Ecosystems

, Volume 10, Issue 2, pp 226–238 | Cite as

The Effects of Adjacent Land Use on Nitrogen Dynamics at Forest Edges in Northern Idaho

Article

Abstract

The effects of immediately adjacent agricultural fertilization on nitrogen (N) at upland forest edges have not been previously studied. Our objective was to determine whether N from fertilized agriculture enters northern Idaho forest edges and significantly impacts their N status. We stratified 27 forest edge sampling sites by the N fertilization history of the adjacent land: current, historical, and never. We measured N stable isotopes (δ15N), N concentration (%N), and carbon-to-nitrogen (C/N) ratios of conifer tree and deciduous shrub foliage, shrub roots, and bulk soil, as well as soil available N. Conifer foliage δ15N and %N, shrub root δ15N, and bulk soil N were greater and soil C/N ratios lower (P < 0.05) at forest edges than interiors, regardless of adjacent fertilization history. For shrub foliage and bulk soil δ15N, shrub root %N and C/N ratios, and soil nitrate, significant edge–interior differences were limited to forests bordering lands that had been fertilized currently or historically. Foliage and soil δ15N were most enriched at forest edges bordering currently fertilized agriculture, suggesting that these forests are receiving N fertilizer inputs. Shrub root %N was greater at forest edges bordering currently fertilized agriculture than at those bordering grasslands that had never been fertilized (P = 0.01). Elevated N at forest edges may increase vegetation growth, as well as susceptibility to disease and insects. The higher N we found at forest edges bordering agriculture may also be found elsewhere, given similar agricultural practices in other regions and the prevalence of forest fragmentation.

INTRODUCTION

Forest edge creation and the juxtaposition of edges with more intensive land uses, such as agricultural fields, may influence forest nitrogen (N) dynamics. This could have significant consequences, given that 62% of forested lands are located within 150 m of a forest edge in the continental United States (Riitters and others 2002). In regions experiencing high atmospheric N deposition, forest edges (from 0–25 to 0–150 m) have received greater amounts of N, as measured in canopy throughfall water, than forest interiors (Draaijers and others 1988; Weathers and others 2001; Spangenberg and Kölling 2004; Devlaeminck and others 2005). This pattern has been attributed to increased wind turbulence (Draaijers and others 1988) and vegetation density at edges (Weathers and others 2001). N deposition to forest edges could be even greater in forests immediately adjacent to fertilized agricultural fields due to drift of N fertilizers. Riparian forests are known to take up nitrate (NO3-N) from adjacent agricultural fields via groundwater (Peterjohn and Correll 1984; Jordan and others 1993; Hill 1996), but this phenomenon has not been studied in upland forests. One exception is that Honnay and others (2002) found higher levels of soil ammonium (NH4-N) at forest edges than interiors in five Belgian forest fragments adjacent to agricultural crops and attributed this pattern to the influx of agricultural fertilizers.

Nitrogen inputs from adjacent agricultural fields may positively impact forest growth but may also have undesirable ecological consequences (Matson and others 2002). In the interior northwestern US, forest N fertilization has increased the growth of Douglas-fir (Pseudotsuga menziesii var. glauca) (Balster and Marshall 2000; Garrison and others 2000), but it has also resulted in increased incidence of root rot, a commonly occurring disease of Douglas-fir and grand fir (Abies grandis) caused by Armillaria ostoyae (Entry and others 1991). Some North American forests receiving excess N have experienced eventual declines in productivity, increases in mortality, and shifts from coniferous to deciduous tree species (Fenn and others 1998). Finally, if forests are unable to retain elevated N inputs, the excess N lost via NO3-N leaching or release of the greenhouse gas nitrous oxide could have adverse affects upon other terrestrial or aquatic ecosystems and human health (Matson and others 2002).

The effects of N deposition upon forest ecosystems are commonly measured via foliage and soil N concentrations (%N) and carbon-to-nitrogen (C/N) ratios. Conifer and deciduous tree foliar %N have shown strong positive correlations with atmospheric N deposition (Flückiger and Braun 1998; Boggs and others 2005), and conifer foliar %N increased by as much as 45% with decreasing distance to a point source of high ammonia emissions (Kätzel and Löffler 1997). Foliar and soil C/N ratios have been significantly lower in forests receiving higher amounts of N deposition relative to other nearby forests in California and the Colorado Front Range (Fenn and others 1996; Baron and others 2000). Forest soil and litter C/N ratios are indicators of N retention, because NO3-N leaching increases as C/N ratios decrease (Emmett and others 1998; Gunderson and others 1998; Lovett and others 2002).

Nitrogen stable isotopes (δ15N) are also useful for understanding N dynamics. δ15N is the ratio of 15N to 14N relative to that of atmospheric N2 and varies as a function of both the δ15N composition of N inputs to a system and internal N transformations (Högberg 1997; Kendall 1998). Forest vegetation and soil 15N enrichment have been positively correlated with soil net N mineralization and nitrification potentials (Garten and Van Miegrot 1994) and stream NO3-N levels (Pardo and others 2002). 15N enrichment of soils is a common result of N fertilization due to resulting increases in nitrification and losses of 15N-depleted N via ammonia volatilization or nitrate leaching (Feigen and others 1974; Högberg 1990; Högberg and Johannisson 1993). The highest plant and soil δ15N were found in agricultural sites and the lowest in forests within a region in Germany (Norra and others 2005). Plant and soil δ15N were also effective indicators of the intensity of past agricultural use in forests of northeastern France due to the 15N enrichment associated with past fertilization (Koerner and others 1999). Higher δ15N in agricultural fields may not necessarily be attributable to the 15N composition of the fertilizer itself, because manufactured fertilizers have δ15N values similar to atmospheric N2 (0‰) and large variation, with mean δ15N ranging from −1.9‰ to +1.9‰ for ammonium sulfate and urea (Shearer and others 1974).

The objective of our study was to determine whether N from fertilized agricultural fields enters low elevation conifer forests in northern Idaho and whether this significantly impacts the N status of these forests. In this northern Idaho landscape, forests are found adjacent to land uses including agricultural fields, previously cultivated grasslands, and native prairies, and we stratified our forest edge sampling sites across this land use–N status gradient. We tested the hypotheses that forest plant and soil δ15N, %N, and available N are greater and C/N ratios lower (1) at forest edges than interiors and (2) at forest edges bordering currently fertilized agriculture than at edges bordering grasslands fertilized historically or never. We measured conifer foliage δ15N, %N, and C/N at 27 sampling sites, and at a subset of 9 sites we measured soil available N and δ15N, %N, and C/N of bulk soil and deciduous shrub foliage and roots.

