ARTICLE pubs.acs.org/est
Role of Soil Health in Maintaining Environmental Sustainability of Surface Coal Mining Peter M. Acton,† James F. Fox,*,† J. Elliott Campbell,‡ Alice L. Jones,§ Harold Rowe,|| Darren Martin,† and Sebastian Bryson† †
Civil Engineering Department, University of Kentucky, Lexington, Kentucky 40506, United States Engineering Department, University of California Merced, Merced, California 95343, United States § Geography Department, Eastern Kentucky University, Richmond, Kentucky 40475, United States University of Texas at Arlington, Arlington, Texas 76019, United States
)
‡
bS Supporting Information ABSTRACT: Mountaintop coal mining (MCM) in the Southern Appalachian forest region greatly impacts both soil and aquatic ecosystems. Policy and practice currently in place emphasize water quality and soil stability but do not consider upland soil health. Here we report soil organic carbon (SOC) measurements and other soil quality indicators for reclaimed soils in the Southern Appalachian forest region to quantify the health of the soil ecosystem. The SOC sequestration rate of the MCM soils was 1.3 MgC ha1 yr1 and stocks ranged from 1.3 ( 0.9 to 20.9 ( 5.9 Mg ha1 and contained only 11% of the SOC of surrounding forest soils. Comparable reclaimed mining soils reported in the literature that are supportive of soil ecosystem health had SOC stocks 2.55 times greater than the MCM soils and sequestration rates were also 1.63 times greater. The high compaction associated with reclamation in this region greatly reduces both the vegetative rooting depth and infiltration of the soil and increases surface runoff, thus bypassing the ability of soil to naturally filter groundwater. In the context of environmental sustainability of MCM, it is proposed that the entire watershed ecosystem be assessed and that a revision of current policy be conducted to reflect the health of both water and soil.
1. INTRODUCTION Surface coal mining in the Southern Appalachian Forest Region (SAFR) has received recent scientific and public attention due to potentially unsustainable policies associated with the energy production practice and its environmental consequences. The SAFR—located in southern West Virginia, eastern Kentucky, southwestern Virginia, and portions of eastern Tennessee, United States—is a mixed, mesophytic, hardwood forest region responsible for 23% of coal production in the United States and is noted for abundant reserves of high-volatile, low-sulfur, bituminous coal. An estimated 24 billion metric tons of high quality coal remain in the region making continued mining an attractive source of energy.1 However, environmental consequences associated with mountaintop coal mining (MCM) practices have been highlighted in international research circles.24 MCM methods make use of dynamite blasting and large scrapers, excavators, and dump trucks to remove as much as 100 m of geologic overburden material to reach coal seams through methods including steep-slope mining, mountain top removal, contour mining, and area mining. Excess overburden, comprised of sandstone, limestone, and shale fragments, is either replaced or stored in neighboring valleys termed valley fills. The net result is over 300 000 ha of temperate forest in the SAFR that has been r 2011 American Chemical Society
drastically disturbed.4 Palmer et al.4 review the effects of the MCM practices on stream ecosystem health and call for policy change. Surface mining that was performed in the region during the first three-quarters of the 20th century was typically in easily accessible seams, and reclamation that followed tended to favor end dump (uncompacted) fill on steep slopes. The result of this practice was that surface mining in the region was associated with landslides and severe surface erosion on the reclaimed sites. These problems prompted passing of the Surface Mining Control and Reclamation Act (SMCRA) in 1977 which required that fill be compacted and steep slopes be stabilized. Following from SMCRA, reclamation of MCM sites has emphasized compaction of overburden to reduce soil loss and establishing a sustainable mixture of grass seed and legumes to reduce surface erosion. Topsoil is typically not added to the reclaimed sites because of difficulties in on-site storage of the material during the
Received: August 8, 2011 Accepted: October 26, 2011 Revised: October 25, 2011 Published: October 26, 2011 10265
dx.doi.org/10.1021/es202764q | Environ. Sci. Technol. 2011, 45, 10265–10272
