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 Similar articles in PubMed 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 (9) Citing Articles via Google Scholar Google Scholar Articles by Johnson, D.W. Articles by Todd, D.E. Search for Related Content PubMed PubMed Citation Articles by Johnson, D.W. Articles by Todd, D.E., Jr. GeoRef GeoRef Citation Agricola Articles by Johnson, D.W. Articles by Todd, D.E. Related Collections Watershed and Landscape Processes Soil Chemistry Forest Soils

TECHNICAL REPORT
Landscape and Watershed Processes

The Effects of Throughfall Manipulation on Soil Leaching in a Deciduous Forest

^a Environmental and Resource Sciences, Fleischmann Agriculture Bldg/370, Univ. of Nevada, Reno, NV 89557
^b Environmental Sciences Division, Oak Ridge National Lab., P.O. Box 2008, Building 1059, Oak Ridge, TN 37831-6422

* Corresponding author (dwj{at}unr.edu)

Received for publication February 19, 2001.

	ABSTRACT

TOP ABSTRACT INTRODUCTION SITE AND METHODS RESULTS DISCUSSION REFERENCES

 
The effects of changing precipitation on soil leaching in a deciduous forest were examined by experimentally manipulating throughfall fluxes in the field. In addition to an ambient treatment (AMB), throughfall fluxes were reduced by 33% (DRY treatment) and increased by 33% (WET treatment) using a system of rain gutters and sprinklers on Walker Branch Watershed, Tennessee. Soil leaching was measured with resin lysimeters in the O horizons and with ceramic cup lysimeters in the E (25 cm) and Bt (70 cm) horizons. Large and statistically significant treatment effects on N fluxes were found in the O horizons (lower N fluxes in the DRY and higher N fluxes in the WET treatment). Together with the greater O horizon N content observed in the DRY treatment, this suggested that N was being immobilized at a greater rate in the DRY treatment than in the AMB or WET treatments. No statistically significant treatment effects on soil solution were found in the E horizons with the exception of (Ca²⁺ + Mg²⁺) to K⁺ ratio. Statistically significant treatment effects on electrical conductivity (EC), pH, Ca²⁺, Mg²⁺, K⁺, Na⁺, SO^2-₄, and Cl^- were found in the Bt horizons due to differences between the DRY and other treatments. Despite this, calculated fluxes of Ca²⁺, Mg²⁺, K⁺, Na⁺, SO^2-₄, and Cl^- were lowest in the DRY treatment. These results suggest that lower precipitation will cause temporary N immobilization in litter and long-term enrichment in soil base cations whereas increased precipitation will cause long-term depletion of soil base cations.

Abbreviations: AMB, no treatment, ambient throughfall • DO, dry, open (location in the DRY treatment between troughs) • DRY, treatment where throughfall flux was decreased by 33% • DT, dry, treated (location in the DRY treatment under troughs) • EC, electrical conductivity • NuCM, Nutrient Cycling Model • ortho-P, ortho-phosphate • SWF, soil water flux • TDE, Throughfall Displacement Experiment • WET, treatment where throughfall flux was increased by 33% • WO, wet, open (location in the WET treatment removed from drip points) • WT, wet, treated (location in the WET treatment near drip points)

	INTRODUCTION

TOP ABSTRACT INTRODUCTION SITE AND METHODS RESULTS DISCUSSION REFERENCES

 
CLIMATE CHANGE could affect the cycling of nutrients and productivity of forest ecosystems in a number of ways. To date, most studies have emphasized the effects of increased temperature or elevated carbon dioxide; but changes in precipitation may have equal or greater effects (Kirschbaum, 1995). Specifically, changes in precipitation could cause (i) changes in productivity due to changes in both water and soil nutrient availability, (ii) changes in water yield, and (iii) changes in water quality (Johnson et al., 1998, 2001; Ulrich, 1983; Lamersdorf et al., 1995). Changes in soil N availability can have short-term effects on plant productivity because forests are often N limited. Ulrich (1983) noted pulses of NO^-₃ and Al during warm, dry years in the Solling forest ecosystem in Germany. He hypothesized that drought intensifies N mineralization and nitrification during summer, resulting in NO^-₃ pulses during rewetting periods. In acidic soils, the NO^-₃ pulse is accompanied by a pulse in soil solution Al concentrations. Results of a large field research project involving the imposition of drought through roofs in the forest subcanopy to five forest ecosystems in Europe failed to support Ulrich's (1983) hypothesis, but did suggest that natural periods of drought could produce such pulses (Lamersdorf et al., 1995).