METHODS

Study Area and Sampling Design

Twenty-seven sampling sites were established within a 4,763-km2 study area encompassing Latah and Benewah Counties in northern Idaho (Figure 1). The study area is located at the interface of the Palouse prairie and Clearwater Mountains and is dominated by forest (60%), agriculture (21%), and shrub and grasslands (17%) (Scott and others 2002). This study focused on low-elevation forests dominated by ponderosa pine (Pinus ponderosa) or Douglas-fir (P. menziesii var. glauca). Other tree species occurring at sampling sites were grand fir (A. grandis), western larch (Larix occidentalis), lodgepole pine (P. contorta), and western redcedar (Thuja plicata). The forest sites ranged in elevation from 820 to 1,050 m above sea level, whereas the region’s agricultural and urban areas are at a base elevation of 800 m. The soils are deep, well-drained silt loams of the orders Mollisols, Inceptisols, and Alfisols that were formed in loess, with volcanic ash, granite, and metasedimentary rock contributing to soil development at some study sites (Weisel 1980; Barker 1981). Annual precipitation averages 63 cm (1915–2005) in Potlatch, Idaho, which is centrally located in the study area (Western Regional Climate Center, http://www.wrcc.dri.edu). Most precipitation is received from November to May, whereas only 8 cm is received on average during the warmest portion of the growing season from July to September. Mean minimum and maximum temperatures are −6.2 and 2.1°C in January and 7.6 and 28.2°C in July. Regional deposition of inorganic N from 1994 to 2004 averaged 1.2 (0.1) and 3.2 (0.1) kg/ha/year in wet and dry deposition (SE given in parentheses), respectively, at the Palouse Conservation Farm in adjacent Whitman County, Washington, as reported by the National Atmospheric Deposition Program (http://www.nadp.sws.uiuc.edu).
Figure 1.

Map showing the location of sampling sites in Latah and Benewah Counties in northern Idaho.

The 27 forest sampling sites were stratified by adjacent fertilization history: currently fertilized, historically fertilized, or never fertilized. Nine sites were located adjacent to currently fertilized agricultural crops. Although the current crops were wheat and Kentucky bluegrass, it is possible that some fields had been in a legume rotation in the past. Rates of N fertilizer application vary by crop and other site conditions, but range from 168 to 182 kg/ha for spring wheat and 174–224 kg/ha for bluegrass seed (Mahler 2001; Mahler and Guy 2002). A common fertilization regime in the region is to inject anhydrous ammonia into the soil in the fall (except in bluegrass turf) and apply dry fertilizers in the spring in the form of urea or ammonium (Roy Patten, UI Plant Science Farm, personal communication). To confirm that fertilizers did not have distinct δ15N values, we tested locally obtained samples of dry fertilizers (n = 2). Urea (46% N) had a mean δ15N of −1.0‰, and ammonium phosphate sulfate (16% N) had a mean δ15N of −2.5‰. Shearer and others (1974) report values of +3.6 and −2.7‰ for samples of anhydrous ammonia. Another nine sampling sites were located adjacent to previously cultivated grasslands currently dominated by exotic pasture grasses. Many of these sites were enrolled in the Conservation Reserve Program, and we were able to confirm that most had been out of agricultural production for approximately 20 years. The remaining nine sites, where adjacent lands had never been fertilized, were divided among unplowed native prairie (three), forest meadows (two) and forest clearcut harvests (four). The forest meadows had been disturbed by logging camp activities during the 1920s.

Each sampling site was characterized by a 150- to 400-m length of forest edge. Two points were randomly located along each length of edge, at least 40 m apart, and at least 50 m from another edge of that forest. Forest interior sampling points were located 50 m from the edge points, perpendicular to the forest edge. All edges were at least 50 years old, with the exception of the clearcut edges that had been created within the past 2–10 years. Edge orientation varied, with edges facing north to east at 10 sites and south to west at 17 sites. Sampling sites were located at least 1 km apart.

We located ten of the sampling sites by randomly generating points throughout forests of the two counties, using the Idaho GAP Analysis land cover dataset (Scott and others 2002). We then identified forest edges located within a 1-km radius of each random point using recent digital imagery. The remaining sites were chosen purposively to achieve a balance of edge types for comparison. There were only three prairie-forest edges and two dry meadow-forest edges within the study area, and all of these were included. The four clearcut sites were chosen exclusively on the property of the University of Idaho Experimental Forest because this was the only owner of clearcut edges able to guarantee that sites would not be disturbed by timber harvesting during this study. The remaining eight sites were chosen by identifying agriculture- or grassland-forest edges on digital imagery.

Plant Sampling and Site-Level Measurements

Two canopy dominant or co-dominant trees (ponderosa pine or Douglas-fir) were chosen randomly from all potential trees at each forest edge and interior location (see Table 1 for details). Tree species were paired between edge and interior locations within each sampling transect, and ponderosa pine was preferentially sampled. At four sampling transects both species were sampled, resulting in four trees per location. We included tree species in our analysis, because %N and δ15N are consistently higher in ponderosa pine than Douglas-fir. Shade-intolerant ponderosa pine maintains a lower leaf area, with a higher N concentration per needle, than the more shade-tolerant Douglas-fir. At the forest edge, trees were chosen from within a search area centered on the randomly located point and extending 20 m along the edge and 10 m into the forest. At the forest interior, the search area was from 50 to 60 m from the forest edge and was extended as far as 75 m in a few cases. The diameter at breast height (DBH) range of sampled trees was restricted to 30–65 cm, and we only sampled trees with live crowns and no forks below 1.3 m.
Table 1.

Characteristics of Trees Sampled for Foliar δ15N, %N, and C/N Ratios

Adjacent fertilization history

Forest location

Ponderosa pine (#)

Douglas-fir (#)

DBH (cm)

Currently fertilized

Edge

24

12

47.9 (1.8)

Interior

24

12

44.9 (1.5)

Historically fertilized

Edge

20

18

43.4 (1.7)

Interior

20

18

43.7 (1.1)

Never fertilized

Edge

22

18

46.2 (1.3)

Interior

22

18

44.2 (1.2)

DBH = diameter at breast height (1.3 m above the ground). Mean with standard error given in parentheses.

Current-year foliage samples were collected from the selected 228 trees between August 17 and September 9, 2004. Foliage was typically obtained from the upper two-thirds of the canopy, using a shotgun to remove a small branch. In a few cases when ponderosa pines had extremely high canopies, foliage was sampled from the lower third of the canopy. Ponderosa pine foliar %N has increased with foliage height in even-aged plots, but not in multi-aged plots (Nagel and O’Hara 2001). Our sampling sites were multi-aged, so we do not expect that this sampling height difference significantly affected our results. We measured the DBH of each tree due to a previously demonstrated relationship between DBH and conifer foliar δ15N (Koyama and others 2005) and to account for potentially larger trees at the forest edge relative to interior. Foliage samples were kept frozen until dried for 48 h at 70°C. Dried foliage was ground to a fine powder with a ball mill and analyzed for δ15N (‰), %N (mg/mg × 100), and %C (mg/mg × 100) in a CE Instruments NC 2500 elemental analyzer combined with a Finnigan-MAT Delta+ continuous-flow mass spectrometer at the Idaho Stable Isotopes Laboratory. δ15N is determined as follows, where the standard is atmospheric N2 (0.3663 atom% 15N) (Hogberg 1997):
$$ \delta ^{{{\text{15}}}} {\text{N }}(\permille) = {\left( {{\left[ {\frac{{^{{{\text{15}}}} {\text{N}}}} {{^{{{\text{14}}}} {\text{N}}_{{{\text{sample}}}} }} - \frac{{^{{{\text{15}}}} {\text{N}}}} {{^{{{\text{14}}}} {\text{N}}_{{{\text{standard}}}} }}} \right]} - 1} \right)} \times 1,000. $$
Forest site productivity may be confounded with land use adjacent to a forest or vary between the forest edge and interior due to fine-scale topographical variation. For this reason, we recorded forest habitat type series (Cooper and others 1991) at each edge and interior sampling point. Habitat type series identifies different biophysical settings that are named for the most shade-tolerant and moisture-demanding tree species that is present and reproducing successfully. Habitat type series is related to forest site index, another common measure of site productivity (R. A. Monserud 1984; unpublished data). Series present across the sampling sites included, from warm/dry to warm/wet, ponderosa pine, Douglas-fir, grand fir, and western redcedar.