Environmental Science & Technology mining process, and its high erodibility when applied over a compacted subsurface. While SMCRA might have been successful in terms of avoiding slope instability and reducing surface erosion, the legislation appears to have had little forethought in terms of soil ecosystem health. Soil health is defined as the ability of the soil to perform ecological soil functions and in general obtain sustainable agronomic, pastoral, and silviculture capabilities.58 To this end, soil health on reclaimed MCM sites is the ability of the soil to accumulate and protect ecosystem soil carbon, promote root action, and transfer liquids and gases, as well as provide sustainable mixed hardwood, deciduous forest capabilities, which are indicative of the Appalachian Region. Measuring the ability of the soil to perform ecological soil functions can be difficult and expensive over long time scales, and thus indicators of soil quality are typically used to provide a measure of soil health.5 In the case of reclaimed surface mining soils, Shukla et al.6 performed a statistical cluster analysis of twenty soil properties for reclaimed soils and showed that soil organic carbon (SOC) stock, soil bulk density, and aggregate stability are highly indicative of soil quality. SOC stock and the build-up of SOC over time provide a justifiable, long-term, integrated measure of soil quality;8,9 soil bulk density is indicative of the soil’s ability to support root action and transfer liquids and gases;10,11 and aggregate stability reflects the ability of the soil to protect SOC from natural losses through the creation of barriers between microbes and enzymes; and stable aggregates create a pool to allow for the accumulation of ecosystem carbon.5,9,12 To meet slope stability regulations emphasized in SMCRA, typical MCM reclamation practices in the SAFR include heavy compaction of overburden at the land surface followed by grass and legume postmining land cover. By adhering to the SMCRA policy, the soil health of the reclaimed mine soils has been neglected. The resulting reclaimed grassland soils are typically described as low quality with respect to natural fertility, organic matter content, and their hydrologic function,4,13 particularly in comparison with the mixed, mesophytic, hardwood forests of the region that support some of the most ecologically rich ecosystems in North America.14,15 However, the soil health of reclaimed MCM soils in the SAFR has not been well documented in the scientific literature. Extensive environmental research was performed by the U.S. EPA in response to litigation related to Clean Water Act regulations that applied to mountaintop mining.1 Although some consideration of microbial aspects of soil carbon and nitrogen was reported,1 the majority of the focus was on the impacts of MCM on water in order to address the litigated issues. Previous literature in the eastern United States has placed emphasis upon tree stand recovery on surface mining soils and also presented soil data that suggests that reclaimed surface mining soils show little soil development and are depleted with respect to their carbon storage capacity.1619 However, literature that describes the soil health of MCM reclaimed grassland soils in the SAFR and documents indicators of soil quality has not been presented. Maintaining a high level of soil quality in postmining disturbed systems should be regarded as a high priority in order to promote a well-functioning terrestrial ecosystem, support agricultural and silvicultural land-uses, and store and sequester soil carbon. Further, it is well understood that downstream water quality is partially a function of upland soil health, and specific hydrologic and ecological functions may be solely attributed to soil conditions.20 Therefore, proper reclamation of upland soils
ARTICLE
may prove to be an economical alternative to mitigate lingering water quality problems. Thus, it is critical that the health of the reclaimed soil ecosystem is incorporated into environmental policies with respect to MCM. The objective of this study was to quantify the impact of MCM and current reclamation practices in the SAFR on the soil health of the postmining soils by measuring SOC stocks and other related soil quality indicators. Thereafter, a critical examination of the current policy is presented, with emphasis on the apparent lack of consideration for both soil health and the interplay among soil, water, and the subsequent watershed quality. A chronosequence consisting of six separate grassland reclaimed soils spanning 14 years of age were sampled and used to quantify increases in SOC stocks with age, as well as several other key soil health indicators. The generated results were then examined temporally and modeled alongside comparison studies reported in the literature to assess the SOC uptake rate of SAFR mine soils during the first 20 years and after of soil development.
