SEARCH

SEARCH BY CITATION

Keywords:

  • Epidemiology;
  • Conductivity;
  • Cause–effect;
  • Weight of evidence;
  • Macroinvertebrate

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. CASE STUDY METHODS
  5. RESULTS
  6. Conclusions
  7. DISCUSSION
  8. SUPPLEMENTAL DATA
  9. Acknowledgements
  10. REFERENCES
  11. Supporting Information

Increased ionic concentrations are associated with the impairment of benthic invertebrate assemblages. However, the causal nature of that relationship must be demonstrated so that it can be used to derive a benchmark for conductivity. The available evidence is organized in terms of six characteristics of causation: co–occurrence, preceding causation, interaction, alteration, sufficiency, and time order. The inferential approach is to weight the lines of evidence using a consistent scoring system, weigh the evidence for each causal characteristic, and then assess the body of evidence. Through this assessment, the authors found that a mixture containing the ions Ca+, Mg+, HCOmath image, and SOmath image, as measured by conductivity, is a common cause of extirpation of aquatic macroinvertebrates in Appalachia where surface coal mining is prevalent. The mixture of ions is implicated as the cause rather than any individual constituent of the mixture. The authors also expect that ionic concentrations sufficient to cause extirpations would occur with a similar salt mixture containing predominately HCOmath image, SOmath image, Ca2+, and Mg2+ in other regions with naturally low conductivity. This case demonstrates the utility of the method for determining whether relationships identified in the field are causal. Environ. Toxicol. Chem. 2013;32:277–287. © 2012 SETAC


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. CASE STUDY METHODS
  5. RESULTS
  6. Conclusions
  7. DISCUSSION
  8. SUPPLEMENTAL DATA
  9. Acknowledgements
  10. REFERENCES
  11. Supporting Information

Benchmark values that are analogous to conventional laboratory-based water quality criteria can be derived from field studies. An example is the benchmark for conductivity in streams contaminated by leachates of crushed rocks 1, 2. However, because observed relationships between exposure and response are not necessarily causal, a method has been developed for assessing their causality 3. The method uses a criterion-guided weight-of-evidence process that includes a set of characteristics of causation and a formal weighting method applied in six steps. In the present study, we apply that method to demonstrate its utility and to determine the causal nature of the exposure–response relationship used to derive the conductivity benchmark. Other potentially causal factors and their potential to confound the relationship are assessed in a companion article 4. In the article evaluating confounding, we found that the model used to determine the conductivity benchmark is not appreciably affected by other co-occurring stressors. Our concern in the present study is whether elevated ionic concentrations are a cause of the extirpation 5 of macroinvertebrates, and not whether an ionic mixture is the cause in a particular stream or the only cause occurring in the streams represented in the data sets. (Note that the concentration of a defined mixture of ions is the cause, but specific conductivity, hereafter referred to as conductivity, is the exposure metric.) By extirpation, we mean the depletion of a population of a genus to the point that it is no longer a viable resource or is unlikely to fulfill its function in the ecosystem 5.

Our specific hypothesis is that increasing the concentration of ions of Ca+, Mg+, HCOmath image, and SOmath image causes the loss of sensitive invertebrate genera, especially mayflies, from streams. This concern originated from studies of streams below valley fills that contained impaired biotic communities where the degree of impairment was most strongly correlated with conductivity 6, 7. Our analysis uses general knowledge from laboratory experiments and also field observations from streams in Appalachia. Although the effects of elevated ionic concentration on freshwater organisms are well established, most studies have focused on Na+ and Cl, which are associated with marine ecosystems or salts from marine deposits 8–11. However, ionic constituents can be quite different when land disturbance increases ionic concentration. In coal mining areas, the leaching of calcareous overburden results in ionic mixtures containing more HCOmath image/COmath image plus SOmath image than Cl 1, 6. The physiological challenges for organisms are thus quite different 9, 11 compared to ionic regulation in Na+ and Cl-rich waters. Furthermore, freshwater streams are naturally very dilute; background conductivities are often less than 100 µS/cm 1, and species have evolved to occupy that niche 9, 10, perhaps at the expense of other compensatory mechanisms.

CASE STUDY METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. CASE STUDY METHODS
  5. RESULTS
  6. Conclusions
  7. DISCUSSION
  8. SUPPLEMENTAL DATA
  9. Acknowledgements
  10. REFERENCES
  11. Supporting Information

Causal assessment methodology

This causal assessment uses epidemiological methods to assess the general causal hypothesis that increased ionic strength has caused extirpation of stream benthic invertebrates. The method uses all relevant and good-quality evidence in a weight-of-evidence process. The evidence is organized by six characteristics of causation: co-occurrence, preceding causation, interaction, alteration, sufficiency, and time order 12. The causal assessment process is described in a companion article 3 and involves six steps that generate and evaluate whether causation is or is not supported by the body of evidence. A key feature of the process is the weighting and weighing of evidence 3, 13. The evidence is weighted using a system of plus (+) for supporting conductivity as a cause, minus (−) for weakening, and zero (0) for no effect. (Both neutral evidence and ambiguous evidence have no effect on the inference.) A single score is applied to register the logical implication of the relevance of good-quality evidence. Especially strong evidence receives an additional score, based on logical properties (e.g., the effect is inconsistent with the mode of action of the agent) or the quantitative strength of the evidence (e.g., high correlation coefficients). An additional score is lodged if there is consistency among independent studies. A convincing body of evidence requires strong evidence for several characteristics of a causal relationship 3, 12.

Assessment endpoints

The entities of concern are benthic macroinvertebrates. The effect is extirpation of genera from streams in their natural range as defined previously 5. Because the endpoint is the extirpation of multiple genera, a single measurement endpoint is sometimes needed to represent those multiple individual responses. Depending on the type of evidence, different biological measurement endpoints are used. In particular, the number of ephemeropteran genera is used in many of the quantitative analyses because many Ephemeroptera appear to be sensitive (Fig. 1). Also, to the extent that replacement does not occur, the total number of genera is a summary of the consequences of extirpation. The assessment is for general causation 3 of the absence of genera, not for any specific time or location.

thumbnail image

Figure 1. The genera in the order Ephemeroptera, as a group, are extirpated at lower conductivity levels than many other taxonomic groups. The plot is a species sensitivity distribution (SSD). Open circles represent the 95th percentile extirpation concentration (XC95) for a genus. The closed circles are XC95 values for the genera of the order Ephemeroptera. The genus at 230 µS/cm is Cinygmula and at 3,923 µS/cm is Caenis. Other orders represented in the lower fifth of the SSD include Plecoptera, Trichoptera, and Diptera. Data source: Watershed Assessment Data Base (WABbase).

