOCHEMISTRY

Landscape Patterns of Net Nitrification in a Northern Hardwood-Conifer Forest

^a Institute of Ecosystem Studies, P.O. Box AB, Millbrook, NY 12545
^b Dep. of Forest Science, Oregon State Univ., Corvallis, OR, 97331

^* Corresponding author (venterea{at}soils.umn.edu)

	ABSTRACT

TOP NOTES ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

 
The production of NO^-₃ in forest soils is an important process influencing the form of N available for plant uptake and the potential for off-site N losses. We sampled mineral and organic soil horizons in 100 plots distributed across a 3160-ha hardwood-conifer forest in central New Hampshire to examine patterns of net nitrification rate (NR) and NO^-₃ concentration associated with physiographic features and vegetation abundances. Net NR and NO^-₃ concentrations each varied by a factor of 150 across the forest. Greater rates of net NO^-₃ production were associated with higher plot elevation, extent of southerly aspect, greater abundance of sugar maple (Acer saccharum Marsh.) and striped maple (Acer pensylvanicum L.), and lower abundances of coniferous trees. Regression models incorporating these factors accounted for 52% of the variance in NR across the entire valley. Higher soil water contents, N mineralization rates, and total N concentrations in higher elevation plots and higher N mineralization and respiration rates in more south-facing plots contributed to the landscape patterns. Incorporation of measured soil C/N ratios together with landscape factors accounted for an additional 10% of the variation. Regression models incorporating vegetation, physiographic, and soils data explained 62 to 73% of the variance in soil C/N ratios across the valley. Our results demonstrate how multiple biotic and abiotic factors combine to generate wide variation in NO^-₃ production, and provide a quantitative basis for estimating how vegetation shifts and climate change may influence N cycling in heterogeneous landscapes characteristic of many temperate forest ecosystems.

Abbreviations: BA, basal area • CV, coefficient of variation • HBEF, Hubbard Brook Experimental Forest • NR, nitrification rates • S, southness

	INTRODUCTION

TOP NOTES ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

 
THE PRODUCTION OF NO^-₃ via nitrification in forest soils is an important process influencing both the form of N available for plant uptake and the potential for N losses via NO^-₃ leaching or gas emission. In temperate forests, perhaps the most important impact of nitrification is its influence on NO^-₃ inputs to headwater streams, which in turn may affect downstream water quality (Goodale and Aber, 2001; Likens and Bormann, 1995). Patterns of soil NO^-₃ production across a broad forested landscape are likely to influence the variation in stream NO^-₃ levels among different watersheds and the spatial distribution of soil NO^-₃ inputs within each watershed. These patterns may influence the potential for in-stream retention or transformation of NO^-₃ in lower order streams before discharge to larger stream systems where NO^-₃ processing rates may be lower (Peterson et al., 2001). Nitrification also generates soil acidity and may promote the leaching of base cations from the plant-soil system (Likens and Bormann, 1995).

Nitrification is influenced by a variety of soil factors including water content, pH, temperature, texture, oxygen and NH⁺₄ availability, and the numbers and types of active nitrifying microbes (Schmidt, 1982). Across any single forest landscape, NRs may vary widely due, in part, to variation in vegetation type which affects the competition for N between plants, fungi and bacteria (Romell, 1935), the quality of organic substrates undergoing decomposition (Melillo et al., 1982; Scott and Binkley, 1997), the presence of secondary plant compounds (Fierer et al., 2001), and the soil chemical environment in general (Finzi et al., 1998a; Jenny, 1941). Some studies have indicated that the soil C/N ratio, which is greatly influenced by vegetation inputs, may provide a strong index of soil nitrification and NO^-₃ leaching potential (Lovett et al., 2002; Christ et al., 2002; Ollinger et al., 2002; Goodale and Aber, 2001; Lovett and Rueth, 1999). In addition to vegetation, variations in localized climatic influences due to physiography (e.g., elevation, aspect, slope) affect soil temperature and water content over the short term, and soil development over the long term. Thus, it is not surprising that several studies have found that NRs within temperate forest ecosystems can vary spatially by factors of 10 to 1000 (Ollinger et al., 2002; Lovett and Rueth, 1999; Zak et al., 1989; Federer, 1983).

Several studies have examined the influence of vegetation on soil N dynamics by comparing data from distinct stands dominated by individual species or vegetation types (e.g., hardwoods vs. conifers) (Ollinger et al., 2002; Giardina et al., 2001; Verchot et al., 2001; Lovett and Rueth, 1999; Finzi et al., 1998b; Zak and Pregitzer, 1990). A major limitation of this categorical approach is that many temperate forests in North America are composed largely of heterogenous mixtures of species including different deciduous and coniferous trees in close proximity (Ollinger et al., 2002; Schwarz et al., 2001; Finzi and Canham, 1998). There is very little information relating soil N cycling process rates or soil C/N ratios directly to vegetation abundances in multiple-species stands. In a study in the Catskills Mountains of New York state, Lovett et al. (2002) found that soil C/N ratio in mixed stands was positively related to abundance of red oak (Quercus rubra L.) and negatively related to abundance of sugar maple. With respect to physiography, a few studies have examined interactions between topography and vegetation in controlling nitrification (Christ et al., 2002; Bohlen et al., 2001; Knoepp and Swank, 1998; Zak et al., 1991). But few, if any, studies have utilized comprehensive sets of vegetation, physiographic, and soils data for robust analysis of potential controls over net NR across a broad and heterogenous landscape. Such an analysis may be useful not only as a means of characterizing extant patterns, but also for understanding how NO^-₃ leaching and N oxide gas emission rates may respond to climate change directly or to vegetation shifts due to disease or climate change.

