SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Rodrigue, J. A.
Right arrow Articles by Burger, J. A.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Rodrigue, J. A.
Right arrow Articles by Burger, J. A.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Rodrigue, J. A.
Right arrow Articles by Burger, J. A.
Related Collections
Right arrow Soil Fertility and Productivity
Right arrow Forest Soils
Right arrow Ecosystem Restoration
Published in Soil Sci. Soc. Am. J. 68:833-844 (2004).
© 2004 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA

DIVISION S-4—SOIL FERTILITY & PLANT NUTRITION

Forest Soil Productivity of Mined Land in the Midwestern and Eastern Coalfield Regions

J. A. Rodrigue and J. A. Burger*

Dep. of Forestry (0324), 228 Cheatham Hall, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061

* Corresponding author (jaburger{at}vt.edu).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Our goal was to determine the effects of surface mining on forest land productivity in the eastern coalfields of the USA before the passage of the Surface Mining Control and Reclamation Act of 1977 (SMCRA), and to determine the extent to which selected mine soil properties influenced forest productivity. The site productivity of 14 mined and eight nonmined sites in the eastern and midwestern coalfields were compared. Results show that site productivity of nonmined sites and 12 of the 14 mined sites was similar. Sites with low productivity were shallow, had high coarse fragment contents, and had lower fertility. Regression analysis identified five influential soil properties affecting site quality, which included soil profile base saturation (BS), total coarse fragments, total available water, C horizon total porosity, and soil profile electrical conductivity (EC). These five properties explained 52% of the variation in tree growth. Forests on most prelaw mined sites were just as productive as the forests on unmined adjacent sites and can be used as a benchmark to assess the impacts of current reclamation on mine soil quality and forest productivity.

Abbreviations: AWHC, available water holding capacity • BS, base saturation • EC, electrical conductivity • ECEC, effective cation exchange capacity • SI, site index • SMCRA, Surface Mining Control and Reclamation Act • VDP, variance decomposition performance • VIF, variance inflation factors


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
SURFACE MINING has been disturbing land, forests, and waterways of the midwestern and eastern USA for over a century. Since the implementation of the SMCRA in 1978, U.S. Office of Surface Mining statistics show that 500000 ha have been disturbed by mining in the East (OSM, 1999). We estimate that an equivalent amount was disturbed before 1978, although the exact amount is not known. Before federal regulations controlling mining, the drastic nature of mining disturbance prompted some mine operators, landowners, and surrounding communities to reclaim mined areas (DenUyl, 1955), and several states with mining activity enacted regulations to control the mining process and minimize adverse effects (Davidson, 1981; Sandusky, 1980). In the midwestern and eastern states, many sites were reclaimed with trees to control erosion and reduce sedimentation. However, there was little or no expectation for commercial wood production, and the productivity of these mined lands is still largely unknown. Most mined sites planted to trees developed good vegetative cover for erosion control, but many other environmental problems remained, including degraded water quality, toxic spoils, uneven landscapes, acid drainage, high walls, and subsidence.

The SMCRA was enacted in 1977 to address human safety, land productivity, and environmental problems that occurred during mining and reclamation. However, in the process of attaining these goals, reforestation disincentives were created because the reclaimed landscape is difficult to plant to trees and it is commonly unproductive for forestry (Burger, 1999). Postlaw emphasis was placed on water quality and erosion control (Boyce, 1999) at the expense of site productivity, reforestation, C sequestration, and productive land uses. In many cases, reclamation in the Appalachian region results in mine soils that are alkaline, highly compacted, and covered with competitive grasses, which makes it difficult to re-establish forests and causes them to grow poorly (Burger, 1999). Nonetheless, the Code of Federal Regulations (1997, Section 715.13[a]) interpreting SMCRA requires that states restore disturbed land to conditions capable of supporting the uses that they supported before mining.

An example of the mine soil productivity problem was documented by Torbert et al. (2000), who reported 11-yr results of a test planting of three pine species on a prelaw mined site and a postlaw mined site. They established their study during the transition period when the new law was first implemented. Trees planted on the prelaw mined site were planted on the flat bench that remained after contour coal extraction, while the postlaw mined site was reclaimed to its "approximate original contour" required by the new law. The height and diameter growth of all three pine species [loblolly (Pinus taeda L.), Virginia (Pinus virginiana Mill.), and white (Pinus strobus L.)] was greater on the prelaw mined sites than on the postlaw mined sites. The heights on the prelaw mined sites averaged 7, 5.6, and 3.7 m, while the heights on the postlaw mined site averaged 6.7, 5.3, and 3.1 m for loblolly, Virginia, and white pine, respectively. The diameter growth on the prelaw mined site averaged 11.2, 9.7, and 5.3 cm, while on the postlaw mined site the diameters averaged 8.9, 7.4, and 3.6 cm. Projecting these growth rates to a harvest age of 20 yr indicates that stumpage value on the prelaw site will be approximately twice that of the postlaw site.