Download figure to PowerPoint

Data sets

Several sources of field data were used to develop evidence. Field data sets were obtained for West Virginia and Kentucky, USA 2 in Appalachia including ecoregions 68 (Southwestern Appalachia), 69 (Central Appalachia), and 70 (Western Alleghany Plateau) 14. The Watershed Assessment Database (WABbase), which was obtained from the West Virginia Department of Environmental Protection, is described by Cormier et al. 1 and was used to derive the conductivity benchmark. Additional information sources were used, including (1) toxicity tests from peer–reviewed literature 7; (2) information on the effects of ionic mixtures on freshwater invertebrates from standard texts and physiological reviews 8–11, 15–23; (3) a U.S. Environmental Protection Agency (U.S. EPA) Region 3 data set from Gregory J. Pond, which includes the original data found in Pond et al. 6 and data collected for a Programmatic Environmental Impact Assessment 24; (4) data on the composition of Marcellus shale brine from Amy Bergdale, U.S. EPA Region 3, based on analyses by drilling operators; (5) data from the Kentucky Division of Water database 2; and (6) geographic and related information from the West Virginia Department of Environmental Protection and public sources 1, 2.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. CASE STUDY METHODS
  5. RESULTS
  6. Conclusions
  7. DISCUSSION
  8. SUPPLEMENTAL DATA
  9. Acknowledgements
  10. REFERENCES
  11. Supporting Information

Pieces of evidence for each characteristic of causation are presented, followed by a summary and scores for evidence of each characteristic.

Co–occurrence

Because causation requires that causal agents interact with unaffected entities, they must co–occur in space and time.

Co–occurrence between conductivity and extirpation of genera

All 163 benthic invertebrate genera appearing in the West Virginia species sensitivity distribution (SSD) list are observed at some sites below 100 µS/cm except Hydroporus (lowest occurrence at 168 µS/cm); therefore, low conductivity is not a limiting factor 1, 2. However, 24.5% of genera are never observed at >1,500 µS/cm (Table 1). Among the 10% most susceptible genera are representatives of Trichoptera, Plecoptera, Ephemeroptera, and Diptera, highlighting that many groups are affected.

Table 1. After mining or reclamation, specific conductance is greater, and selected metric values are less even though total rapid bioassessment protocol (RBP) 6 habitat scores are similar to unmined conditionsa
StreamSampling yearSpecific conductance (µS/cm)Total genus richnessEPT genus richnessEphemeropteran genus richnessTotal RBP habitat score
  • a

    Based on Pond et al. 6. Empty cells indicate data not available.

  • b

    RBP from Spring 2000.

  • c

    Identified for this study from the West Virginia Department of Environmental Protection.

  • d

    Single measurement.

  • e

    Mean value.

    EPT = Ephemeroptera, Plecoptera, Trichoptera; MTM-Valley Fill = mountaintop mining overburden; WABbase = Water Analysis Database; GLIMPSS = genus–level index of most probable stream status; WVSCI = West Virginia Stream Condition Index.

Unmineda
 Rushpatch19996042179147
 20067040197144
 Spring19995133178163b
 20066637218149
 White Oak19996432179161
 20068830208163
Reclaimed mineda
 Ballard19991,20133123148
 20061,1952093149
 Stanley Fork19991,3871420145
 20062,0102860155
 Sugartree20001,8542240141
 20071,9102040154
Unminedc
 Ash Fork199844a    
 200339b41206 
 200651a24144 
 200737a27229 
MTM-Valley Fill permit awarded 1994, 1996c
 Boardtree19981,396d    
 20033,015e2052 
 20073,390d    
MTM-Valley Fill permit awarded 1996c
 Stillhouse1998511d    
 20033,199e830 
 20073,970d    

Scoring—This evidence supports the causal relationship (+); extirpation of 40 genera in West Virginia and 46 in Kentucky in streams with conductivity >1,500 µS/cm is a strong effect (+). The two independent data sets and analyses corroborated one another (+). The total score assigned is + + + (Supplemental Data, Table S1).

Co–occurrence of cause and Ephemeroptera

We constructed a contingency table of the presence/absence of any individual of the order Ephemeroptera at sites near background conductivity (≤200 µS/cm) and high conductivities (>1,500 µS/cm; Table 2). It shows that Ephemeroptera co–occur with low conductivity but that all ephemeropteran species are absent from more than 55% of sites where conductivity is high. This analysis emphasizes the difference between high- and low-conductivity sites with respect to a clear effect endpoint, the absence of all individuals of the order Ephemeroptera.

Table 2. Presence of Ephemeroptera contingent on stream conductivity from three data setsa
 Ephemeropteran presentbEphemeropteran absentcTotal
  • a

    Percentage in parentheses is calculated from the ratio of total observations in a dataset within the designated conductivity range. For example, at least one individual of the order Ephemeroptera was observed in 99.2% (852/859) of sites <200 µS/cm in the WABbase data set.

  • b

    Sites with >1 ephemeropteran individuals.

  • c

    Sites with 0 ephemeropteran individuals.

  • d

    Sampling methods 1.

  • e

    Sampling methods 6.

  • f

    Sampling methods 1, 37.

    WVDEP = West Virginia Department of Environmental Protection; WABbase = Water Analysis Database.

WVDEP data set—WABbased
 Near background conductivity (≤200 µS/cm)852 (99.2)7 (0.8)859
 High conductivity (>1,500 µS/cm)50 (45)61 (55)111
 Total90268970
U.S. EPA Region 3 data sete
 Conductivity ≤300 µS/cm7 (100)0 (0)7
 Conductivity >1,500 µS/cm4 (19)17 (81)21
 Total111728
Kentucky data setf
 Conductivity ≤300 µS/cm150 (97.4)4 (2.6)154
 Conductivity >1,500 µS/cm9 (69.2)4 (30.8)13
 Total1598167

We repeated the analysis with the U.S. EPA Region 3 data set and the Kentucky data set, with similar results despite the differences in sampling and identification protocols for the Kentucky data set (Table 2). To ensure a sufficient number of samples in these smaller data sets, the low-conductivity category was <300 µS/cm, and the high-conductivity category was >1,500 µS/cm. In each case, high-conductivity sites were much more likely to lack Ephemeroptera (Table 2).

Scoring—This evidence supports the causal relationship between conductivity and extirpation of genera (+). Where conductivity is high, individuals of the order Ephemeroptera are less likely to occur. A change of 50% or more is large (+). The evidence is corroborated in three independent data sets collected from different streams at different times by different researchers using different sampling protocols (+). The total score assigned is + + + (Supplemental Data, Table S1).

Co–occurrence in nearby catchments

Studies of matched mined and unmined streams provide another way to examine co-occurrence. Pond et al. 6 compared sites in three unmined watersheds with three nearby reclaimed mined watersheds below valleys filled with mountaintop mining overburden (MTM-Valley Fill; Table 3). The conductivity is lower in the unmined sites compared to the reclaimed mined sites, and all of the biological metrics are greater in the unmined sites, even though habitat scores are similar. The number of ephemeropteran genera is two- to threefold greater in the unmined sites. From the WABbase we identified two other valley–filled tributaries, Boardtree and Stillhouse Branch, and one unmined tributary, Ash Fork, in the Twentymile Creek Watershed, West Virginia, USA. The conductivity is lower, and all of the biological metrics are greater in the unmined sites compared to mined sites (Table 3). In a study of 28 watersheds with excellent habitat quality, streams with valley fills had greater conductivity and reduced diversity and fewer sensitive genera 25. These studies show that where conductivity is greater, the biological diversity is less.

Table 3. Number of genera contingent on stream conductivitya
 West VirginiaKentucky
Genera presentGenera absentGenera presentGenera absent
  • a

    When conductivity is low, all genera but one are observed, but at high concentrations 24% are absent when 200 specimens are identified (West Virginia, USA) and 44% when all individuals are identified (Kentucky, USA) in a sample. This shows that many genera are affected by ionic stress, not just mayflies, and that low conductivity is not a limiting factor. Percentage in parentheses is calculated from the ratio of total observations in a data set within the designated conductivity range. For example, at least one individual of a genus was observed in 99.9% (162/163) of sites <150 µS/cm in the Watershed Assessment Database (WABbase). Data from WABbase and Kentucky Division of Water database.