Foliage and roots were sampled from a common deciduous shrub, snowberry (Symphoricarpos albus), at three sites in each of the three fertilization history categories, in late May 2005. The three sites in the never-fertilized category included two forests adjacent to prairie and one adjacent to a meadow. Foliage was sampled from two randomly located shrubs at 4 and 50 m from the forest edge within each sampling transect. Three branches of leaves were clipped from each shrub, rinsed with deionized water, and kept frozen. A subset of leaves was dried at 70°C for 24 h. Fine roots (<1 mm diameter) located within the top 10 cm of soil were sampled from one of the shrubs at each location. Roots were rinsed with deionized water to remove soil particles and kept frozen until dried at 70°C for 48 h. Dried foliage and roots were ground and analyzed in the same manner as the conifer foliage.

Soil Sampling

Soils data were collected at the same nine sites as the snowberry data. Within each transect, samples were collected at 4 and 50 m from the forest edge within the forest and 50 m from the forest edge outside the forest. An exception was that NH4-N and NO3-N were not measured outside the forest at the three sites adjacent to agriculture. Although we measured bulk soil δ15N and soil available N, which can be related to nitrification rates and N losses, a limitation of this study is that we did not directly measure these processes.

We measured soil NH4-N and NO3-N concentrations using ion exchange resins. Three Unibest-PST 1 mixed-bed ion exchange resin capsules saturated with H+ and OH (Skogley 1992; Dobermann and others 1994) were buried at a systematic spacing at each sampling point, 10 cm below the surface of the mineral soil in late May 2004. A straight vertical cut was made into the soil and each capsule was placed into a small hole created within this face. Resin capsules were left buried for 141–145 days. Upon removal, the capsules were rinsed with double-deionized water to remove soil particles and were kept moist and refrigerated. Lab analyses were completed by Unibest International Corporation (Pasco, Washington, USA). In the lab, the capsules were shaken with 20 ml of 2 N HCl for 20 min. The extract was poured off and the addition of HCl and shaking was repeated two more times. The three extracts were mixed and analyzed on a Timberline TL-200 Ammonium Analyzer, which used sodium hydroxide to convert ammonium to ammonia gas. The sample flowed past a gas permeable membrane and ammonia was absorbed into a buffer solution, which then flowed into a conductivity cell for the determination of NH4-N. The same sample was then rerun with the addition of zinc, which converted the nitrate to NH4-N, NH4-N concentration was determined, and NO3-N was calculated by subtraction (M. Moore, Unibest, personal communication). The resulting NH4-N and NO3-N concentrations (mg/l) were converted to resin absorption quantity, expressed as μmol of N absorbed per cm2 of capsule surface per day (μmol/cm/day) (Skogley and others 1996). The method detection limit was 0.07 mg/l (∼3.96e-05 μmol/cm/day) for NH4-N and 0.08 mg/l (∼1.19e-04 μmol/cm/day) for NO3-N.

Three bulk soil samples were collected at each sampling location in October 2004. The samples were collected systematically, 0.5 m apart, beginning 1 m from the resin capsule sampling area, on transects parallel to the forest edge. Organic matter was cleared and mineral soil was sampled with a bulk density corer from a 5- to 10-cm depth. Each sample was a composite of two core samples. Soil samples were kept frozen until homogenized sub-samples were freeze dried for approximately 36 h. Samples were weighed after drying to determine bulk density (g/cm3). The dried soil was passed through a 2-mm sieve and then ground and analyzed in the same manner as the foliage samples. N content was also determined on a volume basis by multiplying by bulk density values.

Statistical Analysis

Statistical models of the foliage, root, and soil response variables were fit using the linear mixed-effects model (lme) function of the open-source statistical language R (R Development Core Team 2005, Pinheiro and Bates 2000). Mixed effects models include random effects that can account for correlated error among observations. In this case, due to the hierarchical, nested structure of our data, residuals from data in the same site, transect, or plot may be correlated. We included three nested random effects in each model: sampling site, transect nested within site, and plot nested within transect. In models of conifer foliage δ15N, %N, and C/N ratios, the fixed effects (predictor variables) were, in order: tree species, tree DBH, forest habitat type series, forest edge or interior location, adjacent land fertilization history, and interactions among tree species, forest location, and fertilization history. The fixed effects in the models of shrub and soil variables were forest location, adjacent fertilization history, and their interaction. We did not test for a site productivity effect with these variables because only two habitat type series were represented in the subset of nine sites. Only data from forest locations were included in models of soil variables; data from adjacent grasslands were used for reference only. One sampling transect had very high NO3-N values, but the model was robust to this influential point so it was retained.

Separate models were fit to include each level of nested random effects. Analysis of variance (ANOVA) was used to determine which model had the lowest Akaike’s information criterion (AIC) and thus best fit (Burnham and Anderson 1998). F tests from a sequential ANOVA were used to test for statistical significance at α = 0.05, using the best-fitting model. When adjacent fertilization history or interaction terms were statistically significant, we tested for pairwise differences (α = 0.05) among categories. We used the “simultaneous comparisons based on parameter estimates” (csimtest) function of the open-source statistical language R, which adjusted for multiple tests using the logical constraint method (Westfall 1997; Bretz and others 2004; R Development Core Team 2005).

For each model, residual and quantile–quantile plots were assessed, and Box–Cox tests were used to determine whether data transformations were suggested. Data transformations are noted in Tables 2 and 3. An R2 analog was used to assess model fit. The R2 value represents the proportion of the root mean square error (RMSE) explained by the fixed effects and is equal to the variance of the response variable minus RMSE2, divided by variance of the response variable. RMSE is the overall standard deviation of the model residuals. The Intraclass Correlation Coefficient (ICC) represents the proportion of the RMSE explained by each of the random effects and is equal to the random effect variance, divided by the random effect variance plus the variance of the residuals associated with fixed effects.
Table 2.