The main objective of this study was to examine landscape patterns of net nitrification in a mixed hardwood-conifer forest during the middle of the growing season. We measured potential net NR in soils collected from 100 plots distributed across the Hubbard Brook Valley in central New Hampshire. Our analysis was primarily aimed at determining how much of the overall valley-wide variation in net nitrification could be explained by plot-scale landscape factors alone, and how much additional variation could be explained by supplementing landscape data with soils information. The soils data also were used to evaluate the underlying proximal controls mediating any observed landscape patterns.

	METHODS AND MATERIALS

TOP NOTES ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

 
Site Description
The study was conducted in the Hubbard Brook Experimental Forest (HBEF), which encompasses a 3160-ha valley within the White Mountain National Forest of central New Hampshire, USA (43° 56' N lat., 71° 45' W long.) (Fig. 1 ; http://www.hubbardbrook.org/). The HBEF has been the site of numerous biogeochemical investigations centered largely on a series of small (10–40 ha) watersheds in the northeast corner of the valley (e.g., Likens and Bormann, 1995). The study described here was part of a multi-investigator effort to characterize variation in ecosystem properties across the entire valley. The HBEF is composed of 30% (basal area, [BA]) yellow birch (Betula alleghaniensis Britton), 17% sugar maple, 12 to 13% each of red spruce (Picea rubens Sarg.), and American beech (Fagus grandifolia Ehrh.), approximately 7% each of paper birch (Betula papyrifera Marsh.) and balsam fir (Abies balsamea), and <6% each of red maple (Acer rubrum), Eastern hemlock (Tsuga canadensis), white ash (Fraxinus americana), and striped maple (Schwarz, 2001). The mean air temperatures are 19°C in July and -9°C in January. Annual precipitation of approximately 1400 mm is spread evenly throughout the year, of which approximately 30% occurs as snow (Likens and Bormann, 1995). Soils are mainly acidic (pH 3.5–5.5) and consist of well-drained, Typic Haplorthods of sandy loam texture derived from glacial till (USDA, 1996).

View larger version (28K): [in this window] [in a new window]   Fig. 1. Representation of Hubbard Brook Experimental Forest, showing 100 plot locations utilized in present study. Numbers are plot designations previously established by Schwarz et al. (2001). Dashed lines are approximate boundaries of areas used in previous watershed-scale studies (Likens and Bormann, 1995).

Soil Sampling
Soils were collected on four occasions in June and early July of 2000 (12–14 June, 19–21 June, 26–28 June, and 3–5 July). At each plot, five soil cores were taken at evenly spaced locations along a 12-m diam. circle concentric within the 25.2-m diam. plot boundary. Any loose litter (O_i horizon) was removed before insertion of a 5-cm diam. by 17-cm polyvinyl chloride (PVC) coring tool to a depth of at least 7 cm and at most 15 cm. The entire thickness of the O horizon and the thickness of the mineral horizon to the bottom of each core sample were determined. Soil cores were separated in the field into organic (O_e + O_a) and mineral horizons, and then aggregated to form a single organic and single mineral soil composite per plot. After collection, soils were stored in polyethylene bags at 4°C for 1 to 3 d. Six of the 100 plots (plots 31, 56, 196, 299, 312, and 325) were sampled during each sampling event to examine the extent of temporal variability in net NR occurring during the 4-wk sampling period.

Soil Analysis
Within 1 to 3 d of sample collection, soils were passed through an 8-mm sieve, manually homogenized, and weighed before analysis. Two subsamples (approximately 5 g each) of each mineral and organic soil composite were extracted overnight in 25 mL of 1.2 M KCl. Extracts were filtered (Whatman no. 42 paper, Whatman Ltd., Maidstone, Kent, UK) into 60-mL polyethylene bottles and stored at 4°C before analysis of inorganic N (NH⁺₄–N and total NO^-₂–N + NO^-₃–N) levels using an automated colorometric analyzer (3000 series, Perstorp Analytical, silver Springs, MD). Concurrently with the initial extraction, two additional 50-g subsamples of each composite were placed into separate 1-L glass jars. The jars were sealed with polyethylene plastic wrap, and a small piece (approximately 4 cm²) of filter paper moistened with deionized water was suspended above the incubating soil to minimize drying. Jars were incubated at room temperature (20–25°C) for 28 d, at which time the entire jar contents were extracted overnight with 250 mL of 1.2 M KCl. Extracts were filtered and stored as above before determination of final inorganic N concentrations. Potential net nitrification rates on a dry mass basis (mg N kg^-1 d^-1) were calculated from the net increase in total NO^-₂–N + NO^-₃–N concentration occurring during the incubation period. Potential net mineralization rates were calculated from the net increase in total inorganic N. Mean values of the two laboratory replicates were reported for each composite sample.

Three times during the above incubations (at 8, 19, and 27 d), seals on the incubation jars were removed for approximately 30 min to allow equilibration of the jar headspace with ambient (room) air. Jars were then sealed with metal screw-on lids equipped with butyl rubber septa. After 20 to 28 h, 9-mL headspace samples were removed by plastic syringe and transferred to evacuated 9-mL glass vials, which were stored at room temperature before determination of CO₂ concentrations using gas chromatography with thermal conductivity detection. Rates of CO₂ production (mg C kg^-1 soil d^-1) were calculated from the concentration at the end of the 20- to 28-h incubation and concentrations in room air samples collected during the 30-min equilibration period. Rates were corrected to account for the dry mass of soil present and the headspace volume occupied by soil. No consistent temporal trends during the incubation period were evident, so the mean of the three measurements was used in subsequent analysis.