Near background conductivity (<150 µS/cm)16211040
(99.9)(0.01%)(100%)(0%)
High conductivity (≥1,500 µS/cm)123405846
(75.5%)(24.5%)(55.8%)(44.2%)

Scoring—This evidence supports the causal relationship (+); the number of genera is two to three times greater at the low-conductivity sites for most metrics; few or no Ephemeroptera were observed at three-fourths of the sites (+). The results are consistent and independently corroborated (+). Total score assigned is + + + (Supplemental Data, Table S1).

Preceding causation

Each causal relationship is a result of a web of preceding cause and effect relationships that begins with sources and includes pathways of transport, transformation, and exposure. Evidence of sources of a causal agent increases confidence that the causal event actually occurred and was not a result of a measurement error, chance, or hoax 26.

Complete source-to-cause pathway from the literature

Because exposure to dissolved ions does not require transport or transformation (i.e., organisms are directly exposed to ions in water immediately below sources), only evidence of the occurrence of sources of ionic inputs is assessed for this type of evidence. Potential sources for a mixture of ions HCOmath image/COmath image plus SOmath image greater than Cl in the region include surface and underground coal mining, effluent from coal preparation plants and associated slurry impoundments, effluent from coal fly ash impoundments, scrubbers at coal-fired electric plants, and demineralization of crushed rock 7, 23, 24, 27. In particular, high-conductivity leachate has been shown to flow from valley fills created during coal mining operations 6, 18, 24. In contrast, mixtures are more likely to be dominated by Cl when they are associated with winter road maintenance 28, 29, brines from natural gas and coalbed methane operations 30, treatment of wastewater 31, and human and animal waste 23, 32. Ecological studies have shown that conductivity increases only slightly following clear–cutting and burning. Dissolved mineral loading may be increased slightly by timber harvesting but also declines quickly as vegetation reestablishes 33. Golladay et al. 34 and Arthur et al. 35 found increases in nitrogen and phosphorus export in logged catchments in Appalachia but minor differences in calcium, potassium, or sulfate concentrations between logged and undisturbed watersheds. Likens et al. 36 actually found sulfate concentrations to decrease following clear-cutting and experimental suppression of forest growth by herbicides.

Scoring—This evidence from the literature indicates that there are sources of the mixture of dissolved ions that are widespread in the region and can be differentiated from sources of other mixtures (+). Multiple studies are consistent in the description of the ion types associated with different sources (+). Strength is not scored. Total score is + + (Supplemental Data, Table S2).

Co–occurrence of sources and conductivity from the region

Conductivity is shown to increase after the construction of valley fill coal mining operations in two catchments, Boardtree and Stillhouse (Table 3). Conductivity is elevated where surface mining operations occur in a watershed and not in an adjacent unmined watershed (Table 3), and overall concentration of ions is greater in mined watersheds with valley fills than in unmined watersheds 6. Similar results are reported in mined and unmined sites in Kentucky and in Virginia 25, 37. Principal component analysis sorted mined and residential sites from reference sites primarily on the basis of specific conductance and pH 37.

Scoring—This evidence supports the causal relationship (+). The conductivity at mined sites is 10 to 50 times greater than at unmined sites (+). The source of increased conductivity is independently corroborated and consistent (+). Total score is + + + (Supplemental Data, Table S2).

Characteristic composition of identified sources

Correlation and regression analyses suggest that, in ecoregions 69 and 70, conductivities above 500 µS/cm contain high levels of the ions of Ca2+, Mg2+, HCOmath image, and SOmath image 7 (scatterplots in Supplemental Data, Fig. S1ae), which is consistent with surface coal mining and valley fill sources 6, 37. In the WABbase data set, 98% of the sample sites were characterized by anions with (HCOmath image + SOmath image)/Cl ≥1 1, and conductivity is less correlated with Cl than with the other ions (Table 4). In mined and unmined sites, the dominant cations are Ca2+ and Mg2+, and anions are HCOmath image and SOmath image 6, 38 from largely calcareous geology. This excludes sources dominated by NaCl including saline effluents from human and livestock wastes 23, 31, 32, road salt 28, 29, and produced brines from gas extraction (A. Bergdale, personal communication U.S. EPA, Wheeling WV; Supplemental Data, Table S3) 2, 30. The median difference is very large; 99% of anions are HCOmath image and SOmath image in both mined and unmined sites, whereas >99% of the anions are Cl in brines from gas extraction in Marcellus shales (Supplemental Data, Table S3). Therefore, this causal assessment relates primarily to mixtures of ions typical of alkaline coal mine drainage and associated valley fill discharges rather than sources that are not coal-related.

Table 4. Spearman rank correlation of stream parameters in WABbase data seta
 ConductivityAlkalinitySulfateChlorideHardnessMagnesiumCalcium
  • a

    This shows that in this region conductivity is most highly correlated with ions other than chloride.

    b Data set as described in 2 for sites with all seven measurements, n = 1,118,

    WABbase = Water Analysis Database.

Conductivity10.780.890.640.950.930.92
Alkalinity0.7810.60.560.780.70.79
Sulfate0.890.610.410.850.90.8
Chloride0.640.560.4110.50.430.5
Hardness0.950.780.850.510.960.99
Mg0.930.70.90.430.9610.91
Ca0.920.790.80.50.990.911

Scoring—This evidence supports the causal relationship (+) by showing that there are sources of high conductivity with a consistent matrix of ions. Both mined and unmined sites have similar proportions of Ca2+, Mg2+, HCOmath image, and SOmath image but very different concentrations. The difference between the ionic composition of mined watersheds and watersheds with other sources of ions such as brines is very large (+). The evidence from the WABbase data set and two other Appalachian studies consistently supported the ionic makeup associated with land disturbance, especially surface mining (+). The data for mined and unmined watersheds are from a peer–reviewed publication 6, and the brine values are from reports from extraction permittees in West Virginia (A. Bergdale, personal communication U.S. EPA, Wheeling WV). Although the brine analyses are not peer reviewed, the findings are qualitatively similar to other nonpeer–reviewed reports of the makeup of such brines. Total score is + + + (Supplemental Data, Table S2).

Correlation of conductivity with sources

In ecoregion 69 14, conductivity increased with particular sources based on scatter plots of conductivity for proportions of nine land cover classifications 38. Data were analyzed from 190 small (<20–km2) catchments draining to the Coal, Upper Kanawha, Gauley, and New Rivers, West Virginia, USA 38. The two land-use types, percentage area in valleys filled with mountaintop mining overburden (r = 0.66) and the summed percentage area in valley fill, abandoned mine land, and mine land (r = 0.57), are most strongly and positively correlated with conductivity. In contrast, percentage area in forest is negatively correlated with ion concentrations (r = −0.55). Percentage area in urban/residential (r = 0.10) is not well correlated and in this region is confounded somewhat by mining land uses. The ions that are more strongly correlated with percentage area in valley fill are total Ca2+ and Mg2+ (also captured together as hardness [r = 0.70]), HCOmath image measured as alkalinity (r = 0.47), and SOmath image (r = 0.70). Noticeably, Cl is not strongly correlated with any land use variable, apparently due to the low range of Cl concentrations, except at one site 38.