Unlike many other agricultural crops, there is no productivity standard in the regulations for forestland; a minimum number of trees surviving for the 5-yr bond period is the only performance standard associated with the tree component of forestland uses. Therefore, forestland productivity is commonly degraded in the process of mining and reclamation (Burger, 1999). Mine operators choose rock overburden material for the surface that supports vigorous herbaceous ground covers used for temporary erosion control. Research in the Appalachian coalfield region has shown that the type of overburden suitable for the temporary ground cover is not usually the best choice for long-term forest uses (Torbert, 1995). On midwestern sites, where topsoil is usually recovered and replaced, excessive grading compacts the C horizon and topsoil creating conditions unfavorable for tree establishment and growth (Pope, 1989).

To satisfy the intent of SMCRA, land reclaimed for forestry should logically meet a minimum productivity standard similar to that required for other crops; however, little is known about mine soil quality requirements for trees. Prelaw mined sites are growing forests in the midwestern and the eastern coalfields over a wide environmental gradient (Burger et al., 1998; Andrews, 1992; Plass, 1982). Productivity comparisons between prelaw mined sites and nonmined forest sites should show the extent to which mined sites can be reclaimed to premined productivity levels. Furthermore, characterization of mined sites growing productive forests should provide insight into soil and site conditions needed for reclaiming mined land for forestry uses. Therefore, the objectives of this study were to: (i) compare the site productivity of a range of surface mined sites to adjacent nonmined forests; and (ii) determine soil and site properties that influence tree growth and long-term productivity on reforested mined sites.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Fourteen forest sites across seven states, each with an average size of 2.5 ha of uniform and contiguous forest cover, were located on reclaimed mined lands in the midwestern and Appalachian coalfields (Fig. 1) . The 14 sites ranged from 20 to 55 yr old. The canopy layer species ranged from pure hardwood and conifer stands to mixed stands (Table 1). These sites also covered a broad spectrum of spoil types. The measurement sites were chosen to represent a cross-section of site, soil, and stand conditions. Adjacent, nonmined, native forest sites were also located and measured. These undisturbed control sites represented land and forest conditions similar to those present on the mined sites before they were disturbed. A study of premined topography and landscape using maps and onsite field surveys was used to confirm that site features and soil types were similar and comparable. All nonmined sites were mature, well-stocked, native forest stands, but all had been harvested to some degree at some point in their history. After the boundaries of each study site were established, a 20 by 20 m grid was superimposed across the site. Attempts were made to place grid lines perpendicular to spoil banks on open-pit mined sites where more than one spoil bank existed to ensure that the sites' microtopography was taken into account. A 20-m buffer strip was maintained on all edges of each forest site. Sampling was based on the intersections of the grid (Fig. 2) . Field data collection took place between May and August 1999, with the exception of two sites that were measured in August 1998.



View larger version (41K):
[in this window]
[in a new window]
  		Fig. 1. General location of study sites in the midwestern and Appalachian coalfields.

 

View this table:
[in this window]
[in a new window]
  		Table 1. Location, description, age, and site quality of study sites located in the Midwestern and Appalachian coalfields.

 


View larger version (102K):
[in this window]
[in a new window]
  		Fig. 2. Typical site and plot layout used at each study site.

 
Site productivity for forest growth is a function of soil, geology, topography, and climate. Site productivity is commonly estimated using site index (SI), the average height of codominant canopy trees at a selected age (e.g., Age 50) (Carmean, 1975). White oak (Quercus alba L.) was used as the indicator species and was always measured when present. To compare across sites growing different species, SI was standardized to white oak SI (Wenger, 1984). This site productivity estimate reflects the influence of soil and site characteristics while removing the effect of differences in tree species' growth patterns, tree age, and stocking levels. Sampled trees were healthy, intermediately shade-tolerant, in dominant or codominant positions in the canopy, and had been in free-to-grow positions for most of their lives. On each of four randomly chosen plots, tree height and total age were measured on one tree of the three main species in the canopy layer. Regional SI curves were used to estimate their height at Age 50 (Carmean et al., 1989). To make direct comparisons between mined sites and nonmined sites, SI estimates for each species were converted to a SI for white oak (base age 50 yr) using Doolittle's (1958) species comparison curves. Site index relationships among species that naturally grow together in the same stand are well correlated and their relationships are well documented (Carmean, 1975). White oak was used as the indicator species because it is a common species throughout the region; it is equally productive across the region; it is commonly planted on reclaimed sites; and it is an important commercial species.