2. MATERIALS AND METHODS Complete methods description can be found in the Supporting Information. 2.1. Soil Health Indicators and SOC Stocks. Soil sampling was performed on six ridge-top grassland reclaimed MCM sites in Southeastern Kentucky. Sampled sites were 0, 2, 6, 8, 10, and 14 years since reclamation. The study sites were undisturbed, second-growth hardwood, deciduous forests prior to mining. All reclaimed sites were grasslands with intermittent shrubs and, upon reclamation, were compacted heavily and seeded with a grass/legume mixture to establish a self-maintaining grass system. At each site, multiple soil pits were excavated and soil field observations were performed (see Table S1 for site-specific parameters). Thereafter, litter and soil samples from 0, 510, 1025, and 2550 cm depth intervals were collected across the pit face. For comparison with the unmined forests of the region, 24 soil pits were also excavated in old- and secondgrowth forests in close proximity to the reclaimed sites. Bulk density measurements, as well as samples for water-stable aggregate and pH analysis, were collected from depth intervals of 010 and 1050 cm. Excavated spoil was also collected and analyzed in the laboratory to determine the volumetric ratio of rock and soil. In the laboratory, soil samples for isotopic and elemental analysis were oven-dried at 55 °C until a constant mass was reached. The samples were lightly pulverized in a mortar and pestle and then dry sieved using a 2-mm sieve. Manual examination was employed to remove all root and plant material within the sample. The processed samples were ground to a fine powder. Powdered samples were weighed into silver capsules that were subsequently acidified repeatedly with 6% sulfurous acid (H2SO3) in order to remove carbonate phases. Samples were analyzed using a Costech 4010 elemental analyzer interfaced with a Thermo FinniganConflo III device to a Thermo Finnigan Delta Plus XP isotope ratio mass spectrometer (IRMS). δ13C is expressed in delta notation as 0 1 13 12 ð C= CÞ sample δ13 C ¼ @ 13 12 1A1000 ð1Þ ð C= CÞstandard All isotopic results are reported in per mil (%) relative to Vienna Pee Dee Belemnite for δ13Corg. Average standard 10266
dx.doi.org/10.1021/es202764q |Environ. Sci. Technol. 2011, 45, 10265–10272
Environmental Science & Technology
ARTICLE
deviation for the isotopic standards was 0.04% for δ13C. Average standard deviation for the elemental standard (acetanilide) was 0.82% for %C. Average standard deviation for %Corg and δ13Corg of unknowns was 0.07% and 0.04%, respectively. One obstacle facing SOC measurements in reclaimed mining soils is the presence of geogenic organic carbon (GOC), i.e., carbon from crushed coal fragments within the mining soils.17,2125 The carbon separation technique was used to estimate the contribution of SOC and GOC end-members to the soils.2225 A simple end-member mass balance mixing model was applied as follows: δ13 CTOC ¼ PSOC δ13 CSOC þ PGOC δ13 CGOC
ð3Þ
where δ CTOC is the isotopic value of the total organic carbon in the samples and δ13CSOC and δ13CGOC are the mean carbon isotopic delta values of the soil organic carbon and geogenic organic carbon, respectively. PSOC and PGOC are the fraction of organic carbon derived from each source. Visible coal fragments in the soil samples pooled for each site were analyzed to establish GOC end-members, and litter was used to approximate the SOC end-member for each site. It is known that the isotopic signature of organic carbon can increase with decomposition from litter to SOC and subsequently deeper in the soil column.26 The SOC was less than 14 years in age so it was considered that enrichment would be small; however, since this value is unknown for our sites, two uncertainty bounds were calculated including zero enrichment with depth in the soil column and a one per mil isotopic enrichment. SOC stocks were estimated by using the carbon distributions with depth and estimates for soil bulk density with correction for rock content at each site. Monte Carlo analysis was performed in order to account for the variability of the results. To perform the analysis, a random number generator was used within Microsoft Excel 2007 and samples were drawn from assumed normal distributions for δ13C of the end-member sources and transported soil particulate organic matter. For each spatial location, 1000 realizations were performed, and each realization was solved independently. Carbon density was spatially calculated using the following equation: 13
102 Sf z Cd ¼
n
∑ SOCi Fbi Δzi i¼0 n
∑ Δzi i¼0
SOCðtÞ ¼ SOCe þ ðSOCi SOCe Þ ekt
ð2Þ
and PSOC þ PGOC ¼ 1