Results of Statistical Models for Conifer Foliage Variables

Predictor variables

δ15N

%N1

C/N

Tree species

<0.01

<0.01

<0.01

DBH

0.14

0.13

0.11

Habitat type series

0.01

0.36

0.35

Edge/interior location

<0.01

0.01

0.01

Fertilization history

0.29

0.03

0.01

Species × location

0.46

0.82

0.77

Species × fertilization history

0.69

0.01

<0.01

Location × fertilization history

0.18

0.38

0.40

Model diagnostics

  R2 (fixed effects)

0.24

0.74

0.76

  ICC site

0.55

0.12

0.10

  ICC transect

<0.01

NA

NA

  ICC point

0.17

NA

NA

Statistically significant P values from mixed effects model ANOVAs are shown in bold (α = 0.05).

R2 is an R2 analog for mixed effects models (see text for details).

Intraclass correlation coefficient (ICC) represents the proportion of the root mean square error explained by each random effect.

11/%N transformation was applied.

NA not included in the best-fitting model.

Table 3.

Results of Statistical Models for Shrub and Soil Variables

 

Shrub foliage

Shrub roots

Bulk soil

Soil available N

Predictor variables

δ15N

%N

C/N

δ15N

%N

C/N

δ15N

N1

C/N

NO3 2

NH4+ 2

Edge/interior location

<0.01

0.10

0.24

0.01

0.01

<0.01

0.01

0.03

0.02

0.55

0.33

Fertilization history

0.66

0.31

0.39

0.68

0.05

0.06

0.60

0.29

0.80

0.45

0.11

Location × fertilization history

<0.01

0.24

0.55

0.11

0.02

<0.01

0.03

0.95

0.11

0.02

0.32

Model diagnostics

  R2 (fixed effects)

0.10

0.02

-0.02

0.06

0.33

0.37

0.02

-0.11

-0.11

0.02

0.09

  ICC site

0.33

<0.01

0.07

0.39

<0.01

<0.01

0.16

0.51

0.51

<0.01

0.05

  ICC transect

0.29

0.63

0.50

NA

0.77

0.78

0.37

0.05

0.05

0.37

0.02

  ICC point

NA

NA

NA

NA

NA

NA

0.12

0.18

0.18

NA

0.35

Statistically significant P values from mixed effects model ANOVAs are shown in bold (α = 0.05).

R2 is an R2 analog for mixed effects models (see text for details).

Intraclass correlation coefficient (ICC) represents the proportion of the root mean square error explained by each random effect.

1Bulk soil N is expressed in units of mg/cm3.

21/NO3 or NH4+ transformation was applied.

NA not included in the best-fitting model.

RESULTS

Forest Edge Effect

Conifer foliage δ15N and %N were significantly greater and C/N ratios significantly lower at the forest edge than interior, regardless of adjacent fertilization history (Tables 2, 4). Shrub foliage δ15N was significantly greater at the forest edge than interior (Tables 3, 4), but only in forests adjacent to currently fertilized agriculture (P = 0.02) or historically fertilized grasslands (P < 0.01). Shrub root δ15N was significantly greater at the forest edge than interior, regardless of adjacent fertilization history (Tables 3, 4). Shrub roots were consistently more enriched in 15N than foliage from the same shrubs. Shrub root %N was significantly greater and C/N ratios significantly lower at the forest edge than interior (Tables 3, 4). However, this difference in shrub root %N was only found in forests adjacent to currently fertilized agriculture (P = 0.04) or historically fertilized grasslands (P = 0.01). For shrub root C/N ratios, this difference was only significant in forests adjacent to historically fertilized grasslands (P < 0.01).
Table 4.

Nitrogen Data Summarized by Forest Location and Adjacent Fertilization History

 

Currently fertilized

Historically fertilized

Never fertilized

Edge

Interior

Edge

Interior

Edge

Interior

δ15N (‰)

  Conifer foliage

−1.19 (0.27)

−2.85 (0.22)

−2.03 (0.28)

3.82 (0.21)

2.74 (0.33)

−3.67 (0.37)

  Shrub foliage

−1.12 (0.62)

−2.35 (0.51)

−1.23 (0.41)

3.89 (0.22)

−1.46 (0.47)

−1.93 (0.52)

  Shrub roots

0.12 (0.60)

−0.72 (0.45)

−0.04 (0.48)

1.91 (0.36)

0.75 (0.59)

−0.86 (0.49)

  Bulk soil

4.39 (0.18)

3.51 (0.34)

4.44 (0.23)

3.19 (0.24)

3.12 (0.29)

3.38 (0.44)

%N

  Conifer foliage

1.14 (0.03)1

1.07 (0.03)

1.07 (0.03)1

1.03 (0.03)

1.02 (0.03)2

1.01 (0.03)

  Shrub foliage

2.31 (0.13)

2.15 (0.09)

2.10 (0.08)

1.99 (0.08)

1.94 (0.08)

1.97 (0.08)

  Shrub roots

1.64 (0.15)1

1.43 (0.12)

1.29 (0.15)

1.00 (0.12)

1.04 (0.07)2

1.10 (0.09)

  Bulk soil3

1.65 (0.10)

1.40 (0.10)

1.95 (0.15)

1.47 (0.08)

2.45 (0.23)

2.13 (0.16)

C/N ratio

  Conifer foliage

45.11 (1.42)1

47.66 (1.32)

48.20 (1.49)1

50.01 (1.41)

51.02 (1.64)2

51.54 (1.78)

  Shrub foliage

20.66 (1.08)

21.81 (0.89)

22.24 (1.01)

23.25 (0.92)

23.96 (0.99)

23.76 (1.08)

  Shrub roots

28.39 (2.57)

32.26 (2.89)

34.96 (3.44)

49.20 (5.58)

45.44 (3.01)

42.22 (3.17)

  Bulk soil

16.28 (0.60)

16.83 (0.66)

16.00 (0.49)

18.89 (0.60)

16.06 (0.71)

16.44 (0.66)

Soil NO3-N 4

1.37 (0.40)

1.13 (0.43)

0.47 (0.07)

0.52 (0.06)

0.56 (0.04)

0.74 (0.13)

Soil NH4-N 4

4.98 (0.22)

4.89 (0.34)

4.57 (0.24)

6.13 (1.00)

4.00 (0.10)

4.32 (0.25)

Mean with standard error given in parentheses.

Edge-interior pairs shown in bold signify that values are significantly different between the edge and interior (P < 0.05).

1,2Different numbers signify statistically significantly differences (P < 0.05) among forest edges bordering different fertilization histories, within each row.

3Bulk soil N is expressed in units of mg/cm3.

4Units of umol/cm2/day have been multiplied by 1,000.

Bulk soil δ15N was significantly greater at the forest edge than interior (Tables 3, 4, Figure 2), but only in forests adjacent to historically fertilized grasslands (P = 0.01). Bulk soil N (mg/cm3) was significantly greater and C/N ratio significantly lower at the forest edge than interior, regardless of adjacent fertilization history (Tables 3, 4). Bulk soil %N, uncorrected for bulk density, did not vary between the forest edge and interior. Soil NO3-N was significantly greater at the forest edge than interior only in forests bordering currently fertilized agriculture (P = 0.02, Figure 3). Soil NH4-N did not differ between forest edge and interior locations. There was considerable variability in soil NH4-N within sampling points, as illustrated by ICC values (Table 3).
Figure 2.