Soil pits were dug at four randomly located plot centers on each site (Fig. 2). Pits were described using standard soil description techniques. Loose samples and duplicate bulk density samples were collected from each soil horizon of the profiles of both mined and nonmined sites. All minesoils were Entisols with AC profiles. The nonmined soils were Inceptisols, Ultisols, and Alfisols (Table 1). Soil properties were analyzed by horizon and profile average values were weighted by horizon depth. Soil samples were air-dried, sieved (2 mm), and weighed to determine coarse fragment content. Particle size was determined using the hydrometer method (Gee and Bauder, 1986). Bulk density and porosity, also corrected for coarse fragment content, were determined using soil cores. Porosity was determined using a tension table with a 50-cm water column (0.005 MPa). Available water holding capacity, water held between 0.03 and 1.5 MPa, was determined with a pressure plate apparatus (Klute, 1986). Soil pH was determined using a 1:2 soil/water mixture (McLean, 1982). Electrical conductivity was determined by a 1:5 soil/water extract (Rhoades, 1982). Total C was determined using a Leco C analyzer (LECO Corp, St. Joseph, MI), and pedogenic C was estimated by the Walkley-Black wet oxidation procedure (Nelson and Sommers, 1982). Exchangeable acidity was determined using the potassium chloride extraction technique (Thomas, 1982). Exchangeable cations (Ca, Mg, K, Na, Fe, Al, and Mn) were extracted with 1 M ammonium acetate (Thomas, 1982). Effective cation exchange capacity (ECEC) was estimated by summing the charge associated with exchangeable acidity and exchangeable Ca, Mg, K, and Na. Base saturation was calculated as the proportion of the ECEC occupied by base cations (Thomas, 1982). Total N was determined by Kjeldahl digestion and analyzed with a Bran and Luebbe TRACCS 2000 spectrophotometer (Bran+Luebbe Inc., Delavan, WI) (Bremner and Mulvaney, 1982). Phosphorus was extracted with sodium bicarbonate (Olsen and Sommers, 1982). Phosphorus and cations were determined using a Jarrell-Ash ICAP-9000 spectrophotometer (Jarrell-Ash, Grand Junction, CO).

Differences in SI between nonmined and mined study sites were tested using t tests. Regression analysis was used to determine the effects of mine soil properties on site productivity (SAS Institute, 1999). Soils on each site were thoroughly characterized both in the field and laboratory by measuring 37 properties that we hypothesized influenced tree growth (Table 2). Values of soil properties known to be nonlinear in their relationship with SI were transformed for linear regression analysis, and soil variables expressed as ratios or percentages were transformed using the arcsine function. The data set was then analyzed for multicollinearity, and soil properties high in multicollinearity were removed (filtered).


View this table:
[in this window]
[in a new window]
  		Table 2. Soil properties of the full profile, A horizon, and C horizon measured on all study sites. Those properties shown in italic type were used as independent variables in regressions with forest site index. Those properties shown in regular type were screened out of the variable list due to their multicollinearity with other soil variables.

 
The filtering process involved iterative steps using SAS software (SAS Institute, 1999). Each step involved examination of information from the variance-decomposition proportions (proportions of variation in SAS) and variance inflation factors (VIFs). The variance decomposition proportions (VDP) calculated the proportion of variance inflation contributed by each variable. A high VDP indicated that multicollinearity existed between variables. Those variables identified by the VDP were then crosschecked with the VIFs, which measure the combined effect of all the variables on the variance of one of the variables (Montgomery et al., 2001). More than one high VIF reinforces the presence of multicollinearity. For each step, the variable that contributed most to the multicollinearity was eliminated from the model. The filtering step was then repeated, removing one variable at each step until multicollinearity among variables was removed.

Three regression model selection procedures (R-square, Stepwise, MaxR) were used on the filtered mine soil dataset to determine the combination of mine soil properties that accounted for the variation in SI on the mined sites. The use of multiple model selection procedures allowed analysis of the data using each procedure's strengths, the examination of multiple models, and better knowledge of variable relationships. After regression analysis, the best model was selected based on criteria that included minimizing mean squared error (MSE) and maximizing individual variable significance, adjusted R2, R2, and biological significance. Results from statistical tests termed different in this paper have a significance level of p ≤ 0.10.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Mined Site Characterization and Productivity
Mined sites included in this study were located in seven states, were mined with different techniques, and were planted with a variety of tree species (Table 1). On the flat terrain of the midwest, most mining was done in open pits and overburden was cast in piles and left unleveled. However, IL-2 was leveled with a dragline; KY-1 and KY-2 were top graded, meaning that the tops of cast piles were struck off with a dozer. Other midwestern sites (IL-1, IN-1, IN-2, KY-3, KY-4) were left unleveled. The eastern sites were a mixture of open-pit and contour mined sites, with a highwall remaining on one side adjacent to a relatively flat, narrow bench with overburden piled on the outslope. OH-1, OH-2, and PA-1 were open-pit mined. OH-1 was top graded and OH-2 was left unleveled. PA-1 was open pit-mined and leveled by a dragline. The others were contour-mined (WV-1, WV-2, VA-1). Half of the midwestern sites were planted with hardwood species, while pines were commonly used on eastern sites. The oldest sites were planted in 1938 and the most recent in 1977.