dominant species blend and remained in a grassland state after 14 years. Given the similarity of the study sites, the use of a chronosequence was justified to examine the temporal dimension of SOC uptake and development. Undisturbed forest soils were collected to quantify nearby SOC stocks for comparison. An analysis of soil chronosequences was performed to assess the temporal changes in SOC content with emphasis on the time to reach equilibrium SOC content. A first-order kinetic relationship is commonly applied to this analysis,27 e.g., application to SOC development on reclaimed surface mining sites in southeastern Ohio.21 The equation applied is as follows:
ð4Þ
where Cd is the soil organic carbon density (MgC ha1), S is the volumetric fraction of soil material, z is the depth of soil measurements (m), SOCi is the measured soil organic carbon content (gC/100g soil), Fb,i is the dry bulk density (Mg m3), and Δzi is the depth increment (m). The measured bulk density included a significant amount of rock fragments, and the volumetric fraction term corrects the calculation to account for only soil particles. 2.2. Temporal Assessment and Modeling. The climate of all MCM study sites was assumed to be constant given that all sampled sites were located within 40 km and were within 60 m of vertical elevation. Parent material of MCM soils consisted of a mixture of shale and gray and brown sandstone, and was also spatially constant. MCM soils were also seeded with a similar C3
ð5Þ
where SOC(t) is the soil carbon density in time, SOCe is the equilibrium carbon density, and SOCi is the initial carbon density. Although this relationship does not physically account for soil and plant biogeochemistry, it does allow for the comparison of SOC sequestration timeframes between data sets. Soil chronosequences were chosen from Akala and Lal21 and Chaterjee et al.28 and also modeled using eq 5 in order to provide comparison with the MCM sites from the SAFR. Akala and Lal21 and Chaterjee et al.28 were chosen because the soil reclamation practices reported in these papers were highlighted as being supportive of soil ecosystem health. Specifically, the reclamation methods promoted the ability of the soil to accumulate and protect ecosystem soil carbon which is reflected in SOC sequestration rates as high as 4 Mg ha1 y1; promoted root action and soil development which is reflected by soil bulk density decreases as high as 25% over a 25 year period; and increased soil fertility as indicated by increasing nitrogen contents of reclaimed soils. The soil reclamation methods also show the ability to support both sustainable grassland pasture and forest postmining land uses. The sites reported in the Akala and Lal21 and Chaterjee et al.28 were fully mined similarly to the SAFR sites studied here and the forests were cleared, topsoil scraped, overburden removed, and coal extracted. Unlike the reclamation methods typically used in the SAFR, the Southeastern Ohio sites reported in Akala and Lal21 and Chaterjee et al.28 were graded with loose, noncompacted overburden and stored topsoil was spread over the graded surface. When applying the first-order kinetic model (eq 5), the same numerical values in the equation from Akala and Lal21 were applied for comparison with this study. Chaterjee et al.28 was included in the analysis and, similar to the current study, the equation needed to be numerically solved. The constant, k, was solved by differentiating both sides with respect to time and substituting a calculated average initial uptake rate into the variable dSOC(t)/dt. The carbon density obtained at equilibrium is representative of a value that is actually achievable within the soils. In the case of the Chaterjee et al.28 study this value was clearly evident in a temporal plot of the SOC data. In the current study the model was applied twice to the data set while changing the equilibrium carbon density. Low and high values of 50 and 90 MgC ha1, respectively, were used to generate lower and upper bounds. The higher bound was generated with regard to the undisturbed soil carbon density representative of the SAFR,29 while the lower bound represents a lower bound relative to SOC sequestration potential from previous studies. Due to the asymptotic nature of the kinematic function, a consistent method for determining the time frame of equilibrium needed to be established. For this study, a soil was considered stable when 95% of SOC equilibrium was reached. 10267
dx.doi.org/10.1021/es202764q |Environ. Sci. Technol. 2011, 45, 10265–10272
Environmental Science & Technology
ARTICLE
Figure 1. SOC profiles for SAFR soils. Grassland reclaimed and nearby, undisturbed forest soils are represented. New SOC is generated over time in the MCM soils; however, it is restricted to the top 10 cm of the soil column. Undisturbed soils (old-growth) exhibit a gradual decrease of SOC with depth. Uncertainty includes (1 SE.