Concurrently with the above analyses, 15- to 25-g subsamples of each composite were dried overnight at 60°C for determination of gravimetric soil water content. Dried soil was retained and subsequently analyzed for pH, relative water content, and total C and N concentrations. Soil pH was determined in approximately 5-g subsamples by manually mixing soil with 1 M KCl at a mass ratio of 2:1 (soil/solution) for mineral soil and 5:1 for organic soil, and measuring pH in supernatant solution after settling for 1 h. Because of a shortage of sample mass, pH was not determined in 22 of the 100 organic soils. Soil water-holding capacity was determined in 5- to 10-g subsamples by measuring the mass of water retained in initially saturated samples following 24 h of gravity drainage while placed in a funnel-filter apparatus. An index of relative water content was then calculated from the ratio of the gravimetric water content to the water-holding capacity, each expressed on a dry mass basis (g H₂O dry g^-1 soil). Separate subsamples were ground with a Kleco Tissue Pulverizer and analyzed for total N and C (Carlo Erba NS 1500, Carlo Erba, Milan, Italy). Sample masses for total C and N analysis (8–16 mg) were determined using a microbalance accurate to ±0.001 mg (Mettler-Toledo, Columbus, OH).

Data Analysis and Statistics
Slope aspect data were converted from compass readings (^o) to continuous indices of southness (S) (S = cos [^o+ 180°]) and eastness (E = sin[^o]) with values ranging from -1 to +1 for each index, that is, maximum eastness values occurred on sites with a measured aspect of 90°, and maximum southness values occurred on sites with a measured aspect of 180° (Schwarz, 2001). Net NR, NO^-₃ concentrations, and C/N ratios on a per mass soil basis were converted to areal units (e.g., kg N ha^-1) using the total dry mass of soil sampled and total area sampled within each plot. Overall plot areal values were determined from the sum of the areal values for the entire organic horizon and the mineral horizon to the sampled depth (>90% of plots had an average core depth of 13–15 cm). The calculated areal values are meant to provide a measure of process rates and concentrations integrated over the organic and mineral horizons and to provide some estimate of field areal values. The calculated values are subject to some error, since we did not quantitatively account for areal and vertical distributions of soil mass versus nonsoil material (e.g., rocks and roots) within each plot. To examine soil-property controls, multiple regression analyses were performed with net NR and NO^-₃ concentrations expressed on a per mass soil basis as dependent variables against soil properties separately for mineral and organic soils. To examine landscape-scale trends, regression analyses were done using net NR, NO^-₃ concentrations, and C/N ratios expressed on an areal basis as dependent variables. Separate analyses were performed using the following combinations of independent (explanatory) variables: (i) vegetation factors, (ii) physiographic and vegetation factors (hereafter collectively referred to as landscape factors), (iii) landscape factors combined with chemical and physical soil properties, and (iv) landscape factors combined with chemical, physical, and microbial soil properties (process rates). Microbial process rates (i.e., respiration rate and N mineralization rate) were not considered in explanatory models of soil C/N ratio. Models describing soil C/N ratio were developed using untransformed C/N ratio values. Distributions of net nitrification rates and NO^-₃ concentrations expressed both on an areal and per mass soil basis were somewhat skewed to the right. Kolmogorov-Smirnov goodness-of-fit tests showed that distributions resulting from square-root transformations improved the normality of the distributions. The deviations from normality were not extreme, and variable transformation had negligible effects on single-factor correlation analyses, so Pearson product-moment correlation coefficients using untransformed variables are reported for single-factor analysis. Multiple regression models using transformed and nontransformed variables generally yielded similar coefficients of multiple determinations (R²). Square-root transformation generated multiple-factor models with residuals that were more homoscedastic and normally distributed, so their results are reported and presented graphically (Zar, 1996). Stepwise regression analysis with a criterion of p < 0.05 was used to derive multiple factor regression models. The regression models developed here can be expressed in the general form

[1]

where D is the predicted dependent variable, i is the independent variable index, k is the total number of independent variables in the model, a_i is the regression coefficient for independent variable A_i, b is the regression constant, and c is a constant to correct for negative values of D (c 0 for net nitrification rate only). The exponent is equal to 2 for models using square root transformed variables (i.e., for net NR and NO^-₃ concentration), and = 1 for models using untransformed variables (i.e., for C/N ratio). In addition to R² values obtained with transformed variables, R² values were calculated after back-transforming the dependent variables, as an additional and more conservative measure of model efficiency (hereafter referred to as R²_b values). Because of the high ratio between the number of data points (n = 100 plots) and the number of independent variables in each model (5), adjustments based on the number of independent variables did not change R² values substantially and are therefore not reported. Statistical analysis was performed using the Statgraphics software package (Manugistics, Rockville, MD).

	RESULTS

TOP NOTES ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

 
Valley-Wide Variation
Laboratory-measured rates of potential net NR and concentrations of NO^-₃ showed wide variation across the HBEF (Table 1), ranging over two orders of magnitude, with coefficients of variation (CV) of 69 to 97%. Initial soil NO^-₃ concentrations were positively correlated with NR in mineral (r² = 0.58) and organic soil (r² = 0.69) on a mass basis, and on an areal basis (r² = 0.73, Fig. 2) . On a per mass soil basis, NR and NO^-₃ levels were significantly higher in the organic than the mineral soil horizon (p < 0.001). However, the importance of mineral soil processes was demonstrated in that the organic horizon contributed on average 35% of the total areal values because of the greater soil bulk density in the upper mineral horizon ( = 0.47 g cm^-3) compared with the organic horizon ( = 0.09 g cm^-3). The average total C contents were 9.9% (±0.4%) in mineral and 43% (±0.6%) in organic soil. In the six plots that were sampled at each of the four 1-wk sampling periods, there was no consistent temporal pattern in NR. The mean NR in these plots did not differ significantly by week (p > 0.89). There was also very little variation and no evident temporal pattern in soil water content during the 4-wk period, with CV values for each plot ranging from 4 to 17%.

View larger version (27K): [in this window] [in a new window]   Fig. 2. Net nitrification rates versus NO-3–N concentrations determined in soils stored at 4°C for 1 to 3 d following field collection for 100 plots distributed across the Hubbard Brook Experimental Forest.