Mined site productivity (SI) ranged between 16 and 27 m across both coalfields, averaging 24 m (Table 1). On midwestern mined sites, SI ranged from 23 to 28 m (24.6 m average). Site index on eastern mined sites varied between 17 and 29 m (23.8 m average). Nonmined site index was less variable, ranging from 23 to 28 m with an average of 25 m. The greater variation in SI on mined versus nonmined sites was probably due to the mix of mine spoil types and properties represented on our study sites.

Average site productivity of mined sites in the midwestern region was the same as that of their nonmined counterparts (Fig. 3) . Postmining SI across all midwestern sites varied within a range of ±10%, but these differences were not significant. Most importantly, this result indicates that prelaw mined sites in the midwestern coalfields are at least as productive after mining as before. Site productivity of the eastern sites was lower on two out of six. Though statistically similar, the SI of the mined Ohio sites appeared to be 10 to 15% higher than the nonmined condition. Both of the mined sites PA-1 and WV-1 clearly had lower site indices than the nonmined condition. The 32% difference in productivity on WV-1 was attributed to a soil with high coarse fragments and low BS. Average coarse fragment content was 82%, and the BS was 36%, the lowest BS level measured throughout the mined sites. Similar soil environments have been identified on poor-quality sites throughout the eastern coalfield region (Daniels and Amos, 1981; Torbert et al., 1994). Mined site PA-1 was similar to WV-1, with low BS and high coarse fragments. The site indices of PA-1 and WV-1 were 18 and 32% lower than those of their respective nonmined sites (WV-C1, PA-C, Fig. 3).



View larger version (24K):
[in this window]
[in a new window]
  		Fig. 3. Relative productivity of nonmined (zero baseline) and mined sites in the midwestern and eastern coalfields. Grey bars represent mined sites planted to pine species; open bars represent mined sites planted to hardwood species. *Difference between mined and nonmined sites (p ≤ 0.10). {dagger} Site index is the height of a white oak canopy at Age 50.

 
Mine Soil Properties and Site Productivity
Regression Model
Regression analyses were used to determine which mine soil properties most influenced forest productivity. One site property and 36 soil properties were included in the analysis (Table 2). The filtering process for multicollinearity, as described in the Methods section, identified soil properties causing multicollinearity in the C horizon and the total profile. Due to the young age of these soils and poor horizon development, the C horizon typically dominated the soil profiles on each site, resulting in few differences among nutrient values in the C horizon and the total profile. Two-thirds of the multicollinearity analysis involved removal of soil variables related to fertility. For example, exchangeable Ca in the C horizon, exchangeable Ca in the profile, and exchangeable Mg in the C horizon were found to be related. Exchangeable Mg contributed most to the multicollinearity so it was removed. In another example, bulk density and total porosity in the C horizon were correlated; bulk density in the C horizon was removed. After filtering for multicollinearity, the data set contained 22 independent soil and site variables, 12 physical properties, and 10 chemical properties (depicted in italic letters in Table 2) that were subsequently used as independent variables in the regression analyses. Each of the three regression procedures (R2, Stepwise, MaxR) selected the same best model, which included the same five soil variables. All five variables had partial tests that were significant in the model at p ≤ 0.10. The final model had one of the highest adjusted R2, one of the lowest MSE of models that met the partial test criteria, and all variables were biologically relevant to developing forests on surface mines. The model is:

[1]
where SIWO50 is SI for white oak at Age 50; ln arcsine (BS) is the natural log of the arcsine transformed base saturation; ln arcsine (CF) is the natural log of the arcsine transformed coarse fragment content; ln AWHC is the natural log of available water holding capacity; arcsineTPC is the arcsine transformed C horizon total porosity; and ln EC is the natural log of EC.

The final R2 for the model was 0.52. The model contained four variables representing the entire mine soil profile (BS, CF, AWHC, EC) and one variable associated with the C horizon (TPC); three were physical properties (CF, AWHC, TPC) and two were chemical properties (BS, EC). All soil properties included in the model can be measured using standard procedures.