Only mining, especially associated with valley fills, is a substantial source of the ions that are measured as conductivity 38. Disturbances associated with agriculture and human habitation may also contribute, but the densities of agricultural and urban land cover are relatively low, and a clear pattern of increasing conductivity and increasing land use is not evident. Furthermore, despite the bedrock of limestone, dolomite, shale, and calcareous cemented sandstone, natural background conductivity is exceedingly low 1 apparently because the native geology is intact and not crushed.

Although conductivity typically increases with increasing land use 29, at relatively low urban land use, conductivity is highly variable 38. This may be caused by unknown mine drainage, deep mine break–outs, road applications, poor infrastructure conditions (e.g., leaking sewers or combined sewers), or other practices. In contrast, there are clear patterns of increasing conductivity as percentage of area in valley fill increases and decreasing conductivity with increasing forest cover 38.

Scoring—This evidence supports the causal relationship (+). The correlations for percentage area in mountaintop mining with valley fill (r = 0.66), all mining minus valley fill and abandoned mine lands (r = 0.42), and forestry (r = −0.55) 38 are moderately strong based on our a priori scoring criteria 3, 39. The present study has not been independently corroborated, although it is consistent with the findings of Pond et al. 6 and Lindberg et al. 40. The association seems to be specific for extensive geologic disturbances, which in these regions are from mining and valley fills. The total score is + (Supplemental Data, Table S2).

Interaction and physiological mechanisms

Causal agents alter affected entities by interacting with them through a physical mechanism. Evidence that a mechanism of interaction exists for a proposed causal relationship strengthens the argument for that relationship.

Evidence of mechanism of exposure

Aqueous salts are dissolved ions that are readily available for uptake by aquatic organisms as they pass over their respiratory and other permeable surfaces 8, 9, 11, 17, 20, 41–43. Benthic invertebrates that inhabit naturally low-conductivity streams (Table 1) may be subjected to waters that have a greater concentration of ions due to local effluents 2, 7. Therefore, the pollutant is present in a form that is consistent with a well-established mechanism of exposure for aquatic animals.

Scoring—Evidence of a mechanism of exposure is from knowledge that the ions are present in streams 2, 7 and from general knowledge of animal physiology and the anatomy of insects and other aquatic invertebrates 8, 9, 11, 17, 20, 41–43 (+). Because the exposure is by the same mechanism that provides respiration (i.e., maintenance of water flow over permeable membranes), it is strong (+). Many studies support this inference (+). The total score is + + + (Supplemental Data, Table S4).

Biochemical mechanism of effect

Living cells and the organisms they comprise must maintain a relatively narrowly defined internal composition of ions that varies with function and that is different from their environment. Maintaining homeostasis involves osmotic and ionic regulation by cells and tissues. Freshwater organisms, including mayflies, are known to use various physical structures and physiological mechanisms to maintain water content, charge balance, and specific ionic concentrations 8, 9, 11, 17, 20, 21, 41–43. Many freshwater invertebrates, including mayflies, have mitochondrion–rich chloride cells on gills and other surfaces that take up chloride and other ions 11, 17. Exclusion of ions is insufficient to maintain homeostasis, which requires coordinated anion, cation, and proton transport by passive, active, uniport, and cotransport processes 19.

The clearest mechanism of effects is the inhibition of membrane–transport pathways by excessive ambient concentrations of bicarbonate, which interfere with the uptake and balance of necessary chloride and sodium ions. The processes by which bicarbonate interferes with ion exchange by chloride cells on the gills of Ephemeroptera are illustrated in Supplemental Data, Figure S2.

Scoring—This mechanism supports the causal relationship by providing evidence that the bicarbonate ion matrix in the region can create ionic gradients that interfere with proper homeostasis (+). However, direct observations of the ionic regulatory processes or membrane potential measurements are not described in the literature for affected or tolerant species studied in Appalachia. Evidence from the literature about mitochondrion–rich chloride cells in epithelia of insects (particularly in ephemeropterans), amphibians, and fish, logically leads to disruption of ionic regulation in organisms highly dependent on passive ionic regulation by an HCOmath image/Cl antiport anion exchange. Other ion transport systems are also affected by increases in the concentration of the ion mixture, which is measured as increased conductivity in the region of concern. A large body of peer–reviewed physiological studies 8, 9, 11, 17, 20, 21, 41–43 supports this inference (+). The total score is + + (Supplemental Data, Table S4).

Physiological mechanism of effect

In aquatic systems, organisms are capable of coping with different environmental challenges presented by different concentrations of dissolved ions. However, the extent and rate of adaptation to changes of ionic composition and concentration varies depending on the physiological potential of a particular species 9–11. As noted previously, osmotic and ionic cellular mechanisms involve selectively permeable membranes. However, it is the disruption of the ionic gradients throughout a physiological system of specialized tissues and organs with specialized functions that determines whether a genus will occur at a location. Some examples include slight or large differences in ionic composition between cell compartments, cells, or external media that are used to release energy from food; transcribe and translate RNA into proteins; regulate pH and water volume; excrete metabolic waste (ammonia and CO2); enable secretion of enzymes, hormones, and neurotransmitters; guide embryonic development 9, 11; and propagate action potentials in nerves and muscles, thus enabling complex behaviors, and activation of fertilized eggs 9, 15, 22. These physiological functions enable organisms to develop, grow, move, and sense their environment. When the pH or ionic gradients are disrupted, stream invertebrates emigrate or die.

Scoring—This evidence supports the causal relationship (+) by demonstrating that the loss of ionic regulation can affect an animal's physiology leading to severe effects. Studies of the physiology of affected species and tolerant species from Appalachia are not available. The effects of ionic disruption are supported by a large body of peer–reviewed physiological studies, some of which are presented above (+). The total score is + + (Supplemental Data, Table S4).

Alteration

A cause alters or changes a susceptible entity. In this case, the alteration is failure to maintain viable populations of sensitive species. Documenting that a change occurs is evidence of causation, but that evidence is much stronger if a specific effect of a cause is characterized. If the specific effect of a cause has no other causes, it can be diagnostic of that cause. Extirpation has many causes, so evidence of alteration is not diagnostic in this assessment but can provide evidence of specificity.

Change of occurrence of genera

Ephemeroptera and Plecoptera do not occur in mesohaline waters, whereas other insect orders do occasionally occur in brackish water 10 (Fig. 1). In an article focusing on Ephemeroptera 6, a nonmetric multidimensional scaling model strongly associated Cinygmula, Drunella, Ephemerella, Epeorus, and Ameletus with the low-conductivity reference sites and Stenonema, Isonychia, Baetis, and Caenis with the higher conductivity sites. These results are consistent with estimated 95th percentile extirpation concentration (XC95) values for those genera 1, 2. The low-conductivity ephemeropteran group of Pond et al. 6 has XC95 values between 230 and 591 µS/cm, and the high-conductivity ephemeropteran group has XC95 values between 745 and 3,923 µS/cm 1, 2. Another study using data from Kentucky showed similar results 37; however, there is more uncertainty in this study because habitat alteration may have confounded the relationship with conductivity in that data set. Nevertheless, the relative frequency of the sensitive genera identified in the West Virginia study 6 decreased by more than half at mined sites in Kentucky and, except for Baetis, a tolerant genus that was relatively unchanged, the relative frequency of the insensitive genera increased at mined sites with high conductivity. This evidence indicates that some specific genera tend to be consistently less tolerant and other are consistently more tolerant of increased ionic concentrations occurring in the region.