View larger version (33K): [in this window] [in a new window]   Fig. 3. Net nitrification rates (NR) per mass soil in (a) mineral and (b) organic horizon soils versus NR predicted from multiple regression models in the form of Eq. [1] with = 2. Regression coefficients (ai) shown for each independent variable (Ai) in order of decreasing importance to overall model. r = relative water content; C/N = kg C per kg N; Mr = net N mineralization rate (mg N kg-1 d-1); N = g N per 100 g soil; C = g C per 100 g soil. * signifies p < 0.05; ** signifies p < 0.01; *** signifies p < 0.001.

View this table: [in this window] [in a new window]   Table 3. Single-factor correlation results for vegetation abundances and physiographic factors.

View larger version (37K): [in this window] [in a new window]   Fig. 4. Net nitrification rates (NR) and NO-3–N concentrations determined in soils stored at 4°C for 1 to 3 d following field collection in plots comprised of 50% or more basal area of a single species. RS = red spruce; BF = balsam fir; AB = American beech; YB = yellow birch; SM = sugar maple; PB = paper birch. Species with the same letter designations for each parameter are not significantly different (p > 0.05), using least significant differences multiple range test. Coefficients of variation (CV) values are tabulated.

While no single vegetation or physiographic factor alone could account for more than 23% for the variability in NR across the HBEF, the percentage increased to 52% when vegetation and physiographic variables were combined into multiple regression models (Fig. 5a) . Plot elevation was a highly significant explanatory variable (p < 0.001), and S was also significant (p < 0.01), when included in regression models with BA of total conifers, striped maple, and sugar maple. A similar model explained 40% of the variability in areal NO^-₃ concentrations using total BAs of all trees in replace of sugar maple BA (Fig. 6a) . The amount of the total variance in NR and NO^-₃ concentrations explained by multiple regression models increased with the incorporation of soil chemical factors into the models (C/N ratio was significant, Fig. 5b and 6b), and increased further with the incorporation of soil microbial process rates (net N mineralization rate was significant, Fig. 5c and 6c). Models describing NR tended toward underprediction when the measured NR was approximately 1 kg N ha^-1 d^-1 or greater (Fig. 5). Regression models were less able to account for variation in NO^-₃ concentrations than variation in potential rates of net nitrification. Models describing NO^-₃ concentration also tended toward underprediction at measured concentrations >2 kg N ha^-1 (Fig. 6). The trends toward underprediction at high levels of NR and NO^-₃ may in part have derived from nonlinear relationships between soil C/N ratio and NR that were found in this data set (not shown) and that have generally been observed (Ollinger et al., 2002; Goodale and Aber, 2001; Lovett and Rueth, 1999). Attempts to account for this by incorporating nonlinear terms in regression models were not effective in improving the overall models or removing these trends in residuals.

View larger version (37K): [in this window] [in a new window]   Fig. 5. Net nitrification rates (NR) on an areal basis versus NR predicted from multiple regression models in the form of Eq. [1] with = 2, using (a) landscape factors, (b) landscape factors and soil chemical properties, and (c) landscape factors, soil chemical properties, and microbial process rates. Regression coefficients (ai) shown for each independent variable (Ai) in order of decreasing importance to overall model. El = elevation (m); S = Southness; SM = sugar maple basal area; StM = striped maple basal area; CON = conifer basal area (m2 ha-1); C/N = kg C kg -1 N on areal basis; Mr = net N mineralization rate (kg N ha-1 d-1); * signifies p < 0.05; **signifies p < 0.01; *** signifies p < 0.001.

View larger version (34K): [in this window] [in a new window]   Fig. 6. Measured NO-3concentrations on an areal basis versus NO-3 predicted from multiple regression models in the form of Eq. [1] with = 2, using (a) landscape factors, (b) landscape factors and soil chemical properties, and (c) landscape factors, soil chemical properties, and microbial process rates. Regression coefficients (ai) shown for each independent variable (Ai) in order of decreasing importance to overall model. El = elevation (m); S = Southness; BF = balsam fir basal area; StM = striped maple basal area; BA = total basal area (m2 ha-1); C/N = kg C kg -1 N on areal basis; Mr = net N mineralization rate (kg N ha-1 d-1); * signifies p < 0.05; ** signifies p < 0.01; *** signifies p < 0.001.

View larger version (32K): [in this window] [in a new window]   Fig. 7. Measured soil C/N ratios on an areal basis versus C/N ratios predicted from multiple regression models in the form of Eq. [1] with = 1, using (a) vegetation factors, (b) landscape factors, and (c) landscape factors and soil chemical properties as independent variables. Regression coefficients (ai) shown for each independent variable (Ai) in order of decreasing importance to overall model. S = Southness; RS = red spruce basal area; EH = eastern hemlock basal area; StM = striped maple basal area; WA = white ash basal area (m2 ha-1); pH = mineral soil pH; * signifies p < 0.05; ** signifies p < 0.01; *** signifies p < 0.001.

	DISCUSSION

TOP NOTES ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

Other factors may have contributed to the elevation trend. We found higher total N content in mineral soil with elevation. A small watershed study at HBEF which also found higher net NR in higher elevation plots (Bohlen et al., 2001) found that the N content of beech and maple leaf litter increased, and litter C/N ratios decreased, with elevation, and that microbial biomass N in mineral soil increased with elevation. There is some evidence that rates of wet N deposition may increase with elevation at HBEF because of greater precipitation (Martin et al., 2000). Both dry deposition and cloud water deposition are also likely to increase with elevation in forests like the HBEF because of a variety of climatic and vegetation factors (Lovett and Kinsman, 1990). Notwithstanding elevation differences in N deposition, reduced plant competition for N at higher elevation would be expected to increase the amount of N available for soil microbes and thereby enhance nitrification potential (Robertson, 1982).