Bulk mineral soil δ15N (‰), sampled at 5–10 cm depth. Data are shown at forest edge (4 m) and interior (50 m) locations and at 50 m from the forest edge in the adjacent agricultural field or grassland, within each adjacent fertilization history category. Error bars represent ±1 standard error around the mean (n = 3).

Figure 3.

Soil nitrate (μmol/cm2/day), sampled at 10 cm depth over 5 months using ion exchange resin capsules. Data are shown at forest edge (4 m) and interior (50 m) locations and at 50 m from the forest edge in the adjacent grassland, within each adjacent fertilization history category. Data from forests bordering currently fertilized agriculture are shown with and without (NI) data from one influential sampling transect. Error bars represent ±1 standard error around the mean (n = 3).

Fertilization Effect at Forest Edges

Conifer foliage %N and C/N ratio varied significantly with adjacent fertilization history at both forest edge and interior locations, but only for Douglas-fir (Table 2, Figure 4). Douglas-fir foliar %N was significantly greater and C/N ratios significantly lower at both edges and interiors of forests bordering currently or historically fertilized agriculture than in forests bordering grasslands that had never been fertilized (P < 0.02). Shrub root %N was greater at forest edges bordering currently fertilized agriculture than at those bordering grasslands that had never been fertilized (P = 0.01) (Table 3). The same difference in shrub root %N was not found among forest interiors. Conifer and shrub foliage δ15N, shrub root and bulk soil δ15N, %N, and C/N ratios, and soil NO3-N and NH4-N at forest edges did not vary significantly with adjacent fertilization history (Table 3).
Figure 4.

Ponderosa pine and Douglas-fir foliar δ15N (‰), %N, and C/N ratios at forest edge (0–10 m) and interior (>50 m from edge) locations, within each category of adjacent fertilization history. Error bars represent ±1 standard error around the mean (n = 9). Different letters next to data points represent statistically significant differences (P < 0.05) in edge or interior foliage measurements among adjacent fertilization categories.

DISCUSSION

Forest Edge Effect

We found that foliage and soil δ15N and %N were higher and C/N ratios lower at many forest edges as compared to interiors. For shrub foliage and bulk soil δ15N, shrub root %N and C/N ratio, and soil NO3-N, significant forest edge–interior differences were limited to forests bordering lands that had been fertilized currently or historically. For conifer foliage and shrub roots, δ15N was significantly higher at edges than interiors regardless of adjacent fertilization history (Tables 2, 3). Although our statistical analysis results showed significantly different conifer foliage %N and C/N ratios at edges and interiors regardless of the adjacent land use (Table 2), the data in Figure 4 are not entirely consistent with this statistical result. Ponderosa pine %N and C/N ratios do not appear to differ between edge and interior locations in forests bordering lands that have never been fertilized. This inconsistency may have occurred because edge values were consistently higher (%N) or lower (C/N ratio) than interior values.

We found some forest edge versus interior differences even in forests bordering lands that had never been fertilized. Bulk soil N was significantly greater and C/N ratio lower at forest edges than interiors. This suggests that forest edges may be receiving N inputs from atmospheric N deposition. Canopy throughfall N has been greater at forest edges relative to interiors in regions having relatively high levels of atmospheric N deposition (Draaijers and others 1988; Weathers and others 2001; Spangenberg and Kölling 2004; Devlaeminck and others 2005). Conifer foliage and shrub root δ15N were also higher at forest edges than interiors. Increased δ15N at forest edges may be a result of N transformations, which could be related to increased N inputs, or microclimatic differences at forest edges.

There is a large amount of agricultural land within and to the west of the study area that could be a source of regional N deposition. The total N deposition of 4.4 kg/ha/year (1994–2004 average) in our study area is lower than that found in or near large urban areas of the western United States. Total N deposition ranges from 4 to 8 kg/ha/year in the Colorado Front Range (Baron and others 2000), 7.5–15 kg/ha/year for the Phoenix, Arizona metropolitan area and surroundings, and 20–45 kg/ha/year in the San Bernardino Mountains of southern California (Fenn and others 2003). Although N deposition in our study region is low, even low to moderate deposition has had detectable effects in forest vegetation and soils in the Front Range of Colorado (Baron and others 2000). The forests we studied receive relatively low precipitation annually, especially during the second half of the growing season, which may result in comparable responses to those reported in the high-elevation Colorado forests. Relatively low N deposition rates have resulted in vegetation changes in European forests, and Nordin and others (2005) have suggested that the critical load for N in boreal conifer forests could be just 6 kg/ha/year.

Solar radiation, air and soil temperatures, and wind speed decrease with distance from a forest edge, whereas relative humidity and litter moisture increase (Matlack 1993; Chen and others 1995). These microclimatic gradients could potentially affect plant and soil %N, C/N ratios, and δ15N. Increased temperatures at the forest edge could increase microbial activity and thus N mineralization. Mineralization involves very little fractionation and would not be expected to result in 15N enrichment (Kendall 1998) , but it would supply more NH4-N for uptake by vegetation. Increased nitrification of the NH4-N could follow, enriching this NH4-N supply and potentially explaining the higher δ15N of the conifer foliage and shrub roots at edges that have never been exposed to N fertilization (Table 4).

Fertilization Effect at Forest Edges

Bulk soil and shrub foliage δ15N were higher at forest edges than interiors, but only in forests bordering currently or historically fertilized fields. We did not find significant differences in bulk soil δ15N among forest edges bordering currently or historically fertilized fields and those adjacent to lands that had never been fertilized, despite trends in the raw data (Figure 2). We may not have detected differences due to the large amount of variation in δ15N among and within sites, as illustrated by model ICC values (Tables 2, 3), and the large size of the soil N pool.

15N enrichment can occur as a result of N inputs due to related increases in nitrification, which results in retention of 15N-enriched NH4-N and losses of 15N-depleted NO3-N and ammonia (Högberg 1997). N fertilizers have been applied to the current and historical agricultural fields, and the soil δ15N values in these fields are similar to those found at the forest edges that border them (Figure 2). These findings imply that N fertilizer inputs are also responsible for the 15N enrichment and elevated N status found at these forest edges. It may also be possible that soil disturbance, independent of fertilization, led to increased nitrification and higher soil δ15N in the currently and historically plowed agricultural fields. However, there was rarely evidence of soil disturbances 4 m into the forests, where soils were sampled. Habitat type series also explained significant variation in conifer foliage δ15N (Table 2), suggesting that foliar δ15N is more sensitive to environmental variation than foliar %N. This may reflect differences in nitrification or denitrification related to variation in topography and soil moisture (Garten 1993; Sutherland and others 1993).

Our data are consistent with other studies that have found higher δ15N in foliage and soils as a result of N inputs. Norra and others (2005) reported δ15N ranges of 1–6‰ for agricultural soils as compared with −1.5–3.5‰ for forest soils (0–2 cm). We found that mean conifer foliage δ15N was 57% higher at forest edges bordering agriculture than never-fertilized grasslands. This is of similar magnitude to differences in conifer foliage δ15N between stream banks exposed to salmon carcass N inputs and adjacent upslope areas, where bankside foliage δ15N was 24–60% higher (Koyama and others 2005).