3. RESULTS 3.1. Soil Health Indicators and Carbon Stocks. See Table S1 in the Supporting Information for results of all soil measurements and end member analysis results. Figure 1 provides SOC variation with depth in the soil profile for the 14-year chronosequence. For the soil surface, i.e., 05 cm, an increase in SOC content with site age occurs, with the 14 year site being comparable to the undisturbed forest soils. However, SOC development is much less pronounced with depth in the soil column, showing almost no development below 10 cm. Some SOC content increase is shown for the 510 cm samples from the grassland sites, particularly for the 10 and 14 year sites. Surface bulk density (010 cm) decreased from 1.87 to 1.51 g cm3 in the 2 to 10 year reclaimed soils, while the
1050 cm depth interval consistently produced results above 2.0 g cm3 throughout the chronosequence. For the forest soils, 010 and 1050 cm average bulk density measurements were 1.18 and 1.45 g cm3, respectively. The fraction of soil particles from the 010 cm depth interval passing through the 250-μm sieve decreased from 0.83 to 0.64 between the 2 and 10 year sites, respectively, indicating a strengthening of aggregate bonds with time. The results of the pH analysis provided no measurable temporal trend for either depth interval, and, on average, equaled 7.4, which agrees well with crushed substrate in spoil that is replaced to the soil surface.13 Figure S1 in the Supporting Information document visualizes δ13C profiles for each of the study sites. Table 1 summarizes SOC stock estimates for the grassland reclaimed MCM and forest soils in the SAFR. The unmined 10268
dx.doi.org/10.1021/es202764q |Environ. Sci. Technol. 2011, 45, 10265–10272
Environmental Science & Technology
ARTICLE
forest soil results agree with previous literature for soil carbon stocks in this region29,30 ranging from 80.6 to 90.6 Mg ha1 in the second- and old-growth forests. The reclaimed soils are heavily degraded in terms of their SOC stocks in comparison with the undisturbed forest soils. SOC stocks ranged from 1.3 ( 0.9 to 20.9 ( 5.9 Mg ha1 for the grassland soils and on average were 11% of their undisturbed forest conditions. Depth average bulk density is 36% higher in the reclaimed grassland soils as compared to the forest soils. Aggregates are much less stable for the reclaimed mining surface soils as compared to the forest soils. Table 1. SOC Stocks for the SAFRa Age (yr)
a
SOC (MgC ha1)
0
1.27 ( 0.88
2
2.42 ( 0.77
6
12.66 ( 6.92
8
10.58 ( 3.00
10
10.14 ( 1.58
14 second-growth
20.85 ( 5.92 80.56 ( 14.15
old-growth
90.57 ( 9.55
Estimates for 050 cm for reclaimed grassland sites in the current study. The unmined forest soil results agree with previous literature for soil carbon stocks in this region.29,30 The reclaimed soils are heavily degraded in terms of their SOC stocks in comparison with the undisturbed forest soils and on average were 11% of their undisturbed forest condition. Uncertainty reflects (1 SE and includes a range of zero to one per mil δ13C enrichment when performing end-member analysis.
While the literature has shown less emphasis on soil health on reclaimed mountaintop coal mining sites, the soil data sets that do exist tend to show agreement with the results presented here. Amichev et al.17 shows that while forest productivity can reach premined conditions, SOC density is a small fraction (