Both the XC95 values and the SSD 1, 2 demonstrate that a characteristic set of genera, including many Ephemeroptera, was extirpated at relatively low conductivities and another characteristic set was resistant. The relative sensitivities are consistent with the findings of Pond et al. 6 and Pond 37, and with our analyses of data from Kentucky (2 Appendix G of that report). Genera that are sensitive to high conductivity are similar in Kentucky and West Virginia 2. Genera that began to decrease in occurrence at levels <500 µS/cm were identified from the fitted lines on generalized additive model plots for West Virginia and for Kentucky 2. In the WABbase data set, 14 genera with XC95 values less than 500 µS/cm also occur in the Kentucky data set. Among these 14 genera, nine (64.3%) have XC95 values less than 500 µs/cm in the Kentucky data set. A total of 88 (85%) of the 104 genera in Kentucky used to develop the SSD was also used in the West Virginia SSD. Of these 104 genera, 54 showed declines below 500 µS/cm in at least one data set (44 declined in both data sets, four only in Kentucky, and six only in West Virginia). Therefore, the West Virginia and Kentucky data sets had 44 of 54 genera (81.5%) in common that showed declines below <500 µS/cm.

Scoring—This evidence supports the causal relationship (+) by demonstrating that conductivity greater than background levels causes a consistent set of sensitive genera to be extirpated. The number of genera with similar XC95 values (less than 10% difference) in Kentucky and West Virginia with XC95 < 500 µS/cm is 71.4% and for those with a similar pattern of decline it is 81.5% (+). Multiple studies and data sets confirmed the evidence (+). The total score is + + + (Supplemental Data, Table S5).

Models of change of genera

Empirical models based on macroinvertebrate assemblage composition were used to identify probable causes of biological impairments in a case study in Clear Fork Watershed in West Virginia 44. Eight weighted averaging regression models were developed and tested using four groups of candidate stressors based on genus–level abundance. The strongest predictive models were for acidic metals (dissolved aluminum) and conductivity, r2 = 0.76 and r2 = 0.54, respectively. These appear to correspond to effects of acidic and calcareous mine drainages, respectively. This study shows that there is a distinct assemblage associated with neutral to alkaline high conductivity that is different from other causes such as those with acid mine sites with toxic levels of dissolved aluminum.

In another approach 44, nonmetric multidimensional scaling and multiple responses were used to examine the separation of dirty reference groups from clean reference groups based on the biological communities observed in the two groups. Four dirty reference groups were identified consisting of sites primarily affected by one of the following stressor categories: dissolved metals (Al and Fe), excessive sedimentation, high nutrients and organic enrichment (using fecal coliform as a surrogate measure of wastewater and livestock runoff), and increased ionic concentration (using sulfate concentration as a surrogate measure). Of the dirty reference groups, the dissolved metals group was significantly different from the other three dirty reference groups (p < 0.001). The other three dirty reference groups, although overlapping in ordination space to some extent, were also significantly different from one another (p < 0.05). Overall, each of the five reference models (the fifth model was clean reference sites) was significantly different from the others (p < 0.001), indicating that differences among stressors, including ionic concentration, apparently led to unique macroinvertebrate assemblages.

In another study with a different data set collected in West Virginia, nonmetric multidimensional scaling was applied to invertebrate genera, and sites were sorted into distinct ordination spaces characterized by low, medium, and high conductivities associated with surface mines with valley fills 6. A study in Kentucky found similar results 37.

Scoring—This evidence supports the causal relationship (+) by demonstrating that conductivity greater than background levels causes a consistent set of sensitive animals to be extirpated. The prediction was statistically strong (+). The effect is specific enough to clearly separate groups by nonparametric statistical methods in two different data sets. Independent data sets and investigators confirmed that different assemblages of invertebrates occur with different stressors, including neutral–to–alkaline waters with increased concentration of ions (+). The total score is + + + (Supplemental Data, Table S5).

Sufficiency

Because many agents are natural components of the environment (e.g., ions), a causal relationship must show that there are thresholds or patterns of the effect to susceptible entities (e.g., mayflies) associated with the changing magnitude of exposure (e.g., conductivity). In this section, we describe evidence that can be credibly used to evaluate whether the level of ionic concentration is sufficient to cause extirpation. The evidence is primarily from field observations. Several laboratory studies were not used to evaluate sufficiency for the following reasons: (1) the ionic constituents were not similar to those in high conductivity waters in the region of concern; (2) the test species are physiologically tolerant of higher concentrations of ions; or (3) only acutely lethal effects were reported 2. Such toxicity tests serve to show that ionic mixtures are highly toxic at some levels to some test species, but they do not provide evidence that the levels observed in the streams of the regions were sufficient to cause the extirpation of genera. Although available test results were not useful for this causal assessment, such tests are potentially useful for other causal assessments.

Laboratory tests of reconstituted mine discharges

Kennedy et al. 16 tested simulated coal mine discharge waters in Ohio with the ephemeropteran Isonychia bicolor. The ionic matrix was dominated by SOmath image, HCOmath image, and Na+. In 7–d lethality tests, the lowest observed effect concentrations for survival of Isonychia (mid–to–late instars) at 20°C occurred at 1,562, 966, and 987 µS/cm in three tests. These values bracket the field-derived XC95 for Isonychia of 1,180 µS/cm 1, 2. However, when the assay was conducted at 12°C, the lowest observed effect concentration was 4,973 µS/cm, suggesting that longer exposures are needed before effects occur at cold temperatures.

Scoring—The laboratory tests by Kennedy et al. 16 establish that the effect for one insensitive ephemeropteran species, Isonychia bicolor, in the laboratory, occurred at a similar conductivity level to that in the field. A total score of + was assigned (Supplemental Data, Table S6).

Field exposure–response relationships of composite metrics

As Hill 45 suggested, a biological gradient in the field suggests that the exposures reach levels that are sufficient to cause effects. Evidence from several studies was evaluated.

Our analyses, using the WABbase data sets, show that as conductivity increases, the total number of genera and the number of ephemeropteran genera decrease at conductivity levels shown to extirpate sensitive genera (r = −0.41 and −0.61, respectively; Fig. 2). This analysis shows not only the co–occurrence of elevated conductivity and the loss of stream biota but also that there is a regular exposure–response relationship that extends to the lowest observed concentrations (evidence of sufficiency).

thumbnail image

Figure 2. As conductivity increases, the number of total genera (A) and ephemeropteran genera (B) decreases. The fitted lines are locally weighted scatterplot smoothing (LOWESS) lines. The LOWESS line fits simple models to subsets of the data, or a span, in this case a change of 0.75 number of taxa or genera. Association with conductivity is evident despite other causes and randomness. Data source: Watershed Assessment Data Base (WABbase).

