This is a text-only version of the document "Spruce No 1 Mine - Recommended Determination - Appendix 4 - 2010". To see the original version of the document click here.

---


APPENDIX 4 SELENIUM


A4.1 Abstract
Although the Mingo Logan mining company asserts that their plans will not release excess
selenium, the arguments are not well supported by evidence or previous mining in the area
There is abundant selenium in the coal deposits. Conditions that make selenium
bioavailable occur at mine sites in adjacent mines operated by the same company. The
documentation does not demonstrate that the company can effectively distinguish material
requiring isolation from water from those that have low selenium and might be used in
valley fill. Furthermore, they do not provide convincing evidence that water will not
interact with these exposed deposits or that the selenium will be removed when the coal is
removed. Thus, EPA contends that there is a strong likelihood that toxic levels of
selenium will be released if the Spruce l mine continues as planned.


A4.2 Fate of Selenium in the Environment
Selenium naturally exists in different oxidation states including oxyanions [selenate
(SeO42l) and selenite (SeO32l)], reduced selenium [selenide (Se2`)] and elemental selenium
(Se). Selenate and selenite oxyanions, common in oxidizing environments such as the
potentially exposed mine materials in valley fill areas, are the most mobile and toxic of the
selenium species. The presence of aluminum (Al) and/or iron (Fe) oxides in the solid
phase may result in substantially higher Se sorption as selenide (Se2`) oxidizes to selenite
(Se4+) which could then sorb to FeO(OH). However, given the potentially alkaline nature
of the mine materials from the proposed Spruce No. 1 mine, specific adsorption of


selenium species onto geologic (or valley-filled materials) materials is not likely to play an
important role in reducing Se mobility. This is because any near-field Se plume in the
valley-filled materials will be sufficiently alkaline to greatly reduce the number of anion
exchange sites for the Se oxyanions. Oxidation ofthe selenite to the highly mobile
selenate could then take place. Also, given the depositional history and nature of
formation of the rocks above and below the coal layers in West Virginia coal formations,
significant concentrations of sulfates are expected in these valley-filled materials. Some of
the sulfates will form as the metal sulfides in the valley-filled materials are exposed to the
environment. Sulfates are known to compete with Se oxyanions for mineral adsorption
sites making the Se oxyanions more available for leaching to surface Waters and other
environments.


While Se oxyanions have been reported to have lower potential for bioaccumulation in fish
and other aquatic organisms, a reducing environment that may be induced in streams could
reduce the oxyanions to forms that could easily bioaccumulate in aquatic species. A more
desired outcome will be to decrease Se mobility in the valley-filled materials which will
require a reducing environment. It is Well known that reduction of Se is largely controlled
by microbial processes (Fevrier et al., 2007). ln the valley-filled materials, We expect the
subsurface to be physically complex and redox environments (at the microscale), and
associated metabolic processes, to be heterogeneously distributed in both space and time.
In particular, within the deeper deposits, oxygen is less available resulting in a chemically
reducing environment.




A4.3 Predicting Fate of Se in Valley-iilled Materials
In the absence of mechanistic physical, chemical, and microbiological models, the Agency
can only use empirical models of Se partitioning between subsurface compartments as a
basis to predict how Se will behave in these valley-filled materials. The Kd parameter for
Se (and indeed for other toxic metals that may be present in the valley-filled materials) is
an important parameter to estimate the potential for the adsorption of dissolved
contaminants in contact with soil. Kd (L/kg) is typically defined as the ratio of the
contaminant concentration associated with the solids (mg/kg) to the contaminant
concentration in the surrounding aqueous solution (mg/L) when the system is at
equilibrium. It is recognized that generic or default Kd values in the literature can result in
significant errors when used to evaluate contaminant migration or site-remediation options.
Hence site-specific Kd values derived using site-specific data are preferable. However, in
the absence of such data, relevant published values may be used. Consequently, a literature
survey for partition coefficients of Se was conducted. The following criteria and
guidelines developed in the U.S. Environmental Protection Agency (USEPA) document
(U SEPA, 2005) were used to evaluate the published Kd values.


Use “whole” natural media for determination of Kd in natural media systems
(meaning the rejection of Kd values from studies using pure mineral phases or
treated soils).
Use low total metal concentrations (i.e., if Kd were determined at sites with
multiple total metal concentrations, Kd Values corresponding to the lowest metal


concentrations Where Kd is less likely to depend on metal concentrations Where
chosen).
Use pH Values in the natural range of 4-10 (given the expected pH Values of soils
and rock formations that will be Valley-filled).
Partition coefficients with no organic chelates (e.g., EDTA) in the extractant.
Where multiple Kd Values are presented for a system due to experimental variation
of pH or other parameters, choose Kd Values corresponding to the conditions most
closely approximating the geologic conditions of the materials that may be Valley-
filled.


The effective range of Se-Kd values in the ambient vadose zone of sediments is 0 to 0.78
L/kg (Kaplan & Seme, 1995). A Kd of zero means Se is entirely associated with the
solution phase with a very high potential of being released into surface and/ or ground
Waters. Using a conservative estimate of 2 mg/kg (or 2,000 ug/kg) in the valley-filled
materials and the maximum estimated Kd of 0.78 L/kg, at equilibrium; We would expect
0.88 mg/kg (or 880 ug/kg) to remain adsorbed on the solid phase and 1.12 mg/L (or 1,120
ug/L) to be released into the soil solution. If we assume 1 ug = 1 mL (which is reasonable
for dilute systems), then 1,120 ug/mL in solution is equivalent to 1,120 parts per million
(ppm). This is 224 times the 5 ppm criterion for Se in receiving streams. How much of
this will end up in receiving streams or groundwater will require further modeling.