Watershed-level studies of hydrology in humid forest ecosystems often show that evapotranspiration (ET) remains relatively constant despite large variations in precipitation (Luxmoore and Huff, 1989; Likens et al., 1977). Thus, ET consumes a larger proportion of precipitation under dry conditions than under wet conditions, and soil water flux (SWF) is affected disproportionately by changes in precipitation. The magnitude of this effect varies with the amount of precipitation compared with ET in a nominal year. This is illustrated by the results of simulation modeling of the effects of changing precipitation in several forest ecosystems (Johnson et al., 2001). In a red spruce (Picea rubens Sarg.) forest in the Smoky Mountains of North Carolina where ET consumed only 33% of precipitation, simulated precipitation changes of -28% and +28% caused -36% to +35% changes (respectively) in simulated SWF. In a slash pine forest in Florida where ET consumed 88% of precipitation, -33% and +33% changes in simulated precipitation caused -85% and +122% (respectively) changes in simulated SWF. Thus, these simulations suggested that the effects of changing precipitation on SWF are greater in sites where ET consumes a greater proportion of precipitation. On Walker Branch Watershed, the site of the current study, simulated changes of -33% and +33% in precipitation caused -43% and +47% (respectively) changes in simulated SWF (Johnson et al., 1998, 2001).

Reduced SWF can cause increased ionic concentrations in soil solutions, especially for conservative ions such as Cl^-, whose concentrations are not significantly buffered by adsorption reactions (Johnson et al., 2001). For anions such as H₂PO^-₄ and SO^2-₄, adsorption to Fe and Al hydrous oxides may minimize concentration increases due to drought. In the case of HCO^-₃, concentrations are controlled by pH and the partial pressure of CO₂ in soil and pH, and could be affected either positively or negatively by drought. The effects of drought on cations in soil solution should, in theory, be a function of changes in total ionic concentration. As noted by Reuss (1983), increased total ionic concentrations cause trivalent cations to increase disproportionately to di- and monovalent cations (to the 3/2 and third power, respectively) and divalent cations to increase as the square of monovalent cations.

To investigate the potential effects of changing precipitation on forest ecosystems, the Throughfall Displacement Experiment (TDE) was established on Walker Branch Watershed, Tennessee in 1993. Three different throughfall amounts were tested: -33% (DRY), ambient (no change, AMB), and +33% (WET). Throughfall amounts were controlled using a system of rain gutters and sprinklers. In a previous paper, we described the results of 13 months of data on soil solution chemistry from this site and included projections of longer-term trends using the Nutrient Cycling Model (NuCM; Johnson et al., 1998). In that paper, we found higher electrical conductivity (EC) and SO^2-₄, Cl^-, K⁺, Ca²⁺, Mg²⁺, and Na⁺ concentrations in Bt horizon soil solutions (70 cm) in the DRY than either the AMB or WET treatments. We found no treatment effects on pH, HCO^-₃, NH⁺₄, NO^-₃, or ortho-phosphate (ortho-P) concentrations. As is usually the case in N-limited ecosystems, soil solution concentrations of NH⁺₄, NO^-₃, and ortho-P were at or near trace levels at most times, limiting our ability to evaluate treatment effects on these ions.

In this paper, we update the Bt horizon soil solution data set and add data on soil solution from surface (E, 25 cm) soil horizons and on N and P fluxes measured with resin lysimeters from O horizons. We use this data to test our hypotheses, based on the literature review above and previous research at this site. Our hypotheses were that changes in throughfall inputs will cause:

1. concomitant changes in N and P fluxes in O horizons;
   
2. inverse changes in EC and ion concentrations in soil solutions;
   
3. concomitant changes in SO^2-₄ to Cl^- ratios (because of greater changes in conservative anions such as Cl^- than adsorbing anions like SO^2-₄); and
   
4. inverse changes in (Ca²⁺ + Mg²⁺) to K⁺ ratios (because of greater changes in divalent than in monovalent cations).
   

	SITE AND METHODS

TOP ABSTRACT INTRODUCTION SITE AND METHODS RESULTS DISCUSSION REFERENCES

 
Site Description
The throughfall displacement system for the experiment is located on the Walker Branch Watershed (35°58' N, 84°17' W), a part of the U.S. Department of Energy's (DOE) National Environmental Research Park near Oak Ridge, Tennessee (Johnson and Van Hook, 1989). Long-term mean annual precipitation is 1358 mm and mean temperature is 14.2°C. Soils are classified as Fullerton series (fine, kaolinitic, thermic Typic Paleudult). Some chemical and physical properties are given in Table 1. Depth to bedrock at this location is approximately 30 m. The site was chosen because of its uniform slope, consistent soils, and a reasonably uniform distribution of vegetation. White oak (Quercus alba L.), chestnut oak (Quercus montana Willd.), and red maple (Acer rubrum L.) dominate the site, but it contains 16 other tree species (Hanson et al., 1995, 1998, 2001). Stand basal area averages 20 to 25 m² ha^-1. The experimental area was located at the upper divide of the watershed so that lateral flow of water into the soils at the top of the plots would not confound attempts to create a reduced soil water treatment. The site was also chosen because of its southern aspect, which was intended to increase the effects of the reduced moisture.