Elevated soil δ15N persisted at forest edges where adjacent agriculture had been abandoned (Figure 2). The ecological effects appear to decline with time, as plant δ15N and plant and soil %N and C/N ratios in forests bordering historically fertilized grasslands were intermediate between those found in forests adjacent to lands currently and never fertilized (Table 4). These disturbed grasslands have not been grazed, so manure inputs cannot account for the higher δ15N observed in these grasslands or adjacent forest edges. The presence of N-fixing plants, which typically have δ15N close to that of atmospheric N2 (0 to +2‰), could also increase δ15N of the soil available N (Kendall 1998). Only two of the nine disturbed grassland sites had notable patches of the N-fixing plants Lupinus sp. and Vicia villosa, and these were not concentrated in areas where sampling occurred. N-fixing plants were also not present in large numbers in any of the forested areas sampled. Soil δ15N was virtually the same in currently and historically fertilized fields (Figure 2), which is consistent with the expected retention of heavier 15N. Others have found elevated soil δ15N in forests previously in agricultural use over 100 years following last cultivation (Koerner and others 1999; Compton and Boone 2000).

Elevated NO3-N at forest edges bordering agricultural fields may have been a consequence of nitrification of applied fertilizers. Although a considerable amount of NO3-N can be assimilated by forest soil microbes (Stark and Hart 1997), many studies have demonstrated that there is a preference for NH4-N by forest conifers (Lavoie and others 1992; Kronzucker and others 1997; Pritchard and Guy 2005). If NH4-N were assimilated more quickly, this may explain why we saw statistically significant patterns in NO3-N and not in NH4-N. Another possible explanation for this pattern is that NO3-N is more mobile than NH4-N (Binkley 1984) and thus may have been more easily trapped by the ion exchange resins. We also found higher soil NO3-N in grasslands than forests (Figure 3), which is consistent with previous findings that soil NO3-N is higher in disturbed sites than in later successional forests (Likens and others 1969; Walley and others 1996).

Snowberry root mean %N was 37% higher at forest edges adjacent to agriculture than at edges with no history of N fertilization (P = 0.01). We did not find the same difference in shrub root %N among forest interiors, so this result suggests that the higher %N may be attributable to inputs from adjacent N fertilization. Differences in snowberry foliage %N were not statistically significant. Fine roots may show a more pronounced response to fertilization because they take up N first before it is allocated to the foliage. Roots and organic matter have been found to be important N sinks in temperate forest 15N tracer studies (Nadelhoffer and others 2004; Perakis and others 2005). Snowberry foliage %N also showed no statistically significant increase 2 years following application of 220 kg/N/ha in a ponderosa pine forest in central Idaho (VanderSchaaf and others 2004). A similar lack of response of foliar %N to N fertilization was observed in the boreal forest shrub Vaccinium myrtillus (Nordin and others 1998).

Application rates of N fertilizers in our Palouse study area in northern Idaho are similar to those of other agricultural areas in the region, including southeastern Idaho, southeastern Washington, and the Willamette Valley of western Oregon (Puckett 1994). Application rates are even greater throughout much of the midwestern United States (Puckett 1994). Wheat, the dominant crop in our study area, receives a national average of 70–80 kg/ha/year of N per growing season (FAO 2002). Many other common crops in the United States also receive high annual N inputs, with application rates ranging from 50 to 60 kg/ha/year for oats and barley to 150 kg/ha/year for corn and canola and 220 kg/ha/year for potatoes (FAO 2002). Fertilizer application rates are continuing to increase nationally (Howarth and others 2002), so in addition to being potentially widespread, the effects of adjacent fertilization upon forests or other native vegetation could also become more pronounced.

CONCLUSIONS

At forest edges, conifer foliage δ15N and %N, shrub root δ15N, and bulk soil N were greater and soil C/N ratios lower than at forest interiors (P < 0.05), regardless of whether the immediately adjacent land had ever been fertilized with N. This pattern may be the result of regional atmospheric N deposition or microclimatic differences at forest edges. Other N measurements, shrub foliage and bulk soil δ15N, shrub root %N, and soil NO3-N, were only greater at forest edges than interiors in forests bordering lands that had been fertilized currently or historically. Additionally, shrub root %N was greater at forest edges bordering currently fertilized agriculture than at those bordering grasslands that had never been fertilized (P = 0.01). The higher plant and soil δ15N that we found at forest edges, especially those bordering agriculture, may be associated with increased nitrification rates and N losses related to N inputs. Collectively, our findings suggest that some forest edges receive N inputs from adjacent agricultural fertilization.

N inputs to forest edges may result in increased vegetation growth and vigor but may also increase susceptibility to insects and disease. Our findings demonstrate that forest fragmentation may have ecological implications beyond the commonly studied effects upon wildlife habitat, especially in landscapes where forests are in close proximity to intensive land uses. Similar effects may be widespread throughout the continental United States, given the comparable N fertilizer inputs in many other regions and the prevalence of forest edges.

Notes

ACKNOWLEDGEMENTS

We thank Kelsey Sherich and Howard Jennings for their valuable assistance in the field and Andrew Robinson, Sanford Eigenbrode, Douglas Frank and two anonymous reviewers for insightful comments that improved the original manuscript. We also thank the University of Idaho Experimental Forest, Potlatch Corporation, Bennett Lumber, McCroskey State Park, and 17 individual private forest owners for providing access to sampling sites. This research was supported by National Science Foundation IGERT grant 0014304, the USDA McIntire-Stennis Program, and the University of Idaho.