A recent study by Ollinger et al. (2002) found strong relationships between foliar N concentrations and soil NR across a fairly wide spectrum of forest types in New Hampshire. Considering the results of Ollinger, at al. (2002) together with the elevation trends in leaf litter N contents (Bohlen et al., 2001), it is logical to hypothesize that the elevation trends in nitrification that we observed may be well described as a function of variations in foliar chemistry across the HBEF. However, our data also suggest that soil moisture availability will play an important role in this trend. Another recent study in West Virginia found higher rates of potential net nitrification and NO^-₃ export to streams in a higher elevation, less dense, watershed compared with an adjacent watershed with a similar management history (Christ et al., 2002). Significant differences in predominant vegetation between watersheds (greater sugar maple at the high nitrifying site) likely contributed to the differences found by Christ et al. (2002). In a study in the Catskill Mountains, Lovett et al. (2002) found no significant effect of elevation on soil C/N ratio in multiple regression models that included species composition.

At HBEF, elevation trends in species distributions included increases in balsam fir and paper birch, and decreases in eastern hemlock and white ash, with elevation. The common co-occurrence of paper birch, our highest nitrifying species in the species-dominated plots (Fig. 4), with fir at high elevation complicates the interpretation of elevation patterns. This is the case at Hubbard Brook and throughout the northern hardwood forests (Evans et al., 1998). Our data suggest that these vegetation trends may have at least partly counteracted each other with respect to the elevation trend in net nitrification at HBEF (Table 3, Fig. 4). Thus, factors other than tree-species affects—that is, soil water content, N deposition, and N content of litter within species—may have been more important.

The pattern with respect to aspect that we observed was weaker than the elevation pattern. Southness of aspect accounted for only about 5% of the valley-wide variation in net nitrification as a single factor (Table 3), and contributed about 3% of the explanatory power in multiple regression models (Fig. 5 and 6). Relationships between S and rates of soil respiration and N mineralization were stronger than for S and net nitrification (Table 3). Some part of this trend was likely driven by soil temperature. Although all soils were incubated at the same temperature, earlier snow melt in the spring and warmer temperatures throughout the growing season provide more favorable conditions for microbial growth, and thus possibly resulted in higher populations of decomposing and nitrifying organisms in these soils at the time of collection. Multiple regression also indicates that C/N ratios are slightly lower on more south-facing slopes when vegetation factors are held constant (Fig. 7), although the reasons for this are not readily apparent. In contrast with our findings, Gilliam et al. (2001) found lower NO^-₃ concentrations on south- and southwest-facing slopes in forested watersheds in West Virginia and hypothesized that greater rates of weathering-induced nutrient loss and acidification may have contributed to the trend.

The Role of Vegetation
Many studies have shown that forest canopy tree species composition and litter quality are key factors controlling variation in N cycling in temperate forest soils (Lovett et al., 2002; Lawrence et al., 2000; Lovett and Rueth, 1999; Finzi et al., 1998b; Hart et al., 1997; Zak et al., 1989; Pastor et al., 1984). The low rates of net nitrification in the conifer species and relatively high rates in sugar maple that we observed are particularly consistent with previous studies (cited above). Our results, as summarized by the relationships in Fig. 5 through 7 and Table 3, further indicate that some degree of landscape-scale pattern in potential net NO^-₃ production and soil C/N ratio can be identified across a broad forested area which is heterogeneous with respect to both biotic and abiotic influences. Our results also point out the value of utilizing comprehensive vegetation abundance data in analyses, which allow for detection of multiple species influences. Restricting our consideration to species-dominated stands at HBEF would have greatly limited the robustness of the analysis, since there were few plots (six out of 100) comprised of more than 80% of a single species. The high variability in net nitrification potential and NO^-₃ levels within plots comprised of more than 50% of a single species (Fig. 4) further supports the value of our approach in elucidating landscape patterns.

While regression models utilizing comprehensive vegetation abundance data together with basic physiographic information accounted for 40 to 52% of the variation in NO^-₃ concentrations and potential net nitrification rates (Fig. 5a and 6a), supplementation of the landscape information with soil C/N ratio data generated models which explained considerably more of the variation (51–62%, Fig. 5b and 6b). This may indicate that vegetation influences were not fully accounted for by species BA. For example, litter inputs from each species may not have been proportional to their respective BA within each 0.05-ha plot. Also, litter inputs from outside plot boundaries may have had considerable influence on soil N cycling rates within the plot. This is further suggested by multiple regression results for C/N ratio, that is, supplementing plot-scale information with mineral or organic soil pH data also increased the model effectiveness (Fig. 7). Conifer litter inputs from outside the plots, not accounted for in the plot basal area data, may have influenced soil chemistry with respect to both pH (lower) and C/N ratio (higher) resulting in slower rates of net N mineralization and nitrification. This certainly raises questions about ideal sampling protocols for landscape characterizations of soil processes. While a larger plot size would appear to decrease this affect, it would likely also result in higher intra-plot variability in landscapes as heterogeneous as the HBEF, possibly masking the influence of controls that operate on the scale of a few meters or less (e.g., Boerner and Koslowsky, 1989). Little information is available on the relative effectiveness of varying plot size in capturing vegetation influences on N cycling. The results presented here suggest that studies of this nature would be valuable.