Soil Solution and Soil Element Flux Measurements
O Horizon Nitrogen and Phosphorus Fluxes
Resin lysimeters (as described by Susfalk, 2000, and Johnson et al., 2001) were added in 1998 beneath the O horizons of all treatments to collect cumulative soil N and P leaching. The O horizon was too thin for installation of ceramic cup lysimeters, and we have had difficulties with plate tension lysimeters in O horizons at this site in the past because of frequent drying episodes during which lysimeters lose tension. Resin lysimeters are tension free and thus are not subject to problems with wetting and drying. Resin lysimeters are also very cost effective in that only one analysis per year per sampling location is required. Resin lysimeters have been used successfully to obtain indices of leaching in several previous studies and microbial transformations have generally been found to be minimal (Kjønass, 1999; Schnabel, 1983; Schnabel et al., 1993; Sibbesen, 1978; Susfalk, 2000).

The resin lysimeters used initially (in 1998) in this study consisted of a 5.5-cm-long, 4-cm-i.d. PVC pipe within which a resin bag was sandwiched between layers of washed silica sand. Ten grams of oven-dried Rexyn I-300 (H-OH) resin (Fisher Scientific, Lari Lawn, NJ) were placed in a section of nylon pantyhose, using cable ties to secure each end. This resin bag was placed on a 20 g layer of moist, washed silica sand at the bottom of the tube and covered with another 20 g layer of silica sand. In 1999, nylon resin bags were replaced by resin trapped between two Nitex nylon sheets (ICN Biomedicals, Aurora, OH). Susfalk (2000) found that this change did not result in a significant difference in the N or P trapped by the resin lysimeters in a separate study. In order to keep the sand in the tube until installation, the bottom of each PVC tube was covered with cheesecloth held in place with a rubber band. The first set of resin lysimeters were installed in June 1998 by excavating a small hole and tunneling beneath the O horizons. The resin lysimeters were co-located within 2 m of the cluster of ceramic cup lysimeters (n = 9 per treatment and strata). For the DRY and WET treatments, the resin lysimeters were stratified according to position to account for the influence of trough collection (DRY) and drip redistribution (WET) systems. In the DRY treatment, one set of resin lysimeters was placed directly beneath a trough and one set between troughs at each location (n = 9 for each, for a total of 18 replicates). For the WET treatment, one set of resin lysimeters was placed beneath an area that was receiving water from a driphole and one that was not (n = 9 for each, for a total of 18 replicates). The resin lysimeters were removed in June 1999 and replaced with the modified (1999) design. The 1999 resin lysimeters were removed and replaced in June 2000.

After the collection, resins were removed from the lysimeters, placed into 250-mL Erlenmeyer flasks, and extracted with 100 mL of 1.0 M KCl with shaking for 1 h. The extract was filtered (Whatman [Maidstone, UK] No. 1) and stored at 4°C until analysis. The extracts were analyzed for NH⁺₄, NO^-₃, and ortho-P by automated colorimetric analysis at the Water Analysis Laboratory at the Desert Research Institute, Reno, NV. Three 10-g replicates of untreated resins were extracted in the same way and served as blanks. Fluxes were calculated from the amount of NH⁺₄, NO^-₃, and ortho-P extracted from the resins (minus blanks) divided by the surface area of the lysimeters (12.6 cm²). Due to a laboratory error, ortho-P was not measured in the 1998–1999 collection.

Mineral Horizon Soil Solution Collection
Soil solutions from the E (25 cm) and Bt (70 cm) horizons were collected with ceramic cup tension lysimeters (Soil Moisture Equipment Corp., Santa Barbara, CA). The lysimeters were installed prior to the initiation of treatment in a three by three array in each plot at a 7.9-m spacing (nine replicates per treatment and depth). Some pre-treatment collections were made in 1992 and reported earlier (Johnson et al., 1998). Post-treatment sampling was initiated in 1996. Prior to collection of samples, each lysimeter was set to a tension of -40 kPa. Soil solutions were analyzed for electrical conductivity (EC), pH and HCO^-₃ (by titration to pH 5.0), Ca²⁺, Mg²⁺, K⁺, and Na⁺ by atomic absorption; NH⁺₄, NO^-₃, and ortho-P by automated colorimetric analysis; and Cl^- and SO^2-₄ by Dionex (Sunnyvale, CA) ion chromatography. In most cases (64% of all samples), pH values were 5.0 or less and thus HCO^-₃ concentrations were 0. Of the remaining cases where pH values were above 5.0, there was insufficient sample for titrations in 31% of the samples, so these analyses were omitted in favor of analyses for other ions (which take much less volume). Thus, HCO^-₃ analyses are lacking in 11% of the samples overall and it was not possible to include HCO^-₃ in all charge balance calculations. Within the data set containing HCO^-₃, the charge balance of Ca²⁺ + Mg²⁺ + K⁺ + H⁺ + Na⁺ (major cations, neglecting NH⁺₄ and Al) against HCO^-₃ + Cl^- + SO^2-₄ (major anions, neglecting ortho-P and NO^-₃) indicated an anion deficit at higher total cation concentrations in the E horizon soil solutions and a slight cation deficit in the Bt horizon solutions (Fig. 1) . The anion deficit in the E horizons could be due to unmeasured organic anions, and the cation deficit in the Bt horizon could be due to Al, which was not measured (pH values were often low enough for significant amounts of Al to be present in measurable concentrations).