Scattergrams of SI as a function of these five soil properties are presented in Fig. 4 . For four of the five variables, the natural relationships were best approximated with a natural log or asymptotic function (Fig. 4F). For example, the natural relationship between site productivity and base saturation (Fig. 4A) increases rapidly at low base concentrations, but levels off as base concentrations increase and tree nutrient requirements are met. Site index as a function of AWHC is similar (Fig. 4C). Conversely, site index is high at low levels of coarse fragment content (Fig. 4B) and soluble salt concentrations (Fig. 4C). The transformed soil properties represent the linearized form of the natural log transformations (Fig. 4A, B, C, E), which were used in the linear regression analysis. Site index increased proportionally across the range of the C horizon total porosity (Fig. 4D); therefore, the simple linear relationship was used in regression analysis for this soil property.



View larger version (32K):
[in this window]
[in a new window]
  		Fig. 4. Distribution of site index as a function of five selected mine soil properties across 14 mined study sites. Figure 4F shows the nontransformed, general functions for the soil properties that are non-linear with site index.{dagger} Regression depicts predicted site index based on observed values of each soil variable while others held constant at their average. {ddagger} Available water-holding capacity (AWHC) represents centimeters of water held between field capacity and wilting point for the total profile depth on each site.

 
Standardized Coefficients
Standardized coefficients were determined for the five soil variables in the regression model. The coefficients show the relative influence of each variable on SI (Table 3). One standard deviation increase in the independent variable changes the dependent variable, SI, by the dependent variable's standard deviation times the standardized coefficient of the independent variable. For example, an increase in one standard deviation of profile BS (s.d. = 26%) results in a 1.4-m increase in SI. Profile coarse fragments (s.d. = 27%) decreased SI by 1.44 m, while profile available water (s.d. = 9 cm) increased SI by 1.33 m. C horizon total porosity (s.d. = 7%) increased SI by 0.92 m. Profile EC (s.d. = 0.54 dS m–1) reduced site index by 0.74 m.


View this table:
[in this window]
[in a new window]
  		Table 3. Standardized coefficients and statistics for the independent variables in Eq. [1].

 
Interaction effects of multiple soil properties are shown in Fig. 5 , plots of SI as a function of the master variable AWHC and the other four significant soil properties (BS, CF, TPc, EC). Each plot shows the predicted SI across the range of AWHC for five levels of a second soil property while all other variables were kept at their mean value. Across the range of AWHC (18–57 cm) and BS (13–100%), predicted SI ranged from approximately 17 to 28 m. At 35 cm AWHC, SI varied 5 m across the range of BS. There was a proportional increase in site index with increasing BS. In other words, a 20% increase in BS resulted in approximately a 1-m increase in SI across the range of BS. This proportional relationship also occurred for 10% increases in TPC, which resulted in 1.5-m increases in SI. Across the range of AWHC and TPC (41–67%), SI increased 11 m. At constant AWHC, a 4.5-m increase in site index resulted from an increase in TPC from 30 to 70%.



View larger version (39K):
[in this window]
[in a new window]
  		Fig. 5. Estimated site index over a range of profile available water-holding capacity (AWHC) as a function of mined site base saturation, C horizon total porosity, electrical conductivity, and coarse fragments. {dagger}Available water-holding capacity (AWHC) represents centimeters of water held between field capacity and wilting point for the total profile depth on each site.

 
The combined influence of AWHC and EC resulted in an 8-m change in SI (Fig. 5). At constant AWHC, a fivefold decrease in EC resulted in a 3-m increase in SI. The relationship between AWHC and EC was not linear; a decrease in EC from 200 to 160 µS cm–1 resulted in a smaller increase in SI (0.5 m) than a decrease from 80 to 40 µS cm–1 (1.5 m). This was also the case with CF. At constant AWHC, a decrease in CF from 30 to 10% resulted in a greater increase in site index than a decrease from 90 to 70%. Over the range of AWHC, a decrease in CF from 90 to 10% resulted in a 9.5-m increase in SI. For both EC and CF, changes at higher concentrations resulted in small changes in site index. This suggests that within the range of AWHC, higher levels of EC and CF are above an acceptable threshold for forest site productivity.

Mine Soil Properties Influencing Tree Growth
Base Saturation
Base saturation was an important mine soil chemical property correlated with forest site index. Across all mined sites, site productivity increased as BS increased. Base saturation ranged from 13 to 100% and in most cases was higher than that of the nonmined sites (Table 4). The distribution of base saturation was skewed toward high levels (Fig. 4), between 80 and 100%, with the highest levels found, on average, on sites in the Midwest (i.e., IL-1, IL-2, IN-1, IN-2, and KY-1) (Table 4). Base saturation of these midwestern mine soils were the least variable. These sites contained significant amounts of Ca and Mg in relation to other cations such as Al and H. High base saturation levels (>50%) represent adequate base cation availability and a low amount of exchangeable acidity. Unlike typical forest soils of the region, mine soils are commonly composed of freshly weathering material containing higher amounts of permanent charge than the pH-dependent charge associated with organic matter and highly weathered minerals. After several decades, mine soils with BS levels approaching 100% indicate that buffering capacity is still strong, likely due to large amounts of Ca and Mg weathering directly from carbonates. Czapowskyj (1978) found that Ca present in some Pennsylvania mine soils accounted for approximately 80% of BS. Soil reaction ranged from pH 3.2 to 7.9 and was correlated with BS. Native oaks are commonly found across a pH range of 4 to 7, but tend to do best on moderately acid to neutral soils with low salt levels.