REFERENCES

  1. Balster NJ, Marshall JD. 2000. Eight-year responses of light interception, effective leaf area index, and stemwood production in fertilized stands of interior Douglas-fir (Pseudotsuga menziesii var. glauca). Can J Forest Res 30:733–43CrossRefGoogle Scholar
  2. Barker, RJ. 1981. Soil survey of Latah County area, Idaho. USDA Soil Conservation ServiceGoogle Scholar
  3. Baron JS, Rueth HM, Wolfe AM, Nydick KR, Allstott EJ, Minear JT, Moraska B. 2000. Ecosystem responses to nitrogen deposition in the Colorado Front Range. Ecosystems 3:352–68CrossRefGoogle Scholar
  4. Binkley D. 1984. Ion exchange resin bags: factors affecting estimates of nitrogen availability. Soil Sci Soc Am J 48:1181–4CrossRefGoogle Scholar
  5. Boggs JL, McNulty SG, Gavazzi MJ, Moore Myers J. 2005. Tree growth, foliar chemistry, and nitrogen cycling across a nitrogen deposition gradient in southern Appalachian deciduous forests. Can J Forest Res 35:1901–13CrossRefGoogle Scholar
  6. Bretz F, Hothorn T, Westfall P. 2004. Multcomp: Multiple tests and simultaneous confidence intervals. R package. URL: http://www.r-project.org/ [accessed May 15, 2005]
  7. Burham KP, Anderson DR. 1998 Model selection and inference: a practical information-theoretic approach. Springer, New YorkGoogle Scholar
  8. Chen J, Franklin JF, Spies TA. 1995. Growing-season microclimatic gradients from clearcut edges into old-growth Douglas-fir forests. Ecol Appl 5:74–86CrossRefGoogle Scholar
  9. Compton JE, Boone RD. 2000. Long-term impacts of agriculture on soil carbon and nitrogen in New England Forests. Ecology 81:2314–330CrossRefGoogle Scholar
  10. Cooper SV, Neiman KE, Roberts DW. 1991. Forest Habitat Types of Northern Idaho: A Second Approximation. General Technical Report INT-236. USDA Forest Service, Intermountain Research Station. Ogden, UtahGoogle Scholar
  11. Devlaeminck R, Schrijver AD, Hermy M. 2005. Variation in throughfall deposition across a deciduous beech (Fagus sylvatica L.) forest edge in Flanders. Sci Total Environ 337:241–52PubMedCrossRefGoogle Scholar
  12. Draaijers GPJ, Ivens WPMF, Bleuten W. 1988. Atmospheric deposition in forest edges measured by monitoring canopy throughfall. Water Air Soil Pollut 42:129–36CrossRefGoogle Scholar
  13. Dobermann A, Langner H, Mutscher H, Yang JE, Skogley EO, Adviento MA, Pampolino MF. 1994. Nutrient adsorption kinetics of ion exchange resin capsules: a study with soils of international origin. Commun Soil Sci Plant Anal 25:1329–53Google Scholar
  14. Emmett BA, Boxman D, Bredemeier M, Gunderson P, Kjønaas J, Moldan F, Schleppi P, Tietema A, Wright RF. 1998. Predicting the effects of atmospheric nitrogen deposition in conifer stands: evidence from the NITREX ecosystem-scale experiments. Ecosystems 1:352–60CrossRefGoogle Scholar
  15. Entry JA, Cromack K, Kelsey RG, Martin NE. 1991. Response of Douglas-fir to infection by Armillaria ostoyae after thinning or thinning plus fertilization. Phytopathology 81:682–9Google Scholar
  16. Feigen A, Shearer G, Kohl DH, Commoner B. 1974. The amount and nitrogen-15 content of nitrate in soil profiles from two central Illinois fields in a corn-soybean rotation. Soil Sci Soc Am Proc 38:465–71CrossRefGoogle Scholar
  17. Fenn ME, Poth MA, Johnson DW. 1996. Evidence for nitrogen saturation in the San Bernadino Mountains in southern California. Forest Ecol Manage 82:211–30CrossRefGoogle Scholar
  18. Fenn ME, Poth MA, Aber JD, Baron JS, Bormann BT, Johnson DW, Lemly AD, McNulty SG, Ryan DF, Stottlemeyer R. 1998. Nitrogen excess in North American ecosystems: predisposing factors, ecosystem responses, and management strategies. Ecol Appl 8:706–13CrossRefGoogle Scholar
  19. Fenn ME, Haeber R, Tonnesen GS, Baron JS, Grossman-Clarke S, Hope D, Jaffe DA, Copeland S, Geiser L, Rueth HM, Sickman JO. 2003. Nitrogen emissions, deposition, and monitoring in the western United States. Bioscience 53:391–403CrossRefGoogle Scholar
  20. Fluckiger W, Braun S. 1998. Nitrogen deposition in Swiss forests and its possible relevance for leaf nutrient status, parasite attacks and soil acidification. Environ Pollut 102:69–76CrossRefGoogle Scholar
  21. Food and Agriculture Organization of the United Nations (FAO). 2002. Fertilizer use by crop, fifth edition. Rome, Italy. URL: http://www.fertilizer.org/ifa/statistics.asp [accessed December 27, 2005]
  22. Garrison MT, Moore JA, Shaw TM, Mika PG. 2000. Foliar nutrient and tree growth response of mixed-conifer stands to three fertilization treatments in northeast Oregon and north central Washington. Forest Ecol Manage 132:183–98CrossRefGoogle Scholar
  23. Garten Jr. CT. 1993. Variation in foliar 15N abundance and the availability of soil nitrogen on Walker Branch watershed. Ecology 74:2098–113CrossRefGoogle Scholar
  24. Garten CT, Van Miegrot H. 1994. Relationships between soil dynamics and natural 15N abundance in plant foliage from Great Smoky Mountains National Park. Can J Forest Res 24:1636–45Google Scholar
  25. Gunderson P, Callesen I, de Vries W. 1998. Nitrate leaching in forest ecosystems is related to forest floor C/N ratios. Environ Pollut 102:403–7CrossRefGoogle Scholar
  26. Hill AR. 1996. Nitrate removal in stream riparian zones. J Environ Qual 25:743–55Google Scholar
  27. Högberg P. 1990. Forests losing large quantities of nitrogen have elevated 15N:14N ratios. Oecologia 84:229–31Google Scholar
  28. Högberg P. 1997. 15N natural abundance in soil-plant systems. New Phytol 137:179–203CrossRefGoogle Scholar
  29. Högberg P, Johannisson C. 1993. 15N abundance of forests is correlated with losses of nitrogen. Plant Soil 157:147–50Google Scholar
  30. Honnay O, Verheyen K, Hermy M. 2002. Permeability of ancient forest edges for weedy plant species invasion. Forest Ecol Manage 161:109–22CrossRefGoogle Scholar
  31. Howarth RW, Boyer EW, Pabich WJ, Galloway JN. 2002. Nitrogen use in the United States from 1961–2000 and potential future trends. Ambio 31:88–96PubMedCrossRefGoogle Scholar
  32. Jordan TE, Correll DL, Weller DE. 1993. Nutrient interception by a riparian forest receiving inputs from adjacent cropland. J Environ Qual 22:467–73CrossRefGoogle Scholar
  33. Katzel R, Loffler S. 1997. The influence of atmospheric NH3 on Scots pine (Pinus sylvestris L.). In: Mohren GMJ, Ed. Impacts of global change on tree physiology and forest ecosystems. Dordrecht: Kluwer Academic. pp 81–6Google Scholar
  34. Kendall C. 1998. Tracing nitrogen sources and cycling in catchments. In: Kendall C, McDonnell JJ, Eds. Isotope tracers in catchment hydrology. New York: ElsevierGoogle Scholar
  35. Koerner W, Dambrine E, Dupouey JL, Benoit M. 1999. δ15N of forest soil and understory vegetation reflect the former agricultural land use. Oecologia 121:421–5CrossRefGoogle Scholar