View larger version (24K): [in this window] [in a new window]   Fig. 1. Plot of total anions against total cations in soil solutions from the E and Bt horizons.

View larger version (26K): [in this window] [in a new window]   Fig. 2. Soil water potential at 0 to 30 cm in the treatment plots from 1 Jan. 1996 through 31 Dec. 2000.

Forest Floor Sampling
Organic (O) horizons were sampled in February 1999. Sample locations were designated in each treatment plot along upper, middle, and lower slope transects. In each treatment area five locations were designated per transect for a total of 15 sampling locations in each treatment area. To minimize any influence of trough placements on O horizon accumulation or decomposition, post-treatment sampling within the dry treatment plot included twice the area (i.e., one under and one between troughs) per location. Identical sampling locations were avoided between years.

O horizons were sampled within 0.25 m² circular templates. Coarse twigs and large pieces of nonfoliar litter were removed first followed by the collection of the foliar component of the Oi horizon. The Oe and Oa horizons were sampled together because Oa horizons were often too thin to sample separately. All samples were oven-dried for 48 h (100°C) and dry mass was determined. Complete samples were subsequently ground in a large Wiley mill to a 20 mesh and a portion of the homogenized ground material was combusted at 500°C to obtain ash content. Total dry mass per unit ground area (m²) for each organic horizon was corrected for ash content. Other subsamples were ground to a fine powder and analyzed for C and N on a PerkinElmer (Wellesley, MA) 2400 CHNS analyzer at the Desert Research Institute, Reno, NV.

Statistical Analyses
Statistical analyses were performed using the General Linear Model (GLM) in DataDesk software (Data Description, 1988; Velleman, 1997). Within GLM, analysis of variance was used for overall treatment effects and Bonferonni post-hoc tests for differences between AMB, DRY, and WET treatments. Because of high variability, the data for the resin lysimeter fluxes was log-transformed for statistical analysis.

	RESULTS

TOP ABSTRACT INTRODUCTION SITE AND METHODS RESULTS DISCUSSION REFERENCES

 
Hydrologic Fluxes
Precipitation amounts and simulated soil water fluxes for water years 1996 through 2000 are shown in Table 2. Precipitation values for the AMB treatment represent field measurements and those for the DRY and WET treatments represent 0.67 and 1.33 times those values, respectively. As noted in previous papers (Johnson et al., 1998, 2001), these simulations suggest that soil water flux (SWF) is affected disproportionately by treatments: reducing precipitation by -33% in the DRY treatments caused decreases in simulated SWF of -46 to -59%, and increasing precipitation by +33% in the WET treatment increases simulated SWF by +44 to +57% (Table 2). It is also noteworthy that annual and seasonal precipitation input to the TDE site varied widely during the four-year sampling period (Table 2).

O Horizon Fluxes
The fluxes of NH⁺₄, NO^-₃, NH⁺₄ plus NO^-₃, and ortho-P measured by the resin lysimeters are given in Fig. 3 and the statistical analyses are presented in Table 3. (Note that ortho-P values were not measured in the 1998–1999 collection.) As expected, results for the resin lysimeters were sensitive to their placement with regard to the troughs in the DRY treatment and with regard to the drip locations on the WET treatment. Between troughs of the DRY treatment (DRY–open, DO) and between the drip holes of the WET treatment (WET–open, WO) resin-derived fluxes from the O layer were not significantly different from the fluxes in the ambient (A) treatment. Fluxes were, however, strongly affected when resin lysimeters were placed underneath troughs or near drip holes. Fluxes of NH⁺₄, NO^-₃, and NH⁺₄ + NO^-₃ beneath the troughs (DRY–treatment, DT) were significantly (p < 0.05) lower than in the lysimeters located in the open (A, DO, and WO). Fluxes of NH⁺₄, NO^-₃, and NH⁺₄ + NO^-₃ near drip holes (WET–treatment, WT) were significantly (p < 0.05) greater than in the other treatments. Differences in NH⁺₄, NO^-₃, and NH⁺₄ + NO^-₃ fluxes between the WT and other treatments were statistically significant (p < 0.01) in all cases. Fluxes of NH⁺₄, NO^-₃, and NH⁺₄ + NO^-₃ in the DRY, undertrough (DT) treatment were significantly lower than in any other treatment. Differences in ortho-P fluxes were significant only in a few cases (WT–DT, WO–DT, and, accepting the 0.06 probability level, WT–DO). There were no statistically significant effects of date on NH⁺₄, NO^-₃, and NH⁺₄ + NO^-₃ fluxes. (The effects of date could not be tested on ortho-P fluxes because of the lack of data for 1998–1999).