Download figure to PowerPoint

In studies of the effects of valley fills in West Virginia by Pond et al. 6, ephemeropteran genera and conductivity were highly negatively correlated (r = −0.90) with conductivity and less so with habitat (r = −0.64). Pond 37 and Pond et al. 6 also reported that the number of ephemeropteran genera and the number of total genera decrease as conductivity increases. In a recalculation of the Pond et al. 6 data with additional data to create the U.S. EPA Region 3 data set, the ephemeropteran genera and total genera were both moderately negatively correlated with conductivity (r = −0.72 and −0.35, respectively; Fig. 3).

thumbnail image

Figure 3. As conductivity increases, the number of total genera (A) and number of Ephemeroptera genera (B) decreases. The fitted lines are locally weighted scatterplot smoothing (LOWESS) lines (span = 0.75). Data from EPA Region 3.

Download figure to PowerPoint

In a study in Ohio, Kennedy et al. 46 report that as conductivity increases, the percentage of Ephemeroptera decreases from 23.7% at a mean of 399 µS/cm to 0% at 5,376 µS/cm with intermediate effects between those values. In this study, Na+ and Ca2+ were at similar concentrations at the higher conductivity levels.

In a study of 28 headwater streams in Virginia, Timpano et al. 25 showed that the number of Ephemeroptera, Plecoptera, and Trichoptera (EPT) (r = −0.70) and total (r = −0.64) genera declined as conductivity increased even when stream habitat was similar.

Scoring—The field observations show that as conductivity increases, the number of Ephemeroptera and total number of genera decrease and, thus, the concentration of ions in streams is sufficient to cause effects (+). The correlation is strong to moderately strong depending on the data set. The effect was specific for the ionic mixture. The correlations were corroborated with independent data sets from different streams sampled by different investigators (+). A total score of + + was assigned (Supplemental Data, Table S6).

Field exposure–response relationships of composite indices

The relationship between conductivity and the West Virginia Stream Condition Index (WVSCI) score, which is a composite of six family level metrics, was also modeled from the WABbase data set 47. The WVSCI is scored from zero to 100 with a low score indicative of a poorer quality aquatic biological assemblage and poorer stream condition. Mean WVSCI scores from 60 bins were regressed with conductivity (Fig. 4). A strong downward slope of WVSCI was seen with increasing conductivity.

thumbnail image

Figure 4. Stream condition decreases with increasing conductivity. As conductivity increases, the score decreases for a composite index that characterizes stream biota, the West Virginia Stream Condition Index (WVSCI). Points represent mean WVSCI score for conductivity bins. Bars are 90% confidence intervals. The dotted line is the 95% confidence bound for the modeled line. A WVSCI impairment score of 68 intercepts the regression line at 180 µS/cm (dashed arrow). The model estimates a WVSCI value of 64 at 300 µS/cm (solid arrow). Data source: Watershed Assessment Data Base (WABbase).

Download figure to PowerPoint

In Pond et al. 6, the genus–level index of most probable stream status and WVSCI scores were strongly correlated with conductivity (r = −0.90 and −0.80, respectively). Gerritsen et al. 44 identified 180 µS/cm as a plausible stressor–response threshold and 300 µS/cm as a substantial effects threshold for the association of conductivity and the WVSCI biological index using a data set from the WABbase.

Timpano et al. 25 showed that the Virginia Stream Condition Index score decreased as conductivity increased with ordinary least square regression or quantile regression.

Scoring—This set of evidence indicates that, in multiple data sets and by a variety of biological responses and analytical methods, as conductivity levels observed in the region increase, stream condition decreases, and the assemblage of macroinvertebrates is different from best available reference sites in the region. This is supporting evidence of sufficient ionic concentrations in the streams to cause widespread effects (+). The correlations are strong (+). The correlations were corroborated with different methods in four independent studies (+). A total score of + + + was assigned (Supplemental Data, Table S6).

Field exposure–response relationships: Susceptible genera

As conductivity increases, the occurrence and capture probability decrease for many genera in West Virginia and Kentucky 1, 2 at the conductivity levels predicted to cause effects. The loss of these genera is a severe and clear effect.

In the West Virginia data set at 500 µS/cm, 17% of genera (28/163) are extirpated and an additional 68% of genera are declining as conductivity increases at many sites. In the Kentucky data set, 11.5% of genera (12/104) are extirpated at 500 µS/cm, and a total of 67% of genera are in decline. This evidence shows that exposures are sufficient to extirpate susceptible genera in two geographic areas. The associations show that relatively low exposures are sufficient to adversely affect susceptible genera.

Timpano et al. 25 estimated a 20% decline of genera for streams exceeding 652 µS/cm; however, their estimates were based on a small sample size—60 genera with as few as five observations used to estimate the effect. The taxa most sensitive to ionic strength were reported to be ephemeropterans followed by trichopterans.

Scoring—The observed effects logically support the causal relationship between increased conductivity and declining occurrence of susceptible genera and indicate that effects occur at relatively low conductivity levels (+). The effect is strong, with complete extirpation of many genera (+). The results were corroborated with independent data sets from Kentucky and Virginia (+). The total score is + + + (Supplemental Data, Table S6).

Time order

Logically, a causal event occurs before an effect is observed. Evidence of time order could be provided by changes in the invertebrate assemblages after the introduction of a source that increased conductivity.

We could not obtain conductivity and biological survey data collected before and after construction of a valley fill or release of ion-rich effluents from other sources. Hence, this characteristic of causation is scored as no evidence.

Scoring—No evidence.

Conclusions

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. CASE STUDY METHODS
  5. RESULTS
  6. Conclusions
  7. DISCUSSION
  8. SUPPLEMENTAL DATA
  9. Acknowledgements
  10. REFERENCES
  11. Supporting Information

The evaluation of the body of evidence showed that the available evidence supports a causal relationship between mixtures of matrix ions in streams of ecoregions 68, 69, and 70 and resulting biological impairments. That conclusion is based on evidence showing that the relationship of conductivity to the loss of aquatic genera has the characteristics of causation. The six characteristics of causation are summarized below.

Co–occurrence

Loss of genera occurs when conductivity is high but is rare when conductivity is low (+ + +).

Preceding causation

Sources of the ionic mixture are present and are shown to increase stream conductivity in the region (+ + +).

Interaction

Aquatic organisms are directly exposed to dissolved ions. Based on first principals of physics, ionic gradients in high-conductivity streams would not favor the exchange of ions across gill epithelia. Physiological studies over the last 100 years have documented the many ways that physiological functions of organisms are affected by the relative amounts and concentrations of ions (i.e., combinations of ions that some genera do not have mechanisms or the capacity to regulate; + +).

Alteration

Some genera and other response metrics and assemblages are affected at sites with higher conductivity, whereas others are not. These differences are characteristic of high conductivity (+ + +).

Sufficiency

Laboratory analyses report results of effects for a tolerant species, but test durations and most ionic compositions are not representative of exposure in streams. However, regular increases in effects on invertebrates with increased exposure to ions, based on field observations, indicate that exposures are sufficient (+ + +).