View this table:
[in this window]
[in a new window]
  		Table 4. Values of selected soil properties for 22 soils in the midwestern and Appalachian coalfields.

 
Mined site BS and cation nutrition are highly dependent on the parent material. In northern Pennsylvania, Czapowskyj (1978) identified a BS range of 45 to 85% depending on the parent material. This spread in BS contributed to a wide range in poplar (Populus sp.) height growth (3–14 m). The BS of our other mined sites varied widely among plots, with standard deviations as high as 38% (Table 4). WV-1, the site with the lowest forest productivity, had three plots with BS < 30%. Site PA-1 had an average SI estimate of 23 m and an average BS of 51%, but with plots as low as 30%. Other research on Pennsylvania mine soils reported BS as low as 10% (Pedersen et al., 1978). Cummins et al. (1965) reported that >70% of eastern Kentucky spoils sampled were below 60% BS, which would adversely affect agronomic plant growth on these sites.

Coarse Fragments
The second most important variable in the regression model was total profile coarse fragments. Coarse fragments negatively impacted SI, which decreased as coarse fragments increased. Average coarse fragments on mined sites ranged from 14 to 83% (Table 4). A mid-range gap in data shows that the relationship with SI may be unduly influenced by several sites with low values (Fig. 4B). Ten of the 14 mined sites had coarse fragments >50%. The coarse fragment content of all nonmined sites was <50%. Mined sites in this study consisted of cast overburden or contour bench sites high in rock content throughout the profile. Excessive amounts of coarse fragments limit the fine earth volume available for root proliferation, water-holding capacity, and long-term nutrient availability (Torbert et al., 1988; Childs and Flint, 1990; Thurman and Sencindiver, 1986; Lyford, 1964). Other researchers have noted similar ranges in rock content on mined sites and their impact on plant growth (Andrews et al., 1998; Torbert et al., 1988; Pedersen et al., 1978). The amount of rock present on mined sites, even after a period of 20 to 55 yr, depends on rock hardness, blasting techniques, and spoil handling (Daniels and Zipper, 1997). Several researchers reported a reduction of coarse fragments with time in surface horizons where weathering processes are rapid (Daniels and Zipper, 1997; Johnson and Skousen, 1995; Haering et al., 1993). Surface horizons within this study commonly contained lower coarse fragment percentages; however, the high C horizon rock content of some of our oldest sites indicates that weathering processes are only beginning to influence the mine subsoils.

Available Water Holding Capacity
Water availability is the most important growth-promoting factor for many native forest types (Pritchett and Fisher, 1987) and is of significant importance on reclaimed mined sites (McFee et al., 1981; Czapowskyj, 1978). Profile AWHC was the third most important soil variable influencing mined site productivity (Table 3). As AWHC increased, SI increased. Available water-holding capacity, defined in this study as the amount of water available between field capacity and wilting point for the whole profile depth (cm), ranged from 18 to 57 cm across all sites (Table 4).

Mined sites can have poor water retention resulting from high coarse fragment content, lack of fine earth, and poor soil structure, which allow water to drain quickly from the soil profile (Thurman and Sencindiver, 1986, Pedersen et al., 1978). However, AWHC was higher on all mined sites compared with adjacent nonmined sites. On average, the mined sites were 48% higher. Thurman and Sencindiver (1986) also observed similar subsurface water retention on several mined sites compared with local native soils in the Appalachian region. Similarity in AWHC levels was due to deeper mine soils compared with native soils. In our case, simple relationships did exist between available water and total depth, coarse fragments, and C horizon silt and clay percentages. As total depth and C horizon silt and clay percentages increased, profile available water increased. Conversely, as coarse fragments increased, total available water decreased.

Total Porosity
Profile total porosity was the fourth most influential variable on mine soil quality. Higher total porosity resulted in higher site productivity. Total porosity of nonmined forest soils generally range from 30 to 65% (Pritchett and Fisher, 1987). Total porosity ranged from 44 to 67% across all mined sites, with most soils falling in the range of 50 to 60% (Table 4). Total porosity of mine soils reported in other studies ranged from 27 to 83% (Andrews et al., 1998; Johnson and Skousen, 1995; Torbert et al., 1988; Indorante et al., 1981). Total porosity levels were similar to those found in adjacent nonmined soils (Table 4).