View larger version (45K): [in this window] [in a new window]   Fig. 3. Measured fluxes of NH+4, NO-3, NH+4 + NO-3, and ortho-phosphate in resin lysimeters. Note that ortho-P was not measured in 1998–1999. Standard errors are shown. See Table 3 for statistical analyses.

Soil Solution Concentrations
There were no statistically significant treatment effects on E horizon (25 cm) soil solution concentrations of any measured ion (Fig. 4 and 5 ; Table 4). There were significant effects on EC, where the values for the DRY treatment were significantly greater than those for the WET treatment, and for the (Ca²⁺ + Mg²⁺) to K⁺ ratio, where the values for the AMB treatment were significantly greater than for the WET treatment (Fig. 4 and 5; Table 4). Date was a significant factor for pH, EC, Mg²⁺, Na⁺, SO^2-₄, Cl^-, NO^-₃, and SO^2-₄ to Cl^- ratio. Treatment x date was a significant factor for pH and EC, indicating that treatment differences changed over time.

View larger version (27K): [in this window] [in a new window]   Fig. 4. Concentrations of Ca2+, K+, Mg2+, Na+, and Cl- in E horizon soil solutions (25 cm). See Table 4 for statistical analyses.

View larger version (29K): [in this window] [in a new window]   Fig. 5. Sulfate, pH, electrical conductivity (EC), SO2-4 to Cl- ratios, and (Ca2+ + Mg2+) to K+ ratios in E horizon soil solutions (25 cm). See Table 4 for statistical analyses.

View larger version (31K): [in this window] [in a new window]   Fig. 6. Concentrations of Ca2+, K+, Mg2+, Na+, and Cl- in Bt horizon soil solutions (70 cm). See Table 5 for statistical analyses.

View larger version (31K): [in this window] [in a new window]   Fig. 7. Sulfate, pH, electrical conductivity (EC), SO2-4 to Cl- ratios, and (Ca2+ + Mg2+) to K+ ratios in Bt horizon soil solutions (70 cm). See Table 5 for statistical analyses.

Calculated Ion Fluxes in Mineral Soil
Calculated fluxes of Ca²⁺, K⁺, Mg²⁺, Na⁺, Cl^-, and SO^2-₄ for the Bt horizon are shown in Fig. 8 . Error bars are not shown because the flux values are calculated from modeled soil water flux data, and thus it is not possible to conduct statistical analyses of the flux data.

View larger version (35K): [in this window] [in a new window]   Fig. 8. Calculated fluxes of Ca2+, K+, Na+, Mg2+, Cl-, and SO2-4 in the Bt horizons. Values are based on average concentrations and modeled soil water flux. Errors bars represent standard errors associated with averge concentrations only. AMB = ambient, no treatment; WO = WET treatment, placed in the open away from drip points; DO = DRY treatment, placed between troughs; WT = WET treatment, placed near drip points; DT = DRY treatment, placed beneath troughs; DRY-Avg = area-weighted average fluxes in the DRY treatment.

Litter Mass and Nitrogen Content
The mass and N content of standing litter in February 1999 is presented in Table 6. There were no statistically significant differences in mass or N content in Oi horizons, but significant differences in mass and N content in the Oe + Oa, nonleaf component, and in total O horizon were noted (Table 6). The differences were due primarily to greater Oe + Oa mass in the DRY treatment; greater nonleaf mass in the DRY treatment also contributed to the overall difference. The DRY and WET treatments had lower Oi horizon N concentrations and greater Oe + Oa horizon N concentrations than the AMB treatments. Differences in standing litter layer mass were not attributable to differences in litter inputs, which were not significantly different among treatment plots from 1994 through 1998 (Joslin et al., 2000).

	DISCUSSION

TOP ABSTRACT INTRODUCTION SITE AND METHODS RESULTS DISCUSSION REFERENCES

Differences in throughfall fluxes alone could not account for all the observed differences in soil solution concentration; indeed, if throughfall fluxes were the only factor involved, there would be no differences in soil solution concentration (i.e., both ion and water fluxes would change by ±33% and concentrations would stay the same). It is clear from the treatment effects on soil solution concentration that soil water flux must have changed disproportionately in response to changes in throughfall in the DRY treatment, at least. It is not clear why soil solution concentrations in the WET treatment were not more substantially affected than they were. The lack of response in soil solution concentrations in the WET treatment caused soil leaching flux estimates to increase substantially above those in the AMB treatment. As noted above, some of this was certainly due to increased throughfall fluxes. Regardless of the contributions of throughfall fluxes to soil leaching fluxes, however, any change in soil leaching flux represents a net loss from the system, whereas throughfall fluxes include a proportion of atmospheric dry deposition and ions that have been recycled by vegetation (especially in the case of K⁺). Thus, changes in soil leaching fluxes can be expected to have long-term consequences for soil change.