Time order

Conductivity is high and extirpation has occurred after mining permits are issued, but conductivity and biological data before and after mining began are not available (no evidence).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. CASE STUDY METHODS
  5. RESULTS
  6. Conclusions
  7. DISCUSSION
  8. SUPPLEMENTAL DATA
  9. Acknowledgements
  10. REFERENCES
  11. Supporting Information

We have shown how evidence of causal characteristics was used to determine that increased concentrations of a specific ion mixture are a cause of extirpation of benthic invertebrates in Appalachian streams. Although no analysis can prove that an observed association is causal, the strength of the body of evidence is sufficiently convincing to support action. In particular, it was instrumental for assuring the U.S. EPA that high-conductivity effluents in Appalachia were causing extirpation of aquatic life and that this information provided the best science available for decision making 48, 49. A rigorous assessment using multiple lines of evidence was necessary because assessments that use relationships in field data achieve realism at the potential expense of known causation. No statistical analysis can resolve this problem, because, as we all learned, correlation does not equal causation. That aphorism applies to all statistical analyses of associations between variables because they simply quantify the consistency of the association. Hence, we can only hypothesize causal relationships, refute some, and determine how well the evidence supports others 3. We believe that this is best done by a consistent and transparent process of weighing the available and relevant evidence 13. In fact, we believe that most published reports of field studies describe associations and that the present study is one of the few that demonstrates that the relationship is causal.

This causal assessment does not attempt to identify constituents of the mixture that account for the effects. Instead, it shows that the mixture in streams with elevated conductivity and neutral or somewhat alkaline waters in Appalachia can cause and is causing the extirpation of sensitive genera of macroinvertebrates. Laboratory-based physiological evidence suggests that the relative amounts of ions as well as the concentrations of individual ions determine the toxic mechanisms. Failure of HCOmath image-mediated regulation of multiple ions in cells, particularly H+, Na+, and Cl, is one potential mode of action.

This causal assessment does not compare the relative importance of ionic-induced extirpation of genera in Appalachia with other known stressors in the region such as metal toxicity, stream bed siltation, or eutrophication 7, 44. Instead, it determines that addition of the ionic mixture to streams can and does cause extirpation of aquatic invertebrates 1, 2.

Likewise, this assessment does not evaluate how well any model predicts the effects of ionic stress. However, any such model would depend for its causal claims on this assessment.

The causal relationship describes in general how Ephemeroptera and other invertebrates respond to increased concentration of ions in water. A general causal relationship does not require that the species or genera be the same in all applications or at all locations. Therefore, we expect that ionic concentrations sufficient to cause extirpations would occur with a similar salt mixture containing HCOmath image, SOmath image, Ca2+, and Mg2+ in other regions with naturally low conductivity, because the assessment is of general causation for this salt mixture and susceptible stream invertebrates.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. CASE STUDY METHODS
  5. RESULTS
  6. Conclusions
  7. DISCUSSION
  8. SUPPLEMENTAL DATA
  9. Acknowledgements
  10. REFERENCES
  11. Supporting Information