Noncapillary porosity influences a soil's ability to drain and exchange gases (Brady and Weil, 1999). Capillary porosity enhances the ability of soils to retain water under levels of increasing moisture stress. Noncapillary porosity ranged from 13 to 42%, indicating that conditions for gas exchange were more than adequate (Pritchett and Fisher, 1987). Ungraded cast overburden suffers little compaction from traffic that commonly occurs on postlaw mine soils (Sencindiver and Ammons, 2000). Noncapillary porosity on contour-mined sites (WV-1, WV-2, VA-1) was well above 10%, indicating that traffic from mining activity did not excessively compact the mine soils. However, the combination of high coarse fragments, high noncapillary porosity, and excess voids could make these well-drained mine soils droughty during dry periods of the year (Thurman and Sencindiver, 1986). Creating deep mine soils may be one way to offset the potential droughtiness of highly porous mine soils (Wade et al., 1985; Sencindiver and Smith, 1978).

Soluble Salts
Profile soluble salts, estimated by EC, were the least significant variable in the final model (Table 3). However, EC is a common mine soil variable influencing plant productivity (Andrews et al., 1998; Torbert et al., 1988; Davidson, 1986, McFee et al., 1981). High levels of soluble salts inhibit water and carbon dioxide uptake, and also inactivate enzymes affecting protein synthesis, C metabolism, and photophosphorylation (Taiz and Zeiger, 1991). Our regression analysis indicated a decrease in site productivity with an increase in the soluble salt concentration. Andrews et al. (1998) and Torbert et al. (1988) reported a similar trend with white pine planted on mined sites.

Torbert et al. (1988) found a significant relationship between EC and finely textured soils derived from shales and siltstones, with EC increasing with finely textured shales. Electrical conductivity on their study sites in Virginia ranged from 300 to 1700 µS cm–1. To minimize adverse effects of EC, they recommended placing coarse-textured, oxidized sandstone on the surface instead of finely textured, reduced overburden. Our field descriptions also showed that the finely textured C horizons had the highest EC. Textures of plots with the five highest EC readings had textures of silty clay and silty clay loam, while the plots with the five lowest EC values had textures of sandy loam and loam.

McFee et al. (1981) listed EC as one of the most influential soil properties on Indiana mine soils. Electrical conductivity levels were high enough in some black and gray shales and sandstone to retard plant growth. Soluble salt concentrations >1000 to 3000 µS cm–1 were found to be detrimental to plant growth, reducing tree survival and crop yields (McFee et al., 1981; Cummins et al., 1965). Electrical conductivity on our mined sites ranged from 37 to 159 µS cm–1 (Table 4), falling well below established critical limits defined for agronomic purposes. Davidson (1986) found specific conductance important to growth and survival of three pine species and two hardwood species. These collective studies suggest that forest trees may be more sensitive to salty soils than most agronomic crops. Furthermore, the influence of total salts may be manifested through symbiotic relationships with soil biota, namely mycorrhizal fungi. However, cause and effect relationships have not been determined or studied.

Other studies using regression techniques to link measures of site productivity to mine soil and site properties included the five variables in our final model; however, other soil variables such as total depth, soil N, organic C, and soil P were also found to be important for normal tree growth. In our study, some of these properties had colinear relationships with the properties in our regression model. For example, total depth was colinearly related to coarse fragment content. Total depth on reclaimed mined sites is important to tree and plant growth (Andrews et al., 1998; Wade et al., 1985; Pederson et al., 1978; Sencindiver and Smith, 1978). Mine soils deeper than native soils provide trees with more exploitable volume for nutrients, water, and physical stability (Plass, 1982; Sencindiver and Smith, 1978). Our maximum sampling depth was 1.5 m, but even on sites with depths >1.5 m, depth was usually limited by excessively large coarse fragments rather than bedrock or compacted layers. Soils were loose enough to allow tree roots to explore avenues around large coarse fragments.

Other soil properties such as soil N, organic C, and P have been reported as growth limiting on mined sites, but usually within the first 10 yr after disturbance (Andrews et al., 1998; Torbert et al., 1988; Czapowskyj, 1978; Woodmansee et al., 1978; Ashby and Baker, 1968). Organic matter and total N are good indicators of N availability in mine soils (Bendfeldt et al., 2001; Woodmansee et al., 1978). Nitrogen is commonly found limiting on mined sites soon after reclamation when little organic matter is present and N fixing organisms have yet to become established. Soil profile total N levels in this study ranged from 1208 to 5868 kg ha–1. The lower end of this range is equivalent to levels found in agricultural soils of the Piedmont region of the Southeast. The higher end of the range is equivalent to levels found in undisturbed forest soils throughout the USA. (Pritchett and Fisher, 1987). Total N was colinearly related to organic C, so it was not included in the regression analysis.