Evaluation of Hypotheses
The first hypothesis (i.e., changes in throughfall input will cause concomitant changes in N and P fluxes in O horizons) was supported for N but not P. Nitrogen fluxes were significantly greater in the WET treatment and lower in the DRY treatment than in the AMB treatment. This effect was exclusively due to those lysimeters placed either under troughs (DRY) or near drip points (WET); the lack of difference between those lysimeters placed in the open (AMB, WO, DO) is reassuring that the plots were comparable in other ways.

The second hypothesis (i.e., changes in throughfall input will cause inverse changes in EC and ion concentrations in soil solutions) was supported only in part. The DRY treatment did continue to cause increases in ion concentrations in Bt horizon soil solutions, but there were no statistically significant differences in E horizon solutions nor any differences between the WET and AMB treatments in the Bt horizons. The reasons for the greater treatment effect on the Bt than on E horizon solutions are not clear, but may relate to increased differences in soil water flux with increasing depth. Although most roots and therefore most evapotranspiration effects are concentrated in the surface horizons, water removal by roots does occur at depths greater than 25 cm, and, as discussed above, this could magnify treatment effects on soil water flux with depth. This hypothesis remains to be tested, however. The reasons for the lack of soil solution responses to the WET treatment are not known and merit further investigation.

The third hypothesis (i.e., changes in throughfall input will cause concomitant changes in SO^2-₄ to Cl^- ratios) was supported for Bt horizons: SO^2-₄ to Cl^- ratios were lower in the DRY treatment, suggesting that there was some buffering of soil solution SO^2-₄ concentrations by adsorption. The Bt horizon solution SO^2-₄ to Cl^- ratios declined and converged by water year 2000 when soil solution concentrations of Cl^- (and Na⁺) increased in an apparent response to drought (Fig. 6). The Bt horizon SO^2-₄ concentrations showed no change during water year 2000, however, again suggesting buffering and supporting the third hypothesis. Soil solution SO^2-₄ concentrations were not completely controlled by buffering, however: SO^2-₄ concentrations were consistently greater in the DRY than in the other treatments.

The fourth hypothesis (i.e., changes in throughfall input will cause inverse changes in [Ca²⁺ + Mg²⁺] to K⁺ ratios) was supported in part: (Ca²⁺ + Mg²⁺) to K⁺ ratios were significantly lower in WET than in the DRY treatments in the E horizons and significantly lower in the WET than in the AMB treatment in both horizons. In a previous paper (Johnson et al., 1998), we noted no significant treatment effects on (Ca²⁺ + Mg²⁺) to K⁺ ratio.

General Observations
The resin lysimeters allowed us to measure treatment effects on N fluxes for the first time. Treatment effects on N fluxes could not previously be measured because soil solution NH⁺₄ and NO^-₃ concentrations were at or near trace levels at all times. Measurement of forest floor fluxes with tension lysimeters is difficult because the forest floor is thin and subject to extreme wetting and drying cycles. Also, resin lysimeters avoid the problems of exchange or adsorption that may be associated with ceramic lysimeters (Lajtha et al., 1999). Susfalk (2000) noted that resin lysimeters collected substantially greater amounts of NH⁺₄, NO^-₃, and ortho-P than ceramic cup lysimeters in a comparative study in the Sierra Nevada mountains of Nevada. The differences were thought to be due to adsorption by the ceramic and perhaps also because of spoilage of water samples between collections.

With the resin lysimeters, we were able to detect large and statistically significant treatment effects on N fluxes from the forest floor (O horizons). The magnitudes of the N fluxes measured were within values that would be expected in this ecosystem. Nitrogen fluxes via litterfall and crownwash in this forest type are 34 and 3 kg N ha^-1 yr^-1, respectively (Johnson and Van Hook, 1989); thus, the resin lysimeter values suggested that there was a net uptake of N (by soil organisms, tree roots, or both) from the forest floor in the AMB and DRY treatments, as would be expected (with average fluxes of approximately 10 and 5 kg N ha^-1 yr^-1, respectively).

The differences in forest floor N content between treatments provide corroborating evidence that decomposition and N mineralization from litter are being reduced by the DRY treatment. Nitrogen content of litter in the DRY treatment was about 70 kg N ha^-1 greater than the AMB treatment (Table 6), suggesting a mean annual rate of N immobilization in the dry plot organic horizons of 12 Kg N ha^-1 yr^-1. This rate of immobilization is equal to about three times the difference in aerially weighted mineral N flux between the AMB and DRY plots in 1998–1999 (4.3 kg ha^-1 yr^-1; Fig. 3). In the 1999–2000 resin lysimeter collection, the aerially weighted fluxes in the DRY and AMB treatments were virtually identical (7.4 and 6.7 kg ha^-1 yr^-1, respectively), even though the fluxes beneath the troughs remained substantially and significantly lower in the DRY treatment. These differences could reflect error in estimation or they could reflect a leveling off of forest floor mass and N in the DRY treatment. Both the forest floor sampling in 1999 and the resin lysimeter data suggest greater sequestration of N in the forest floor in the DRY treatment.