We thank the West Virginia Department of Environmental Protection for their cooperation and data. Anonymous and named reviewers improved the manuscript: M. Griffith, C. Delos, M. Passmore, J. VanSickle, C. Schmitt, C. Menzie, C. Hawkins, and members of the U.S. Environmental Protection Agency (U.S. EPA) Biological Advisory Committee. The U.S. EPA Science Advisory Board provided review, interdisciplinary insights, and encouragement: D. Patten, E. Boyer, W. Clements, J. Dinger, G. Geidel, K. Hartman, R. Hilderbrand, A. Huryn, L. Johnson, T.W. LaPoint, S.N. Luoma, D. McLaughlin, M.C. Newman, T. Petty, E. Rankin, D. Soucek, B. Sweeney, and R. The article was prepared for publication by D. Kleiser, C. Lewis, S. Moore, and L. Wood of EC Flex, Inc., and L. Kessler, K. Secor, and L. Tackett of IntelliTech Systems. The present study is based on work supported by the U.S. EPA. The views expressed in this article are those of the authors and do not necessarily represent the views or policies of the U.S. EPA.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. CASE STUDY METHODS
  5. RESULTS
  6. Conclusions
  7. DISCUSSION
  8. SUPPLEMENTAL DATA
  9. Acknowledgements
  10. REFERENCES
  11. Supporting Information
  • 1
    Cormier SM, Suter GW II, Zheng L. 2013. Derivation of a benchmark for freshwater ionic strength. Environ Toxicol Chem 32:263271 (this issue).
  • 2
    U.S., Environmental Protection Agency. 2011. A field based aquatic life benchmark for conductivity in Central Appalachian streams. EPA 600/R-10/023F. Office of Research and Development, National Center for Environmental Assessment, Washington, DC.
  • 3
    Cormier S, Suter GW II. 2013. A method for assessing causation of field exposure-response relationships. Environ Toxicol Chem 32:272276 (this issue).
  • 4
    Suter GW II, Cormier SM. 2013. A method for assessing the potential for confounding applied to ionic strength in Central Appalachian streams. Environ Toxicol Chem 32:288295 (this issue).
  • 5
    U.S., Environmental Protection Agency. 2003. Generic ecological assessment endpoints (GEAEs) for ecological risk assessment. EPA 630/P-02/004B. Risk Assessment Forum, Washington, DC.
  • 6
    Pond GJ, Passmore ME, Borsuk FA, Reynolds L, Rose CJ. 2008. Downstream effects of mountaintop coal mining: Comparing biological conditions using family- and genus-level macroinvertebrate bioassessment tools. J N Am Benthol Soc 27:717737.
  • 7
    U.S. Environmental Protection Agency. 2011. The effects of mountaintop mines and valley fills on aquatic ecosystems of the Central Appalachian coalfields. EPA 600/R-09/138F. Office of Research and Development, National Center for Environmental Assessment, Washington, DC.
  • 8
    Evans DH. 2008. Teleost fish osmoregulation: What have we learned since August Krogh, Homer Smith, and Ancel Keys? Am J Physiol Reg I 295:R704R713.
  • 9
    Evans DH. 2008. Osmotic and Ionic Regulation: Cells and Animals. CRC, Boca Raton, FL, USA.
  • 10
    Remane A. 1971. Ecology of brackish water. In Remane A, Schlieper C, eds, Biology of Brackish Water, 2nd ed. John Wiley & Sons, New York, NY, USA.
  • 11
    Bradley TJ. 2009. Animal Osmoregulation. Oxford University Press, New York, NY, USA.
  • 12
    Cormier SM, Suter GW II, Norton SB. 2010. Causal characteristics for ecoepidemiology. Hum Ecol Risk Assess 16:5373.
  • 13
    Suter GW II, Cormier SM. 2011. Why and how to combine evidence in environmental assessments: Weighing evidence and building cases. Sci Total Environ 409:14061417.
  • 14
    Woods AJ, Omernik JM, Brown DD, Kiilsgaard CW. 1996. Level III and IV ecoregions of Pennsylvania and the Blue Ridge Mountains, the Ridge and Valley, and the Central Appalachians of Virginia, West Virginia, and Maryland. EPA 600/R-96/077. U.S. Environmental Protection Agency, Office of Research and Development, Corvallis, OR, USA.
  • 15
    Hagiwara S, Jaffe LA. 1979. Electrical properties of egg cell membranes. Annu Rev Biophys Bioeng 8:385416.
  • 16
    Kennedy AJ, Cherry DS, Currie RJ. 2004. Evaluation of ecologically relevant bioassays for a lotic system impacted by a coal-mine effluent, using Isonychia. Environ Monit Assess 95:3755.
  • 17
    Komnick H. 1977. Chloride cells and chloride epithelia of aquatic insects. Int Rev Cytol 49:285328.
  • 18
    Merricks TC, Cherry DS, Zipper CE, Currie RJ, Valenti TW. 2007. Coal-mine hollow fill and settling pond influences on headwater streams in southern West Virginia, USA. Environ Monit Assess 129:359378.
  • 19
    Nelson DL, Cox MM. 2005. Lehninger Principles of Biochemistry, 4th ed. WH Freeman & Co, New York, NY, USA.
  • 20
    O'Donnel MJ. 2011. Mechanisms of excretion and ion transport in invertebrates. Comprehen Physiol. DOI: 10.1002/cphy.cp130217.
  • 21
    Smith DG. 2001. Pennak's Freshwater Invertebrates of the United States: Porifera to Crustacea, 4th ed. John Wiley & Sons, New York, NY, USA.
  • 22
    Tarin JJ, Cano A. 2000. Fertilization in Protozoa and Metazoan Animals: Cellular and Molecular Aspects. Springer Verlag, Berlin, Germany.
  • 23
    Ziegler CR, Suter GW II, Ben JK. 2010. Candidate cause: Ionic strength. CADDIS, Vol 2: Sources, Stressors, and Responses. U.S. Environmental Protection Agency, Washington, DC.
  • 24
    Bryant G, Mcphilliamy S, Childers H. 2002. A survey of the water quality of streams in the primary region of mountaintop/valley fill coal mining, October 1999 to January 2001. Draft Programmatic Environmental Impact Statement on Mountaintop Mining/Valley Fills in Appalachia—2003 Appendix D. U.S. Environmental Protection Agency, Region 3, Philadelphia, PA, USA.
  • 25
    Timpano A, Schoenholtz S, Zipper C, Soucek D. 2011. Levels of dissolved solids associated with aquatic life effects in headwater streams of Virginia's Central Appalachian coalfield region. Virginia Tech, Blacksburg VA, USA.
  • 26
    Bunge M. 1979. Causality and Modern Science, 3rd revised ed. Dover, New York, NY, USA.
  • 27
    Hem JD. 1985. Study and Interpretation of the Chemical Characteristics of Natural Waters, 3rd ed. USGS Water Supply Paper 2254. U.S. Government Printing Office, Washington, DC, pp 263.
  • 28
    Forman RTT, Alexander LE. 1998. Roads and their major ecological effects. Annu Rev Ecol System 29:207231.
  • 29
    Herlihy AT, Stoddard JL, Johnson CB. 1998. The relationship between stream chemistry and watershed land cover data in the mid-Atlantic region, U.S. Water Air Soil Pollut 105:377386.
  • 30
    U.S., Department of Energy. 2004. A white paper describing produced water from production of crude oil, natural gas, and coal bed methane. U.S. Department of Energy, National Energy Technology Laboratory, Morgantown, WV.
  • 31
    Carey RO, Migliaccio KW. 2009. Contribution of wastewater treatment plant effluents to nutrient dynamics in aquatic systems: A review. Environ Manage 44:205217.
  • 32
    Bernot MJ, Tank JL, Royer TV, David MB. 2006. Nutrient uptake in streams draining agricultural catchments of the Midwestern United States. Freshwater Biol 51:499509.
  • 33
    Swank WT, Douglass JE. 1977. Nutrient budgets for undisturbed and manipulated hardwood ecosystems in the mountains of North Carolina. In Correll DL, ed, Watershed Research in Eastern North America: A Workshop to Compare Results. Smithsonian Institution, Edgewater, MD, USA, pp 343363.
  • 34
    Golladay SW, Webster JR, Benfield EF, Swank WT. 1992. Changes in stream stability following forest clearing as indicated by storm nutrient budgets. Arch Hydrobiol 1:133.
  • 35
    Arthur MA, Coltharp GB, Brown DL. 1998. Effects of best management practices on forest streamwater quality in eastern Kentucky. J Am Water Resour Assoc 34:481495.
  • 36
    Likens GE, Bormann FH, Johnson NM, Fisher DW, Pierce RS. 1970. Effects of forest cutting and herbicide treatment on nutrient budgets in the Hubbard Brook watershed ecosystem. Ecol Monogr 40:2347.
  • 37
    Pond GJ. 2010. Patterns of Ephemeroptera taxa loss in Appalachian headwater streams (Kentucky, USA). Hydrobiologia 641:185201.
  • 38
    Cormier S, Wilkes S, Zheng L. 2012. Relationship of land use and elevated ionic strength in Appalachian watersheds. Environ Toxicol Chem 32:296303 (this issue).
  • 39
    Cormier SM, Suter GW II. 2012. A method for assessing the potential for confounding applied to ionic strength in Central Appalachian streams. Environ Toxicol Chem 32:288295 (this issue).
  • 40
    Lindberg T, Bernhardt E, Bier R, Helton A, Merola R, Vengosh A, Di Giulio R. 2011. Cumulative impacts of mountaintop mining on an Appalachian watershed. Proc Natl Acad Sci U S A 108:2092920934.
  • 41
    Sutcliffe DW. 1962. The composition of Haemolymph in aquatic insects. J Exp Biol 39:325343.
  • 42
    Thorp JH, Covich AP. 2001. Ecology and Classification of North American Freshwater Invertebrates, 2nd ed. Academic, New York, NY, USA.
  • 43
    Wood CM, Shuttleworth TJ. 2008. Cellular and Molecular Approaches to Fish Ionic Regulation. Academic, San Diego, CA, USA.
  • 44
    Gerritsen J, Zheng L, Burton J, Boschen C, Wilkes S, Ludwig J, Cormier SM. 2010. Inferring causes of biological impairment in the Clear Fork Watershed, West Virginia. EPA 600/R-08/146. U.S. Environmental Protection Agency, Office of Research and Development, National Center for Environmental Assessment, Cincinnati, OH, USA.
  • 45
    Hill AB. 1965. The environment and disease: Association or causation. Proc R Soc Med 58:295300.
  • 46
    Kennedy AJ, Cherry DS, Currie RJ. 2003. Field and laboratory assessment of a coal processing effluent in the Leading Creek Watershed, Meigs County, Ohio. Arch Environ Contam Toxicol 44:324331.
  • 47
    Gerritsen J, Burton J, Barbour MT. 2000. A stream condition index for West Virginia wadeable streams. Prepared for the U.S. Environmental Protection Agency by Tetra Tech Inc, Owings Mills, MD, USA.
  • 48
    Stoner N, Giles C. 2011. Improving EPA review of Appalachian surface coal mining operations under the Clean Water Act, National Environmental Policy Act, and the Environmental Justice Executive Order. In July 21, 2011 memorandum to Garvin S, Fleming GK, Hedman, S. U.S. Environmental Protection Agency, Washington, DC.
  • 49
    Ward KJ. 2011. Big news: EPA finalizes water quality guidance to reduce impacts of mountaintop removal mining. Coal Tattoo. The Charleston Gazette, Charleston, WV, USA.
  • 50
    Pond GJ, Bailey JE, Lowman B. 2008. West Virginia GLIMPSS (genus-level index of most probable stream status): A benthic macroinvertebrate index of biotic integrity for West Virginia's wadeable streams. West Virginia Department of Environmental Protection, Division of Water and Waste Management, Watershed Branch, Charleston, WV, USA.

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. CASE STUDY METHODS
  5. RESULTS
  6. Conclusions
  7. DISCUSSION
  8. SUPPLEMENTAL DATA
  9. Acknowledgements
  10. REFERENCES
  11. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
etc_2059_sm_SupplData.doc702KSupplementary Data

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.