However, in the absence of data, we should take the conservative approach and assume
that this leachable mass will end up in streams and/ or groundwater. This may be a realistic
assumption given that the 2 mg/kg used in this example is half of the mean Se
concentration reported for WV coals (see next section). Some horizons above and below
the coal beds have also been reported to have more than 2 mg/kg. As indicated previously,
soil microbial activity may increase the Se Kd. This will happen if mobile Se oxyanions
are reduced to species that would readily adsorb on the solid phase and decrease the
mobility of Se. It is, however, unclear whether the conditions at the proposed valley-filled
materials will be conducive to microbial growth.


Evidence is lacking that might demonstrate that Se in the materials to be deposited in the
Spruce Mine valley-fills will be immobile and poses no environmental threat. This would
require the development of comprehensive and predictive models that quantify the
anticipated dynamics of the biogeochemical systems including variations in the redox
processes operating at the microscale in the proposed valley-filled materials.


A4.4 Differences in Geologic Environment
The selenium content of West Virginia (WV) coals has been thoroughly examined and
these data are available at the West Virginia Geological and Economic Survey (WVGES)
website (http://wwW.wvgs.wvnet.edu/www/datastat/te/index.htm) and are summarized in
Figure 1. From a total of 845 samples, Se concentrations are reported to range from non-
detects to 21.3 mg/kg with a mean concentration of 4.2 mg/kg and standard deviation of
2.83 mg/kg.




Figure 1: Selenium by coal bed in West Virginia coals
Q
_1
91
(E.
Q
:r
-cz
O
cn
Q
U)
m
5
'cm
3
Q.
‘2
5
1 onsfliro
nFI:|
ach
Appa
ASHINGTON
AYNESELIRD A
AYNIIILIRD
X IWICKLIY
' E DS T ONE
~ ITTSI LIKH
,~" ITTLE PITTSHJRGH
." ILK LICK
ARE"
~ KIRBTUWN
J RUSH Clllli
AHOHING
uulnrulvom
, FFER KITTANNIHO
fl, HIDDLI KITTAN NING
_Jr LOWER KITTAHHING
~_." NO. I ILOCK
1‘_.UFFERH0.5 BLDC
/
O. I ILOC
T0 CKTO
PPIR MIKII.
T0 CKTO
OALBUIB
NIFRE Di
HILTON
I`I"I'Li CIIILTON
Ill CLAY
EDRR GRO
LLIAM SON
ii|i.Li5
O 2 Gb!
OWELLTON
IDDLE “AR EQG L
ENS Cl!!
UWER VYQR EAGLE
LENALLIM '|’l.|NN!L
ILBE RT
OUGLAS
UWER D0 LI GLA5
EOER
EVYELL B
EVYELL A
EVYELL
LCH
!CH.L!‘|"
IRE CREEK
OCAHONTAS N0
GCLHONTAS N6
OCQHONTJIE N0
DGAHDNTAS N0
OCQHONTJIE N0
8E2‘l18N¥%iH8 3
OCNIOHTAS H0
OCAHUNTA! N0 2
RIQII
I I I I
` II-7| '
-|.,~ '
Il|-I'
E"
' E z z
" )
I. ii
. L. _ h. . .
I
I
ll ul -I ul Q ll llllul-lub-I » L Q -I ul ul
\l5¢'lIONl¢C¢l0UU|ldU\¢ l4liWll|§llW'l§llU IIIIUCNUYOG' "IIDCOIUUVU Glldhlllldlii
123| QTBBI I
:de
1
I
4
15
rl!
Llpuneuaaanary 13, 2001 (II Izzy
Pittsburgh
Kittann in g
Coalburg &
‘Mnifrede
WVGES, 2002
(coals only)


From the analysis of Se in the different coal beds, coals containing the highest selenium
contents are in a region of south central WV which includes the region of the proposed
Spruce No l mine. ln addition, the coal beds of the Allegheny and Upper Kanawha
Formations exhibit high Se contents compared to coal beds both lower and higher in the
geologic sequence. Formation cores analyzed by USGS (Paybins et al., 2000; Sandra et
al., 2005) show similar trends. The coal beds to be targeted by the proposed Spruce No 1
mine include 5 -Block of the Allegheny Group and down to the Upper Stockton coal bed in
the eastem portion of the project area. ln the westem portion of the project area, they plan
to extract coal through the Middle Coalburg coal bed. These coal beds are enriched in Se
as evidenced by Se distribution data in the Spruce No. 1 column (DT 0417) provided by the
applicant for the NPDES permit application (Figure 2). Due to space limitations, the
profile has been modified to show only coal beds encountered during drilling and the rocks
above and below the coal beds that may be removed and deposited in valley fills during
mining. The complete profile with associated Se for each formation (coal and non-coal
formations) is provided. Similar Se distributions in Coalburg coal and Winifrede coal at
the Bull Creek Mine (Figure 3) were also reported by Vesper & Rhoads (2008).




Figure 2: Geologic column DT0417


Strain Rock Se Sample
Thick. (Feet) Type (mg/kg) |,D. No.
 E 
 &  05  E,
l---l IQ
5-Elm  11  gg  napa
'   _  
§‘;‘;’ at i °-S § 
EI E- *
  ~ I I I
l l l
1111111 0.46 57 EE-1 , 103 
UPPBI' _ 
Sf°=‘<*°" l l! 10 Cm EEE.;-
S1 -E111
 









     





          						



			


































 















	
Breathe Theme |