Implications and Limitations
The regression models developed here are primarily useful in identifying important controlling factors, and in better understanding how multiple factors can contribute to net process rates. Depending on the categories of information available for any particular site or study (i.e., vegetation, physiographic, and different types of soils data), one of the several relationships obtained here should allow for useful cross-site comparisons. The primary aim of the current study was to identify patterns and potential controls over net nitrification, since it is exactly this net rate (i.e., the sum of gross nitrification and gross NO^-₃ immobilization rates), which would be expected to ultimately control NO^-₃ accumulation and leaching potential in forest ecosystems. Our measurements do not provide any direct evidence regarding controls over these separate processes. Thus, factors that have been identified as proximal controls mediating the observed patterns of net nitrification (i.e., soil C/N ratio, N mineralization rates, water content, pH) may have influenced gross NR, gross microbial NO^-₃ immobilization rates, or both.

While our findings are consistent with other studies (Christ et al., 2002; Ollinger et al., 2002; Goodale and Aber, 2001; Lovett and Rueth, 1999) in demonstrating a negative correlation between soil C/N ratio and net nitrification, our data do not clarify precisely how C/N ratio may directly or indirectly mediate net nitrification on an individual process level. There are several possible mechanisms involved, all of which would tend to favor enhanced net nitrification at lower C/N ratios, including: (i) lower soil C/N ratio generally promotes higher rates of net N mineralization during decomposition (Jansson and Persson, 1982), generating a supply of NH⁺₄ above the demands of heterotrophic microbial assimilation, which in the short term provides primary substrate for gross nitrification and in the long term may support the development of robust nitrifying populations, (ii) increased NH⁺₄ availability deriving from increased N mineralization may act to inhibit heterotrophic assimilation of NO^-₃ because of microbial preference for NH⁺₄ (Stark and Hart, 1997; Jansson and Persson, 1982), and (iii) it is possible that high soil C/N ratio may be empirically associated with chemical factors which may inhibit nitrification, such as low pH (as in the current study) or tannins (e.g., Rice and Pancholy, 1973; Fierer et al., 2001), although these relationships have not been well explored. It should also be noted that bulk soil C/N ratios do not necessarily represent the chemical composition of the most labile organic matter that is undergoing decomposition and most strongly influencing N dynamics. Therefore, the soil C/N ratio may be primarily useful in this context as a general index of nitrification and NO^-₃ leaching potential, that is, without necessarily suggesting the exact mechanism(s) of control. Recent studies in Europe and North America have in fact shown that soil C/N ratio is a strong indicator of NO^-₃ leaching potential in forested watersheds (Gundersen et al., 1998; Lovett et al., 2002). Thus, our findings with respect to landscape and soil influences over soil C/N ratio (Fig. 7) have important implications for understanding and possibly predicting NO^-₃ leaching.

The relationship between laboratory-measured rates of net nitrification and NO^-₃ levels found here (Fig. 2) may be interpreted as support for the laboratory assays in representing field N cycling processes, similar to previous results (Ollinger et al., 2002). However, our NO^-₃ data should be interpreted with some caution, since soil samples were extracted for initial NO^-₃ determinations 1 to 3 d after collection. It is possible that some nitrification could have occurred during storage, to an extent that may in fact have been proportional to the subsequently measured net nitrification rates. We attempted to minimize this artifact by storing soils at 4°C, and by sieving the soils immediately before the initial extraction, although we did not test the effectiveness of these procedures in minimizing changes in NO^-₃ levels during storage.

Extension of our models for purposes of predicting NO^-₃ inputs to streams would require consideration of additional factors, including hydrology, which can be complex in many forest systems (e.g., Hagedorn et al., 1999), and other processes regulating the fate of NO^-₃ once produced via net nitrification (e.g., plant uptake and denitrification). This is also suggested by the reduced effectiveness of models obtained here in describing NO^-₃ concentration compared with net NR (Fig. 5 and 6). The impact of seasonal changes in plant nutrient uptake, climate and other factors on the relationships identified here needs to be investigated. More generally, the fact that a substantial portion of the variance remains unexplained by our models indicates that our knowledge of the controls on the nitrification process in forest soils is incomplete, and that predictive models derived from them would be subject to significant uncertainty. However, the relationships observed here might be useful in estimating broad differences in NO^-₃ leaching potential between distinct watersheds within or across landscapes, or between different areas within a single watershed.

Our findings also may be useful in estimating the general response of watersheds or landscapes to alterations in species composition because of climate change, which are predicted to be significant (Iverson and Prasad, 1998), as well as to species shifts because of selective harvest or disease. Indeed, tree species which proved to be strongly related to NO^-₃ production in this study (sugar maple and eastern hemlock), as well as one of the most predominant species at HBEF (American beech), are all presently being threatened by invasive diseases (USDA, 1999; Jenkins et al., 1999; DiGregorio et al., 1999). Thus, for example, our results suggest that decreases in the latitudinal range of coniferous trees under global warming scenarios could promote NO^-₃ leaching potential in some regions, while decreases in sugar maple abundance because of Asian longhorned beetle (Anoplophora glabripennis) infestation could have the reverse effect. In summary, while further data are certainly required in evaluating the generality of our findings to different forest systems, the relationships obtained here provide a quantitative basis for estimating spatial patterns given vegetation, physiographic, and soils data, and an opportunity for testing against data from future studies.

	ACKNOWLEDGMENTS

 
The authors gratefully acknowledge Jason Demers, Brent Mellen, Sabrina La Fave, Adam Welman, Bess Wellborn, and especially Amanda Elliott for help with field sampling and laboratory analysis. Comments provided by Martin Burger, Holly Ewing, Donald Zak, and two anonymous reviewers helped to improve earlier versions of the manuscript. This work was funded in part by HBLTER grant (NSF DEB-9810221), EPA NCERQA (Grant R827674) and a grant from the Andrew W. Mellon Foundation. This research was conducted at the Hubbard Brook Experimental Forest, which is owned and operated by the Northeastern Research Station, USDA Forest Service, Newtown Square, PA. This paper is a contribution to the Hubbard Brook Ecosystem Study and the Institute of Ecosystem Studies.