Implications for Long-Term Soil Change and Plant Productivity
In a previous paper, we explored the implications of the treatments for long-term soil change using the Nutrient Cycling Model (NuCM; Johnson et al., 1998). The model predictions at that time indicated that, compared with the AMB treatment, soils in the DRY treatment would become enriched in base cations over a 30-year period, whereas those in the WET treatment would become depleted. The magnitudes of these changes did not suggest any measurable effect on plant productivity. These predictions remain to be tested. The NuCM also predicted that, because of slightly higher litter production in the WET treatment, forest floor mass and nutrient contents would increase relative to the AMB treatment, and the reverse would occur in the DRY treatment (lower forest floor mass and nutrient contents). This prediction was clearly not supported by the results of this study, where a significant increase in mass and N content in the DRY treatment and no effects in the WET treatment were found.

Perhaps the most significant effects relative to plant productivity noted in this study were the changes in O horizon leaching rates and the corresponding increases in forest floor N content in the DRY treatment. These effects are statistically significant, and excessive immobilization (or reduced mineralization) of N in the forest floor might reduce N availability for growth if sustained over long time periods. However, the rates of N immobilization observed here were relatively small and we hypothesize that they will eventually decline as a new steady state in the forest floor mass is reached.

	ACKNOWLEDGMENTS

 
This research was sponsored by the Program for Ecosystem Research, Environmental Sciences Division, Office of Health and Environmental Research, U.S. Department of Energy and the Nevada Agricultural Experiment Station, University of Nevada, Reno. The contributions by Hanson and Todd were conducted under contract DE-AC05-00OR22725 with University of Tennessee-Battle, LLC. Research was conducted on the Oak Ridge National Environmental Research Park.

	REFERENCES

TOP ABSTRACT INTRODUCTION SITE AND METHODS RESULTS DISCUSSION REFERENCES

 