Intimately linked with soil N content, plant-derived organic C is scarce on recently mined sites and re-accumulates as the forest community develops (Bendfeldt et al., 2001). This process will occur across all mined sites varying by site age, vegetation type, site productivity, soil type, and climate. In this study, organic C was not significantly related to SI as shown by the regression analyses. Over a period of 20 to 60 yr, the range of forest ages included in our study, organic matter accumulated to levels commonly observed in native forest soils. Organic C levels ranged from 1.9 to 8%, encompassing levels of soil organic C reported for soils from the southeastern region of the USA (3–6%) (Brady and Weil, 1999). However, tests for organic C on mined sites can include portions of the geogenic C pool left after mining, which inflates the estimates (Thurman and Sencindiver, 1986; Indorante et al., 1981; Pedersen et al., 1978; Cummins et al., 1965). In any case, organic C had no significant influence on the variation in SI across these mined sites.

Phosphorus, usually present in adequate supply immediately after mining, has been found to decrease as soil weathering takes place (Howard et al., 1988). Andrews et al. (1998) and Torbert et al. (1988) reported P deficiencies in young white pine growing on mine soils, but extractable P was not significantly related to SI in this study. Soil P ranged from 2 to 89 kg ha–1 on our mined sites. A commonly accepted critical P level for agricultural crops using the sodium bicarbonate extractant is 20 kg ha–1 (Olsen and Sommers, 1982), which indicated that 7 out of 13 mine soils may be deficient in P by this standard. The presence of mature forests on these mined sites suggests that P was adequate early on, giving the trees time to acquire and cycle much of their P internally. Other properties such as water supply may have been more limiting, masking any P deficiency.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The SMCRA requires that reclaimed mined sites be capable of the same productive land uses that occurred before mining. Our study demonstrated that some mined sites could be as productive or more productive as their nonmined counterparts. Trees planted on some reclaimed mined sites in the midwest before the passage of SMCRA grew as productively as before mining. Mining in the east degraded the quality of some sites. Sites with degraded productivity had high coarse fragment contents and shallow depths, preventing trees from exploiting the soil volume to meet their nutrient and water requirements. Improper spoil selection also created growth-limiting chemical conditions such as low BS and high salt content.

Tree growth over decades is influenced by myriad biotic, abiotic, and management factors; but mine soil properties had a dominant effect. Soil characteristics that had the greatest effect on tree growth included, profile base saturation, profile coarse fragments, profile available water, C horizon total porosity, and profile EC; these properties explained 52% of the variation in forest SI. Productive mine sites were commonly well-drained, ungraded mixtures of weathered coarse and fine textured materials. Base saturation was commonly >70%, and coarse fragments averaged 59%. Profile available water averaged 30 cm, and C horizon total porosity averaged 55%. Profile EC averaged 87 µS cm–1.

The soil properties identified by this study represent soil attributes fundamentally important to trees for good growth: ample rooting media, proper aeration, and adequate moisture and nutrient supply. These soil properties are variable within a reclaimed mine soil, and individual tree species requirements are specific. Construction of reclaimed mined sites should take into account not only the mechanical processes that are required to ensure successful reclamation, but also the physical and chemical conditions that result. Such considerations are crucial to the interaction between the developing soil and the planted trees.

The results from this study also reinforce concerns about reclaimed mined land conditions created by new regulations based on the SMCRA. Improper selection of spoil material, lack of original soil, biota, and seed pools, overgrading and compaction, and overly competitive ground covers will all adversely influence the excellent forest productivity shown to exist on mine soils created before the enactment of SMCRA.


   ACKNOWLEDGMENTS
 
We thank the Powell River Project, Office of Surface Mining, Virginia Division of Mines, Minerals and Energy, and the U.S. Dep. of Energy (Instrument No: DE-FG26-02NT41619) for financial support of this study. Special thanks to those people involved in the location of study sites: C. Ashby, A. Boyer, F. Brenner, D. Burger, B. Gray, R. Gullic, T. Probert, J. Skousen, J. Vimmerstdedt, and D. Williamson. Thanks also to the reviewers of the manuscript and our helpful colleagues at Virginia Tech.

Received for publication December 11, 2001.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via ISI Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Rodrigue, J. A.
Right arrow Articles by Burger, J. A.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Rodrigue, J. A.
Right arrow Articles by Burger, J. A.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Rodrigue, J. A.
Right arrow Articles by Burger, J. A.
Related Collections
Right arrow Soil Fertility and Productivity
Right arrow Forest Soils
Right arrow Ecosystem Restoration

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Vadose Zone Journal Journal of Plant Registrations
Journal of Natural Resources
and Life Sciences Education		 Journal of
Environmental Quality