	NOTES

TOP NOTES ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

 
R.T. Venterea, current affiliation: USDA-ARS, 439 Borlaug Hall, Dep. of Soil, Water and Climate, Univ. of Minnesota, St. Paul, MN, 55108.

Received for publication May 6, 2002.

	REFERENCES

TOP NOTES ABSTRACT INTRODUCTION METHODS AND MATERIALS RESULTS DISCUSSION REFERENCES

 

• Boerner, R.E.J., and S. Koslowsky. 1989. Microsite variation in soil chemistry and nitrogen mineralization in a beech-maple forest. Soil Biol. Biochem. 21(6):795–801.
• Bohlen, P., P. Groffman, C. Driscoll, T. Fahey, and T. Siccama. 2001. Plant-soil-microbial interactions in a northern hardwood forest. Ecology 4:965–978.
• Bormann, F., T. Siccama, G. Likens, and R. Whittaker. 1970. The Hubbard Brook ecosystem study: Composition and dynamics of the tree stratum. Ecol. Monogr. 40:373–388.
• Christ, M., W. Peterjohn, J. Cumming, and M. Adams. 2002. Nitrification potentials and landscape, soil and vegetation characteristics in two Central Appalachian watersheds differing in NO^-₃ export. For. Ecol. Manage. 159:145–158.
• DiGregorio, L., M. Krasny, and T. Fahey. 1999. Radial growth trends of sugar maple (Acer saccharum) in an Allegheny northern hardwood forest. J. Torrey Bot. Soc. 126(3):245–254.
• Evans, C., E. Miller, and A. Friedland. 1998. Nitrogen mineralization associated with birch and fir under different soil moisture regimes. Can. J. For. Res. 28:1890–1898.
• Federer, C. 1983. Nitrogen mineralization and nitrification: Depth variation in four New England forest soils. Soil Sci. Soc. Am. J. 47:1008–1014.[Abstract/Free Full Text]
• Fierer, N., J. Schimel, R. Cates, and J. Zou. 2001. Influence of balsam poplar tannin fractions on carbon and nitrogen dynamics in Alaskan tiaga floodplain soils. Soil Biology Biochem. 33(12–13): 1827–1839.
• Finzi, A., and C. Canham. 1998. Non-additive effects of litter mixtures on net N mineralization in a southern New England forest. For. Ecol. Manage. 105:129–136.
• Finzi, A., C. Canham, and N. Van Breemen. 1998a. Canopy tree-soil interactions within temperate forests: Species effects on pH and cations. Ecol. Appl. 8:447–454.
• Finzi, A., N. Van Breemen, and C. Canham. 1998b. Canopy tree-soil interactions within temperate forests: Species effects on soil carbon and nitrogen. Ecol. Appl. 8:440–446.
• Giardina, C., M. Ryan, R. Hubbard, and D. Binkley. 2001. Tree species and soil textural controls on carbon and nitrogen mineralization rates. Soil Sci. Soc. Am. J. 65:1272–1279.[Abstract/Free Full Text]
• Gilliam, F.S., B. Yurish, and M. Adams. 2001. Temporal and spatial variation of nitrogen transformations in nitrogen-saturated soils of a central Appalachian hardwood forest. Canadian J. For. Res. 31:1768–1785.
• Goodale, C., and C. Aber. 2001. The long-term effects of land-use history on nitrogen cycling in northern hardwood forests. Ecol. Appl. 11:253–267.
• Gundersen, P., B. Emmett, O. Kjonaas, C. Koopmans, and A. Tietema. 1998. Impacts of nitrogen deposition on nitrogen cycling: A synthesis. For. Ecol. Manage. 101:37–55.
• Hagedorn, F., J. Mohn, P. Schleppi, and H. Fluhler. 1999. The role of rapid flow paths for nitrogen transformation in a forest soil: A field study with micro suction cups. Soil Sci. Soc. Am. J. 63:1915–1923.[Abstract/Free Full Text]
• Hart, S., D. Binkley, and D.A. Perry. 1997. Influence of red alder on soil nitrogen transformations in two conifer forests of contrasting productivity. Soil Biol. Biochem. 29:1111–1123.
• Iverson, L., and A. Prasad. 1998. Predicting abundance of 80 tree species following climate change in the eastern United States. Ecol. Monogr. 4:465–485.
• Jansson, S., and J. Persson. 1982. Mineralization and immobilization of soil nitrogen. p. 229–252. In F. Stevenson et al. (ed.) Nitrogen in agricultural soils. Agron. Monogr. 22. ASA, CSSA, and SSSA, Madison, WI.
• Jenkins, J., J. Aber, and C. Canham. 1999. Hemlock woolly adelgid impacts on community structure and N cycling rates in eastern hemlock forests. Can. J. For. Res. 29:630–645.
• Jenny, H. 1941. Factors of soil formation: A system of quantitative pedology. McGraw-Hill, NY.
• Johnson, C., C. Driscoll, T. Siccama, and G. Likens. 2000. Element fluxes and landscape position in a northern hardwood forest watershed ecosystem. Ecosystems 3:159–184.
• Knoepp, J., and W. Swank. 1998. Rates of nitrogen mineralization across an elevation and vegetation gradient in the southern Appalachians. Plant Soil 204:235–241.
• Lawrence, G., G. Lovett, and Y. Baevsky. 2000. Atmospheric deposition and watershed nitrogen export along an elevational gradient in the Catskill Mountains, New York. Biogeochemistry 50:21–43.
• Likens, G., and F. Bormann. 1995. Biogeochemistry of a forested ecosystem. Springer-Verlag, New York.
• Lovett, G., and J. Kinsman. 1990. Atmospheric pollutant deposition to high-elevation ecosystems. Atmos. Environ. 24A:2767–2786.