  36. Koyama A, Kavanagh K, Robinson A. 2005. Marine nitrogen in central Idaho riparian forests: evidence from stable isotopes. Can J Fish Aquat Sci 62:518–26CrossRefGoogle Scholar
  37. Kronzucker HJ, Siddiqi MY, Glass ADM. 1997. Conifer root discrimination against soil nitrate and the ecology of forest succession. Nature 385:59–61CrossRefGoogle Scholar
  38. Lavoie N, Vézina L-P, Margolis HA. 1992. Absorption and assimilation of nitrate and ammonium ions by jack pine seedlings. Tree Physiol 11:171–83PubMedGoogle Scholar
  39. Likens GE, Bormann FH, Johnson NM. 1969. Nitrification: importance to nutrient losses from a cutover forest ecosystem. Science 163:1205–6PubMedCrossRefGoogle Scholar
  40. Lovett GM, Weathers KC, Arthur MA. 2002. Control of nitrogen loss from forested watersheds by soil carbon: nitrogen ratio and tree species composition. Ecosystems 5:712–8CrossRefGoogle Scholar
  41. Mahler RL. 2001. Northern Idaho Fertilizer Guide: Bluegrass Seed. University of Idaho Cooperative Extension System, Agricultural Experiment Station. Publication CIS 788. Moscow, IdahoGoogle Scholar
  42. Mahler RL, Guy SO. 2002. Northern Idaho Fertilizer Guide: Soft White Spring Wheat. University of Idaho Cooperative Extension System, Agricultural Experiment Station. Publication CIS 1101. Moscow, IdahoGoogle Scholar
  43. Matlack GR. 1993. Microenvironment variation within and among forest edge sites in the eastern United States. Biol Conserv 66: 185–94CrossRefGoogle Scholar
  44. Matson P, Lohse KA, Hall SJ. 2002. The globalization of nitrogen deposition: consequences for terrestrial ecosystems. Ambio 31:113–9PubMedCrossRefGoogle Scholar
  45. Monserud RA. 1984. Height growth and site index curves for inland Douglas-fir based on stem analysis data and forest habitat type. Forest Sci 30:943–65Google Scholar
  46. Nadelhoffer KJ, Colman BP, Currie WS, Magill A, Aber JD. 2004. Decadal-scale fates of 15N tracers added to oak and pine stands under ambient and elevated N inputs at the Harvard Forest (USA). Forest Ecol Manage 196:89–107CrossRefGoogle Scholar
  47. Nagel LM, O’Hara KL. 2001. The influence of stand structure on ecophysiological leaf characteristics of Pinus ponderosa in western Montana. Can J Forest Res 31:2173–82CrossRefGoogle Scholar
  48. Norra S, Handley LL, Berner Z, Stuben D. 2005. 13C and 15N natural abundances of urban soils and herbaceous vegetation in Karlsruhe, Germany. Eur J Soil Sci 56:607–20CrossRefGoogle Scholar
  49. Nordin A, Nasholm T, Ericson L. 1998. Effects of simulated N deposition on understorey vegetation of a boreal coniferous forest. Funct Ecol 12:691–9CrossRefGoogle Scholar
  50. Nordin A, Strengbom J, Witzell J, Näsholm T, Ericson L. 2005. Nitrogen deposition and the biodiversity of boreal forests: implications for the nitrogen critical load. Ambio 34:20–4PubMedCrossRefGoogle Scholar
  51. Pardo LH, Hemond HF, Montoya JP, Siccama TG. 2002. Response of the natural abundance of 15N in forest soils and foliage to high nitrate loss following clear-cutting. Can J Forest Res 32:1126–36CrossRefGoogle Scholar
  52. Perakis SS, Compton JE, Hedin LO. 2005. Nitrogen retention across a gradient of 15N additions to an unpolluted temperate forest soil in Chile. Ecology 86:96–105CrossRefGoogle Scholar
  53. Peterjohn WT, Correll DL. 1984. Nutrient dynamics in an agricultural watershed: observations on the role of a riparian forest. Ecology 65:1466–75CrossRefGoogle Scholar
  54. Pinheiro JC, Bates DM. 2000. Mixed-effects models in S and S-Plus. New York: SpringerGoogle Scholar
  55. Pritchard ES, Guy RD. 2005. Nitrogen isotope discrimination in white spruce fed with low concentrations of ammonium and nitrate. Trees 19:89–98CrossRefGoogle Scholar
  56. Puckett LJ. 1994. Nonpoint and point sources of nitrogen in major watersheds of the United States. US Geological Survey Water-Resources Investigations Report 94–4001Google Scholar
  57. R Development Core Team. 2005. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL: http://www.R-project.org
  58. Riitters KH, Wickham JD, O’Neill RV, Jones KB, Smith ER, Coulston JW, Wade TG, Smith JH. 2002. Fragmentation of continental United States forests. Ecosystems 5:815–22CrossRefGoogle Scholar
  59. Scott JM, Peterson CR, Karl JW, Strand E, Svancara LK, Wright NM. 2002. A Gap Analysis of Idaho: Final Report. Idaho Cooperative Fish and Wildlife Research Unit. Moscow, IDGoogle Scholar
  60. Shearer GB, Kohl DH, Commoner B. 1974. The precision of determinations of the natural abundance of nitrogen-15 in soils, fertilizers, and shelf chemicals. Soil Sci 118:308–16CrossRefGoogle Scholar
  61. Skogley EO. 1992. The universal bioavailability environment/soil test Unibest. Commun Soil Sci Plant Anal 23:17–20CrossRefGoogle Scholar
  62. Skogley EO, Dobermann A, Warrington GE, Pampolino MF, Adviento MaAA. 1996. Laboratory and field methodologies for use of resin capsules. Sciences of Soils 1. URL: http://www.hintze-online.com/sos/1996/Toolbox/Tool1
  63. Spangenberg A, Kolling C. 2004. Nitrogen deposition and nitrate leaching at forest edges exposed to high ammonia emissions in southern Bavaria. Water Air Soil Pollut 152:233–55CrossRefGoogle Scholar
  64. Stark JM, Hart SC. 1997. High rates of nitification and nitrate turnover in undisturbed coniferous forests. Nature 385:61–4CrossRefGoogle Scholar
  65. Sutherland RA, van Kessel C, Farrell RE, Pennock DJ. 1993. Landscape-scale variations in plant and soil nitrogen-15 natural abundance. Soil Sci Soc Am J 57:169–78CrossRefGoogle Scholar
  66. VanderSchaaf CL, Moore JA, Kingery JL. 2004. The effect of multi-nutrient fertilization on understory vegetation nutrient concentrations in inland Northwest conifer stands. Forest Ecol Manage 190:201–18CrossRefGoogle Scholar
  67. Walley FL, Van Kessel C, Pennock DJ. 1996. Landscape-scale variability of N mineralization in forest soils. Soil Biol Biochem 28:383–91CrossRefGoogle Scholar
  68. Weathers KC, Cadenasso ML, Pickett STA. 2001. Forest edges as nutrient and pollutant concentrators: potential synergisms between fragmentation, forest canopies, and the atmosphere. Conserv Biol 15:1506–14CrossRefGoogle Scholar
  69. Weisel, CJ. 1980. Soil survey of Benewah County area, Idaho. USDA Soil Conservation ServiceGoogle Scholar
  70. Westfall PH. 1997. Multiple testing of general contrasts using logical constraints and correlations. J Am Stat Assoc 92: 299–30CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC 2007

Authors and Affiliations

  • Amy Pocewicz
    • 1
  • Penelope Morgan
    • 1
  • Kathleen Kavanagh
    • 1
  1. 1.Department of Forest ResourcesUniversity of IdahoMoscowUSA

Personalised recommendations