• Cook, E.R., M.A. Kablack, and G.C. Jacoby. 1988. The 1986 drought in the southeastern United States: How rare an event was it? J. Geophys. Res. 93(D11):14257–14260.
• Data Description. 1988. DataDesk. Data Description, Ithaca, NY.
• Drungil, C.E.C., T.J. Gish, and K. Abt. 1987. Soil moisture determination in gravelly soils with time domain reflectometry. Trans. ASAE 32:177–180.
• Hanson, P.J., D.E. Todd, and J.S. Amthor. 2001. A six year study of sapling and large-tree growth and mortality responses to natural and induced variability in precipitation and throughfall. Tree Physiol. 21:345–358.[ISI][Medline]
• Hanson. P.J., D.E. Todd, Jr., N.T. Edwards, and M.A. Huston. 1995. Field performance of the Walker Branch Throughfall Displacement Experiment. p. 307–313. In A. Jenkins et al. (ed.) Ecosystem manipulation experiments: Scientific approaches, experimental design, and relevant results. Ecosyst. Res. Rep. #20. Commission of the European Communities, Brussels.
• Hanson, P.J., D.E. Todd, M.A. Huston, J.D. Joslin, J. Croker, and R.M. Augé. 1998. Description and field performance of the Walker Branch Throughfall Displacement Experiment: 1993–1996. ORNL/TM-13586. Oak Ridge Natl. Lab., Oak Ridge, TN.
• High Performance Systems. 2000. Ithink. High Performance Systems, Hanover, NH
• Huff, D.D., R.J. Luxmoore, J.B. Mankin, and C.L. Begovich. 1977. TEHM: A terrestrial ecosystem hydrology model. ORNL/NSF/EATC-27. Oak Ridge Natl. Lab., Oak Ridge, TN.
• Johnson, D.W., D.W. Cole, F.W. Horng, H. Van Miegroet, and D.E. Todd. 1981. Chemical characteristics of two forested Ultisols and two forested Inceptisols relevant to anion production and mobility. ORNL/TM-7646. Oak Ridge Natl. Lab., Oak Ridge, TN.
• Johnson, D.W., P.J. Hanson, D.E. Todd, R.B. Susfalk, and C.F. Trettin. 1998. Precipitation change and soil leaching: Field results and simulations from Walker Branch Watershed, Tennessee. Water Air Soil Pollut. 105:251–262.
• Johnson, D.W., D.D. Richter, G.M. Lovett, and S.E. Lindberg. 1985. The effects of atmospheric deposition on potassium, calcium, and magnesium cycling in two deciduous forests. Can. J. For. Res. 15: 773–782.
• Johnson, D.W., R.B. Susfalk, H.L. Gholz, and P.J. Hanson. 2001. Simulated effects of temperature and precipitation change in several forest ecosystems. J. Hydrol. 235:183–204.
• Johnson, D.W., and D.E. Todd. 1990. Nutrient cycling in forests of Walker Branch Watershed: Roles of uptake and leaching in causing soil change. J. Environ. Qual. 19:97–104.
• Johnson, D.W., and R.I. Van Hook (ed.) 1989. Analysis of biogeochemical cycling processes in Walker Branch Watershed. Springer–Verlag, New York.
• Jones, E.A., D.D. Reed, G.D. Mroz, H.O. Liechty, and P.J. Cattelino. 1993. Climate stress as a precursor to forest decline: Paper birch in northern Michigan, 1985–1990. Can. J. For. Res. 23:229–233.
• Joslin, J.D., M.H. Wolfe, and P.J. Hanson. 2000. Effects of altered water regimes on forest root systems. New Phytol. 147:117–129.
• Kirschbaum, M.U. 1995. The temperature dependence of soil organic matter decomposition, and the effect of global warming on soil organic C storage. Soil Biol. Biochem. 27:753–760.
• Kjønass, O.A. 1999. In situ efficiency of ion exchange resins in studies of nitrogen transformation. Soil Sci. Soc. Am. J. 63:399–409.[Abstract/Free Full Text]
• Lajtha, K., W.M. Jarrell, D.W. Johnson, and P. Sollins. 1999. Collection of soil solution. p. 166–182. In G.P. Robertson et al. (ed.) Standard soil methods for long-term ecological research. Oxford Univ. Press, New York.
• Lamersdorf, N.P., C. Bier, K. Blanck, M. Bredemeier, T. Cummins, E.P. Farrel, L. Rasmusson, and M. Ryan. 1995. Reactions of soil solution chemistry to drought: Results of the EXMAN project. p. 86–95. In A. Jenkins (ed.) Ecosystem manipulation experiments: Scientific approaches, experimental design, and relevant results. Ecosyst. Res. Rep. #20. Commission of the European Communities, Brussels.
• Likens, G.E., F.H. Bormann, R.S. Pierce, J.S. Eaton, and N.M. Johnson. 1977. Biogeochemistry of a forest ecosystem. Springer–Verlag, New York.
• Luxmoore, R.J. 1983. Water budget of an eastern deciduous forest stand. Soil Sci. Soc. Am. J. 47:785–791.[Abstract/Free Full Text]
• Luxmoore, R.J., and D.D. Huff. 1989. Water. p. 197–232. In D.W. Johnson and R.I. Van Hook (ed.) Analysis of biogeochemical cycling processes in Walker Branch Watershed. Springer–Verlag, New York.
• Reuss, J.O. 1983. Implications of the Ca–Al exchange system for the effect of acid precipitation on soils. J. Environ. Qual. 12:591–595.[Abstract/Free Full Text]
• Schnabel, R.R. 1983. Measuring nitrogen leaching with ion exchange resin: A laboratory assessment. Soil Sci. Soc. Am. J. 47:1041–1042.[Abstract/Free Full Text]
• Schnabel, R.R., S.R. Messier, and R.F. Purnell. 1993. An evaluation of anion exchange resin used to measure nitrate movement through soil. Commun. Soil Sci. Plant Anal. 24:863–879.
• Sibbesen, E. 1978. An investigation of the anion-exchange resin method for soil phosphate extraction. Plant Soil. 50:305–321.
• Susfalk, R.B. 2000. Relationships of soil-extractable and plant-available phosphorus in forest soils of the eastern Sierra Nevada. Ph.D. diss. Univ. of Nevada, Reno.
• Topp, G.C., and J.L. Davis. 1985. Measurement of soil water content using time domain reflectometry (TDR): A field evaluation. Soil Sci. Soc. Am. J. 49:19–24.[Abstract/Free Full Text]
• Ulrich, B. 1983. Soil acidity and its relation to acid deposition. p. 127–146. In B. Ulrich and J. Pankrath (ed.) Effects of accumulation of air pollutants in ecosystems. D. Reidel Co.
• Velleman, P.F. 1997. DataDesk Version 6.0 statistics guide. Data Description, Ithaca, NY.

This article has been cited by other articles:

				W. Wilcke and J. Lilienfein Nutrient Leaching in Oxisols Under Native and Managed Vegetation in Brazil Soil Sci. Soc. Am. J., June 2, 2005; 69(4): 1152 - 1161. [Abstract] [Full Text] [PDF]

				J. L. Langlois, D. W. Johnson, and G. R. Mehuys Adsorption and Recovery of Dissolved Organic Phosphorus and Nitrogen by Mixed-Bed Ion-Exchange Resin Soil Sci. Soc. Am. J., May 1, 2003; 67(3): 889 - 894. [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 Similar articles in ISI Web of Science Similar articles in PubMed 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 (9) Citing Articles via Google Scholar Google Scholar Articles by Johnson, D.W. Articles by Todd, D.E. Search for Related Content PubMed PubMed Citation Articles by Johnson, D.W. Articles by Todd, D.E., Jr. GeoRef GeoRef Citation Agricola Articles by Johnson, D.W. Articles by Todd, D.E. Related Collections Watershed and Landscape Processes Soil Chemistry Forest Soils