• Lovett, G., and H. Rueth. 1999. Soil nitrogen transformations in beech and maple stands along a nitrogen deposition gradient. Ecol. Appl. 9:1330–1344.
• Lovett, G., K. Weathers, and M. Arthur. 2002. Control of N loss from forested watersheds by soil C:N ratio and tree species composition. Ecosystems (In press).
• Martin, C., G. Likens, and D. Buso. 2000. Comparisons of long-term precipitation chemistry measurements at the Hubbard Brook Experimental Forest, New Hampshire. Water Air Soil Pollut. 120:359–379.
• Melillo, J., J. Aber, and J. Muratore. 1982. Nitrogen and lignin control of hardwood leaf litter decomposition dynamics. Ecology 63:621–626.[ISI]
• Moore, I., P. Gessler, G. Nielsen, and G. Peterson. 1993. Soil attribute prediction using terrain analysis. Soil Sci. Soc. Am. J. 57:443–452.[Abstract/Free Full Text]
• Ollinger, S., M. Smith, M. Martin, R. Hallett, C. Goodale, and J. Aber. 2002. Regional variation in foliar chemistry and N cycling among forests of diverse history and composition. Ecology 83:339–355.
• Pastor, J., J. Aber, C. McClaugherty, and J. Melillo. 1984. Above-ground production and N and P cycling along a nitrogen mineralization gradient on Blackhawk Island, Wisconsin. Ecology 1:256–268.
• Peterson, B., W. Wollheim, P. Mulholland, J. Webster, J. Meyer, J. Tank, E. Marti, W. Bowden, M. Valett, A. Hershey, W. McDowell, W. Dodds, S. Hamilton, S. Gregory, and D. Morrall. 2001. Control of nitrogen export from watersheds by headwater streams. Science 292:86–90.[Abstract/Free Full Text]
• Rice, E., and S. Pancholy. 1973. Inhibition of nitrification by climax ecosystems. II. Additional evidence and possible role of tannins. Am. J. Bot. 60(7):691–702.[ISI]
• Robertson, G. 1982. Nitrification in forested ecosystems. Philos. Trans. R. Soc. London Ser. B 296:445–457.
• Romell, L. 1935. Ecological problems of the humus layer of the forest. Cornell University Agricultural Experiment Station. Memoir 170. Cornell Univ. Agric. Exp.Stn., Ithaca, NY.
• Schmidt, E. 1982. Nitrification in soil. p. 253–288. In F. Stevenson (ed.) Nitrogen in agricultural soils. Agron. Monogr. 22. ASA, CSSA, and SSSA, Madison, WI.
• Schwarz, P. 2001. Spatial patterns of abundance of northern hardwood-conifer tree species in a forested valley in the White Mountains, New Hampshire, USA. (Ph.D. Dissertation). p. 145. Cornell University, Ithaca, NY.
• Schwarz, P., T. Fahey, C. Martin, T. Siccama, and A. Bailey. 2001. Structure and composition of three northern hardwood-conifer forests with differing disturbance histories. For. Ecol. Manage. 144:201–212.
• Scott, N., and D. Binkley. 1997. Foliage litter quality and annual net N mineralization: Comparison across North America forest sites. Oecologia 111:151–159.[ISI]
• Stark, J., and S. Hart. 1997. High rates of nitrification and nitrate turnover in undisturbed coniferous forests. Nature (London) 385:61–64.
• USDA. 1996. Hubbard Brook ecosystem study: Site description and research activities. USDA Forest Service. Publ. No. NE-INF-96–96R. 2nd ed. USDA Forest Service, Northeastern Forest Experiment Station, Newtown Square, PA.
• USDA. 1999. Asian longhorned beetle (Anoplophora glabripennis): A new introduction. USDA Animal and Plant Health Inspection Service Pest Alert NA-PR-01–99. USDA Forest Service, Northeastern Area, State and Private Forestry, Radnor, PA. Available online at http://www.na.fs.fed.us/spfo/alb. (verified 31 Oct. 2002.)
• Verchot, L., Z. Holmes, L. Mulon, P. Groffman, and G. Lovett. 2001. Gross vs. net rates of N mineralization and nitrification as indicators of functional differences between forest types. Soil Biology Biochem. 33:1889–1901.
• Zak, D., A. Hairston, and D. Grigal. 1991. Topographic influences on nitrogen cycling within an upland pin oak ecosystem. For. Sci. 37:45–53.
• Zak, D., G. Host, and K. Pregitzer. 1989. Regional variability in nitrogen mineralization, nitrification, and overstory biomass in northern Lower Michigan. Can. J. For. Res. 19:1521–1526.
• Zak, D., and K. Pregitzer. 1990. Spatial and temporal variability of nitrogen cycling in northern Lower Michigan. For. Sci. 36:367–380.
• Zar, J. 1996. Biostatistical analysis. Prentice Hall, NJ.

This article has been cited by other articles:

				J. L. Campbell, M. J. Mitchell, B. Mayer, P. M. Groffman, and L. M. Christenson Mobility of Nitrogen-15-Labeled Nitrate and Sulfur-34-Labeled Sulfate during Snowmelt Soil Sci. Soc. Am. J., October 29, 2007; 71(6): 1934 - 1944. [Abstract] [Full Text] [PDF]

				F. S. Gilliam, N. L. Lyttle, A. Thomas, and M. B. Adams Soil Variability along a Nitrogen Mineralization and Nitrification Gradient in a Nitrogen-Saturated Hardwood Forest Soil Sci. Soc. Am. J., January 1, 2005; 69(1): 247 - 256. [Abstract] [Full Text] [PDF]

				P. G. Scowcroft, J. E. Haraguchi, and N. V. Hue Reforestation and Topography Affect Montane Soil Properties, Nitrogen Pools, and Nitrogen Transformations in Hawaii Soil Sci. Soc. Am. J., May 1, 2004; 68(3): 959 - 968. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal