| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Civil & Environmental Engineering, Syracuse Univ., 220 Hinds Hall, Syracuse, NY 13244 USA
b U.S. Geological Survey, 425 Jordan Rd., Troy, NY 12180 USA
cejohns{at}mailbox.syr.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and methodsResultsDiscussionConclusionsREFERENCES |
|---|
in Oa horizons and mineral soil, respectively), exchangeable bases
, and CECe
, indicating the importance of organic matter to the chemistry of these acidic soils. The fraction of CECe occupied by H explained 44% of the variation in pHw. Soil chemical properties at this site vary on spatial scales finer than typical GIS analyses, resulting in relationships with poor predictive power. Thus, interrelationships among soil properties are more reliable for prediction.
Abbreviations: ANOVA, analysis of variance • CEC, cation-exchange capacity • CECe, effective cation-exchange capacity • DEM, digital elevation model • GIS, geographic information system • SOM, soil organic matter
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusionsREFERENCES |
|---|
es, 1986; Likens et al., 1996, 1998; Lawrence et al., 1999). Soil organic matter (SOM) has an important influence on soil physical, chemical, and biological properties (Brady, 1974; Tate, 1987). In addition to providing several essential plant nutrients, SOM is a major source of CEC (Bohn et al., 1989). Through exchange and complexation reactions, SOM influences the acid–base character of both soils and drainage waters (Cronan and Aiken, 1985; Lawrence et al., 1986; Tipping and Hurley, 1992). In cool, humid climates, leaching of organic solutes from the O horizon, and redeposition in the mineral soil is responsible for the accumulation of SOM (Buol et al., 1997). Consequently, the amount and distribution of nutrients in the soil profile is closely related to SOM dynamics. Notably, several studies of forest soils in New England have suggested that SOM provides essentially all of the exchange capacity in the soil (Federer and Hornbeck, 1985; Ross et al., 1991; Johnson et al., 1997).
A means of predicting spatial patterns of soil chemical properties such as SOM concentration, pH, and CEC would be useful in the study of forest ecosystem processes and watershed biogeochemistry (Pregitzer et al., 1983). Jenny (1980) proposed a state-factor model, in which the distribution of a soil property is related to climate, landscape position, organisms, time, and parent material. On small spatial scales, many of the these factors, and the interactions among them, are constant, allowing the researcher to focus on a limited subset of state factors. Here, we focus on the role of landscape position in explaining variations in soil properties.
Geographic information systems technology is a useful tool for economically estimating terrain attributes, quantities that express the position and orientation of points on the Earth surface. With a digital elevation model (DEM), these terrain attributes can be determined for an entire watershed or region at any spatial density up to the model resolution. If terrain attributes can be quantitatively related to soil properties, these properties could be estimated at high spatial density as well. Moore et al. (1993) explored this possibility at a 5.4-ha site in northeastern Colorado, and found significant correlations between several terrain attributes and A horizon thickness, sand and silt concentration, SOM concentration, pH, and extractable P. In their study, terrain attributes could collectively account for as much as one-half of the variation in the soil properties they studied. Because of the importance of leaching and decomposition in acidic forest soils, we hypothesized that landscape factors such as aspect, elevation, and water flowpaths would be related to SOM concentration and other soil chemical properties.
In this study, our goal was to examine the factors influencing soil chemistry in a forested watershed in the Catskills Mountains of New York. Because of elevated inputs of acid deposition, we focused on properties related to acid–base chemistry. We explored both the interrelationships among soil chemical properties and the relationships among soil chemistry and GIS-derived terrain attributes. There are three principal reasons for carrying out such an analysis. First, one can gain insight into factors that most strongly influence soil chemistry. Second, it may be possible to develop quantitative relationships to predict properties that are difficult or expensive to measure (e.g., CEC, base saturation) from measurements of properties that are easier to determine (e.g., terrain attributes, pH, total C). Finally, an understanding of the factors governing soil chemical properties may be useful in developing sound management strategies for forest lands subject to natural disturbances such as wind and ice storms, and anthropogenic disturbances such as acid deposition and forest harvesting. The Catskills are an important area in which to conduct this type of analysis because the region serves as the principal source of drinking water for New York City. Information on the spatial patterns of soil properties is useful in understanding controls on streamwater quality. This information can then be used to help develop watershed management policies to maintain the high quality of Catskills surface waters.
|
Materials and methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusionsREFERENCES |
|---|
3800 km2 of southeastern New York, and consist of flat-lying Devonian sedimentary rocks that have been uplifted and heavily eroded (Kudish, 1979). The bedrock, which consists of 60% sandstone and conglomerates, and 40% mudstone and siltstone (Stoddard and Murdoch, 1991), is overlain by Wisconsinan-aged glacial till that ranges in depth from near zero along ridge tops to 80 m in some valley bottoms.
Input of acidic deposition to the Catskills is among the highest in the USA, the result of both low precipitation pH and high precipitation amounts (Stoddard and Murdoch, 1991). There are no significant sources of air pollution within the region, and the prevailing winds originate from the west and southwest. In 1984 to 1996, SO2-4 concentrations in precipitation at an atmospheric monitoring station at Biscuit Brook, 7.8 km from our study site, declined by 40% (Burns, et al., 1998). There was no significant trend in precipitation NO-3 concentration during the same period. By 1996, NO-3 and SO2-4 deposition was approximately equal.
Six reservoirs in the Catskills supply 75% of the drinking water needs of New York City. Although this water has historically been of excellent quality, there are concerns regarding the possible long-term effects of acid deposition, climate change, and land-use practices on water quality. New York City has obtained a temporary exemption to the Surface Water Treatment Rule, which allows Catskills water to be used without filtration. To maintain this exemption, New York must develop watershed management policies that will insure the long-term quality of the Catskills water supply.
This study was carried out in the Winnisook watershed, a 214-ha headwater catchment located on Slide Mountain, the highest point in the Catskills (Fig. 1) . Water from Winnisook ultimately feeds the Neversink Reservoir, one of the six in the Catskills that serve New York City. The climate is characterized by cold winters and moderately cool summers (Lawrence et al., 1999). At the base of the watershed, the annual average air temperature is 4.3°C, ranging from -8.6°C in January to 16.7°C in July. Average annual precipitation is 1750 mm, with 23% falling as snow. Elevation in the watershed ranges from 840 to 1274 m, with an average watershed slope of 34%. Vegetation at Winnisook is dominated by northern hardwood species, American beech (Fagus grandifolia Ehrh.), yellow birch (Betula alleghaniensis Britt.), and sugar maple (Acer saccharum Marsh.). At upper elevations, pockets of balsam fir [Abies balsamea (L.) Mill.] mixed with white birch (Betula cordifolia Regel) can be found. There is some evidence of forest dieback in the upper reaches of the watershed.
|
Field Methods
A systematic soil sampling strategy was developed for this study (Fig. 1). First, the watershed was divided into eight subcatchments, corresponding with stream water collection points. Subcatchments A through G each contained 12 sampling sites, and the uppermost subcatchment (H) contained six sites. We ultimately abandoned sample collection in subcatchment H because of differences in vegetation, large gaps in the canopy, and evidence of forest dieback. Soil sampling sites were located in four bands running parallel to the stream channel, two on each side of the stream. Sites along the two inner bands were located within 20 to 40 m of the stream, while sites along the two outer bands were located 60 to 110 m from the stream. Sampling sites were located 75 m apart along each of the bands. This approach was developed for the purpose of relating soil chemistry to stream chemistry along a longitudinal gradient, an aspect of the study not discussed here; interested readers may see Ruiz-Méndez (1995) and Lawrence et al. (1999).
At each sampling site, the Oi and Oe horizons were removed. A representative sample of Oa horizon soil was collected by excavating the entire depth to mineral soil. The thickness of the Oa horizon was then measured. A single mineral soil sample was collected at each site by compositing 125-mL subsamples taken at 10-cm intervals to a maximum depth of 50 cm. Only four sampling sites had mineral soil deeper than 50 cm. All samples were air-dried and sieved prior to chemical analysis. Mineral and Oa samples were sieved using 2.00- and 4.76-mm sieves, respectively. Because of complications such as tree-throw, boulder fields, or an absence of soil, we ultimately collected 66 Oa samples and 51 mineral soil samples from the 72 sites.
Laboratory Methods
Soil pH was measured with a combination glass electrode in deionized water (pHw) at solution soil dilution ratios of 5:1 and 2.5:1 for Oa and mineral samples respectively.
Total C and N were determined by combustion gas chromatography on a Carlo Erba EA1108 elemental analyzer (Fisons Instruments, Danvers, MA). For elemental analysis only, soil subsamples were ground to pass a no. 200 sieve (0.075 mm) and dried overnight at 103°C. A 5– to 20–mg aliquot of the dried soil was encapsulated in tin and analyzed.
A mechanical vacuum extractor was used in the determination of exchangeable cations. For the base cations (Ca, Mg, K, and Na), 2.5 g of soil was extracted with 55 mL of 1 M NH4Cl for 12 h. Exact volumes of the extracts were determined gravimetrically. Concentrations of Ca, Mg, K, and Na in the extract solutions were determined by atomic absorption spectrophotometry.
This extraction procedure was repeated with 1 M KCl for the determination of exchangeable acidity, Al, and H (Thomas, 1982). Exchangeable acidity of the extracts was determined titrimetrically using 0.1 M NaOH to a phenolphthalein endpoint. An excess of 1 M KF was added, and the solution allowed to equilibrate for 30 min. Exchangeable Al was then determined by titrating back to the phenolphthalein endpoint with 0.1 M HCl. Exchangeable H was taken as the difference between exchangeable acidity and exchangeable Al.
Effective cation–exchange capacity was estimated by summing the exchangeable acidity and exchangeable base cation concentrations. The term effective refers to the use of neutral salts such as NH4Cl and KCl for the CEC determination. This approach results in the estimation of CEC near the native pH of the soil. In acidic soils, this value can be considerably lower than the total CEC measured in an extractant buffered at pH 7 to 8 (Skyllberg, 1999). Base saturation was calculated as the sum of the exchangeable base cation concentrations, divided by the CECe, and expressed as a percentage. All soil chemical concentrations are expressed on an oven-dry (103°C) mass basis.
Terrain Attributes
The terrain attributes used in this study were generated using a 5 by 5 m DEM. The DEM was generated from a U.S. Geological Survey 71/2-minute quadrangle map with a 6.10-m (20-ft) countour interval. Because of the steep topography in the watershed, we were able to generate a 5 by 5 m DEM without excessive interpolation. Five terrain attributes were calculated for each soil sampling site: elevation, slope, aspect, topographic index, and flow accumulation. Elevation (ELEV) and slope (SLOPE) were directly computed from the DEM. Visual observations in the field, and slope estimated with a clinometer, indicated that the GIS-derived estimates were reasonable.
Aspect was determined as the azimuth value (degrees) of a horizontal projection of the line perpendicular to a plane with the sampling point at its center. This value was converted to the deviation from north (FROMN):
 | (1) |
The topographic index (TOPOIDX) is a hydrologically based index used to characterize the likely soil moisture regime (Moore et al., 1988, 1993; Lawrence et al., 2000)
 | (2) |
is the catchment area (m2) above the point at which TOPOIDX is being computed, and ß is the slope at the point, expressed as an angle. When the catchment area is large, or the slope angle ß is small, one would expect relatively high soil moisture. This corresponds with higher values of TOPOIDX. The flow accumulation (ACC) at any point is the number of DEM cells that drain to that point. With a DEM resolution of 5 by 5 m, each cell occupies 25 m2, so the ACC value may be converted to square meters by multiplying by 25 if desired.
Statistical Methods
STATISTICA, v. 5.5 was used for the statistical analyses (StatSoft Inc., Tulsa, OK). Analysis of variance (ANOVA) was used to test for differences in soil properties and terrain attributes among subcatchments and between inner and outer sampling bands. Separate two-factor analyses were performed for individual soil properties and terrain attributes. In these analyses, subcatchment and band were fixed factors, and interaction terms were included in the analyses. Multiple linear regression was used to develop relationships between terrain attributes and individual soil properties and relationships among soil properties. In the regressions involving terrain attributes, the attributes included in the regression models were determined by forward stepwise variable selection. In the regressions involving only soil properties, pHw and total C were taken as the independent variables. Because of the exploratory nature of this research, we chose a significance level of 0.10 for all hypothesis tests.
|
Results |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusionsREFERENCES |
|---|
. The elevational trend in the mineral soil was somewhat surprising since slope was greater at higher elevation as well (Table 2)
.
|
|
Total C exhibited a large range of values (Table 1). In some strongly mor-type O horizons, total C was 50% or greater. In other sites, where the O–mineral boundary was less distinct, total C was as low as 13.5%. Mineral soil total C was much lower, on average, than in the O horizon, but the highest mineral soil values were similar to the lowest in the O horizon.
Total N patterns were similar to total C patterns, with the average in the Oa horizon about six times the average in the mineral soil (Table 1). The mean C/N ratio declined from 20.8 in the Oa horizon to 12.0 in the mineral soil, indicating that organic matter in the mineral soil has a higher N content than in the Oa horizon.
There were many notable correlations among soil properties (Table 3) . Total C, in particular, was highly correlated with many soil properties. In the Oa horizon, total C was negatively correlated with pHw and exchangeable Al, and positively correlated with exchangeable base cations, CECe, and base saturation. Total C was also highly correlated with total N. In the mineral soil, total C was uncorrelated with pHw, and positively correlated with all of the exchangeable cations, CECe, and base saturation. As in the Oa horizon, total C was strongly correlated with total N in the mineral soil.
|
Horizon thickness was significantly, but weakly, correlated with a number of chemical properties (Table 3). In the Oa horizon, pHw was negatively correlated with thickness, while the rest of the soil properties were either positively correlated or uncorrelated with thickness. In contrast, in the mineral soil there was a positive correlation between thickness and pHw, but negative or insignificant correlations with the other soil chemical parameters.
There were few statistically significant correlations between terrain attributes and soil properties (Table 4) . Those that were significant were weak. The Oa and mineral soil thickness were positively correlated with elevation. In addition, Oa horizon thickness was positively correlated with slope and negatively correlated with topographic index. The soil chemical properties we studied, with a few exceptions, were uncorrelated to terrain attributes (Table 4).
|
Proximity to the stream had little influence on soil properties. Mineral soils sampled from the inner bands closer to the stream channel had 50% lower exchangeable H, and 24% lower exchangeable acidity than soils sampled from the outer bands, further from the stream. There were no significant differences between inner and outer bands in the chemistry of Oa horizons.
|
Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusionsREFERENCES |
|---|
Few studies have attempted to evaluate the potential of terrain analysis for the prediction of soil chemical properties. Moore et al. (1993) used a 15.24-m grid in a terrain analysis of a 5.4-ha toposequence in northeastern Colorado. They were able to explain 48.2, 48.3 and 40.9% of the variance in organic matter concentration, extractable P, and pH, respectively, with four terrain attributes.
In the Winnisook watershed, correlations between the GIS-derived terrain attributes and soil chemical properties were generally poor. As a result, predictive multiple-regression relationships using terrain attributes explained little of the variation in the soil chemical properties we studied (Table 4). The weakness of the relationships between terrain attributes and soil properties at Winnisook may be due to a number of factors, all of which are probably true to some degree. First, the attributes we included in our analysis may not have fully characterized the topographic features of our sampling points. The attributes we used included indices of vertical position (ELEV), inclination (SLOPE, TOPOIDX), spatial orientation (FROMN), and draining area (ACC, TOPOIDX). These parameters together describe the fundamental spatial characteristics of the sites, yet explain only 4 to 25% of the variation in the soil properties we studied. While there may be more sophisticated measures that would improve the quality of our predictions, it seems unlikely that new parameters could explain a substantial amount of the unexplained variance.
For all of the terrain attributes, the ranges in our data were large (Table 2). Thus, the poor relationships we observed cannot be attributed to lack of variation in the values of the terrain attributes. We conclude that the terrain attributes we used cannot be used effectively in the estimation of soil chemical properties at Winnisook. As an example, the relationship between the concentration of exchangeable base cations predicted from the terrain attributes and the observed values is shown in Fig. 2a . While the data points are distributed around the 1:1 line, the weak predictive power is clear.
|
Finally, factors other than topography are also likely to influence soil chemical properties. There are no major variations in parent material, land-use, or time of soil development at Winnisook. However, a distinct increase in acidic deposition with elevation at Winnisook has been observed, and appears to have resulted in increased base cation leaching at higher elevations in the watershed (Lawrence et al., 1999). Also, vegetation in the watershed is highly heterogeneous at both stand and individual-tree scales. Numerous investigators have reported significant chemical differences in soils lying below different tree species (Alban, 1982; Boettcher and Kalisz, 1990; Finzi et al., 1998a, 1998b) and within the canopy of individual trees (Riha et al., 1986). Integration of vegetation information may improve our ability to predict soil chemistry. Such an effort might benefit from the coupling of remote-sensing and GIS technologies.
Importance of Organic Matter
Our data are consistent with the hypothesis that organic matter is a controlling factor of soil chemistry at Winnisook. We observed significant, and generally positive, correlations between total C and most other chemical properties (Table 3). Also, significant negative correlations were observed between pHw and most chemical properties. In the Oa horizon, total C was negatively correlated with pHw. Because of the abundance of weakly acidic functional groups in soil organic matter, high C contents in the Oa horizon are accompanied by lower pH (e.g., Stevenson, 1994). Dissociation of carboxyl groups (—COOH), in addition to lowering pH, results in increased CECe and exchangeable cation concentrations. Accordingly, total C was positively correlated with CECe and exchangeable base cations in the Oa horizon, and pHw was negatively correlated with CECe and exchangeable bases (Table 3).
Because of the high correlations we observed, we can develop relationships to predict exchangeable cation concentrations from total C and pHw. These pedotransfer functions (e.g., Tietje and Tapkenhinrichs, 1993; Bell and van Keulen, 1995) are useful because total C and pHw are relatively easy and inexpensive to measure, compared with exchangeable cations and CECe. Hence, by measuring pHw and total C at a large number of sites, predicted values of exchangeable cations and CECe can be used to increase the number of data points for mapping purposes.
As an example, the sum of exchangeable bases in Oa horizons and mineral soils at Winnisook can be predicted from:
 | (3) |
 |
 | (4) |
 |
These relationships are much stronger than the regression relationships derived from terrain attributes, making them more suitable for modeling and prediction (Fig. 2b).
Cation-Exchange Capacity, pH, and Base Saturation
Through cation exchange, forest soils buffer changes in percolation water chemistry and acid deposition. Together, CEC, pH, and base saturation describe the acid–base status of a soil and its potential ability to buffer acid inputs. At Winnisook, CECe was positively correlated with total C in both the Oa horizon and mineral soil (Table 3, Fig. 3)
. As total C approaches zero, CECe also approaches zero (Fig. 3), suggesting that organic matter is the only significant source of exchange sites in Winnisook soils. This relationship appears to be nonlinear and is consistent with observations made at other sites in the northeastern USA with glacially derived soils of similar age (Ross et al., 1991; Johnson et al., 1997). Total C and pHw together can be used to predict CECe with the following equations:
 | (5) |
 |
 | (6) |
 |
|
54 cmolc (kg C)-1, about three times the CECe of organic matter in the Oa horizon. The negative contribution of pHw to CECe in Eq. [6] is somewhat surprising because at higher pH, greater dissociation of carboxyl groups of soil organic matter should increase CECe (Stevenson, 1994; Bohn et al., 1989). In the Oa horizon, the negative correlation between pHw and CECe can be explained largely by the negative correlation between pHw and total C (Table 3). The partial correlation of CECe and pHw, holding total C constant, was not statistically significant for the Oa horizon
. Furthermore, the regression coefficient for pHw in Eq. [5] was not significantly different from zero. In the mineral soil, however, there was a highly significant negative effect of pHw on CECe
. Other studies of acid forest soils of the northeastern USA have also reported insignificant or negative correlations between CECe and pHw in field samples (Johnson et al., 1991; Ross et al., 1991; Ross and Bartlett, 1995). The mechanism(s) underlying this phenomenon are unclear. When Ross et al. (1991) analyzed data from 148 samples of Spodosols and Inceptisols in a Vermont watershed, they observed no significant contribution of pHw to CECe. However, when individual samples were titrated with base, the CECe increased, as one would expect. Ross et al. (1991) suggested that soils with lower native pHw had inherently lower points of zero charge.
The sum of exchangeable base cations was positively correlated with total C in both Oa and mineral horizons (Table 3, Fig. 4) . To some extent, this relationship was due to the greater CECe associated with greater total C. Thus, the correlation between total C and base saturation was somewhat lower than the correlation between total C and the sum of base cations (Table 3, Fig. 4). Base saturation may be related to organic matter content for several reasons. One possibility is that there is insufficient Al present to saturate high-C, high-CECe soils (Ross et al., 1991). Consistent with this hypothesis, we observed a greater correlation between total C and base saturation in the Oa horizon than in the mineral soil (Fig.4), where Al is much more abundant. Alternatively, higher base saturation in high-C samples may be due to the presence and gradual release of bases from the organic matter during decomposition. In this way, base cations are relocated from the mineral soil to the O horizon through plant uptake and litterfall.
|
|
 | (7) |
|
|
Conclusions |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusionsREFERENCES |
|---|
Stronger predictions of soil chemical properties were possible when other soil properties (intrinsic factors) were used as predictors (Fig. 2). From a modeling perspective, this result is important because extensive measurement of properties that are inexpensive to determine, such as pHw and total C, can be used to augment a smaller number of more expensive analyses, such as CECe and exchangeable bases. Even using these better relationships, there is considerable uncertainty in predicted values, with R2 values in the 0.2 to 0.7 range (Eq. [3]–[6]). Therefore, prediction of chemical properties for a specific point is risky. However, making predictions for an array of points within a watershed for mapping purposes can be effective, since positive and negative errors would tend to cancel.
The relationships between total C, pHw, and the exchange chemistry of Winnisook soil were consistent with other studies carried out in the northeastern USA. Cation-exchange capacity is provided primarily by organic matter, and decreases with increasing pHw in the mineral soil (Fig. 3). Exchangeable Al behaves like a base cation in Winnisook Oa horizons, increasing with increasing pHw. The exchange behavior of Winnisook soils has important implications for the recovery of this ecosystem from chronic acid deposition. For soil pH to increase, the H/CECe ratio in the soil must decrease (Fig. 6). Thus, there is the potential for release of exchangeable H to buffer increases in the pH of percolating water. Furthermore, if increased pH is associated with higher exchangeable Al in the Oa horizon, and lower CECe in the mineral soil, as our data suggest, the base status of Winnisook soils may not readily improve after reductions in acid deposition. The extent to which soil change can be predicted from soil patterns, however, is unclear and merits further study.Cronan Aikin 1985; Paçes 1986
|
ACKNOWLEDGMENTS |
|---|
Received for publication August 19, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionConclusionsREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. M. Schwertfeger and W. H. Hendershot Determination of Effective Cation Exchange Capacity and Exchange Acidity by a One-Step BaCl2 Method Soil Sci. Soc. Am. J., May 1, 2009; 73(3): 737 - 743. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. J. Sullivan, I. J. Fernandez, A. T. Herlihy, C. T. Driscoll, T. C. McDonnell, N. A. Nowicki, K. U. Snyder, and J. W. Sutherland Acid-base Characteristics of Soils in the Adirondack Mountains, New York Soil Sci. Soc. Am. J., December 2, 2005; 70(1): 141 - 152. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. M. Ziadat Analyzing Digital Terrain Attributes to Predict Soil Attributes for a Relatively Large Area Soil Sci. Soc. Am. J., August 25, 2005; 69(5): 1590 - 1599. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. S. Gilliam, N. L. Lyttle, A. Thomas, and M. B. Adams Soil Variability along a Nitrogen Mineralization and Nitrification Gradient in a Nitrogen-Saturated Hardwood Forest Soil Sci. Soc. Am. J., January 1, 2005; 69(1): 247 - 256. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. J. Fernandez, L. E. Rustad, S. A. Norton, J. S. Kahl, and B. J. Cosby Experimental Acidification Causes Soil Base-Cation Depletion at the Bear Brook Watershed in Maine Soil Sci. Soc. Am. J., November 1, 2003; 67(6): 1909 - 1919. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
