Statistical Support Document Effluent Limitations for FGD Wastewater Gasification Wastewater and Combustion Residual Leachate for the Final Steam Electric Power Generating Effluent Limitations Guidelines and Standards


United States
Environmental Protection
Agency	

Office of Water
Washington, DC

September 2015

© CDA	p • •

vcrH	Statistical Support Document:

Effluent Limitations for FGD
Wastewater, Gasification
Wastewater, and Combustion
Residual Leachate for the Final
Steam Electric Power Generating
Effluent Limitations Guidelines
and Standards

-------
STATISTICAL SUPPORT DOCUMENT:

EFFLUENT LIMITATIONS FOR FGD WASTEWATER, GASIFICATION WASTEWATER,

AND COMBUSTION RESIDUAL TKACHATE FOR THE FINAL STEAM ELECTRIC
POWER GENERATING EFFLUENT LIMITATIONS GUIDELINES AND STANDARDS

U.S. Environmental Protection Agency
Office of Water
Washington, DC

September 2015

u

-------
Acknowledgements
and Disclaimer

This report has been reviewed and approved for publication by the Engineering and Analysis
Division, Office of Water, U.S. Environmental Protection Agency. This report was prepared with
the support of Westat under Contract EP-C-10-023, under the direction and review of the Office of
Water. Neither the United States government nor any of its employees, contractors, subcontractors,
or other employees makes any warranty, expressed or implied, or assumes any legal liability or
responsibility for any third party's use of, or the results of such use of, any information, apparatus,
product, or process discussed in this report, or represents that its use by such a third party would
not infringe on privately owned rights.

111

-------
Executive Summary

This report presents the results of the statistical analyses performed on the Flue Gas Desulfurization
(FGD) and gasification wastewater treatment data and the resulting long-term averages, variability
factors, and effluent limitations. The treatment technology options for FGD wastewater presented
in this report are: (i) chemical precipitation system employing both hydroxide and sulfide
precipitation and iron coprecipitation (hereafter referred to as "chemical precipitation"), (ii)
biological (chemical precipitation employing both hydroxide and sulfide precipitation and iron
coprecipitation followed by anoxic/anaerobic biological treatment, hereafter referred to as
"biological"), and (iii) chemical precipitation followed by softening and vapor-compression
evaporation system (hereafter referred to as "vapor-compression evaporator." The treatment
technology option for gasification wastewater is a vapor-compression evaporation system. See the
Technical Development Documentfor the Final Affluent limitations Guidelines and Standards for the Steam
Electric Power GeneratingPoint Source Category (hereafter referred to in this document as Technical
Development Document) for a detailed description of the FGD and gasification systems and the
wastewater associated with each system, rationale for selecting plants used in the calculation of the
effluent limitations for each technology option, and pollutants for regulation for each technology
option.

The following bullets summarize the plants from which data were used to calculate the limits for
each technology option for FGD wastewater together with the pollutants that would be regulated:

¦	The effluent limitations for chemical precipitation technology option are based on data
from Hatfield's Ferry, Keystone, Miami Fort, and Pleasant Prairie. The pollutants that
would be regulated for this technology option are arsenic and mercury.

¦	The effluent limitations for biological technology option are based on Allen and Belews
Creek. The pollutants that would be regulated for this technology option are arsenic,
mercury, nitrate-nitrite as N, and selenium. While these two plants operating the
biological treatment system were used as the basis for the technology option, neither of
these plants include sulfide precipitation in the upstream chemical precipitation system.
For this reason, EPA is transferring the arsenic and mercury limitations calculated based
on the chemical precipitation treatment to the biological treatment for FGD wastewater
(see Section 13 of the Technical Development Document for a detailed discussion of
the transfer of limitations).

¦	The effluent limitations for the vapor-compression evaporation technology option are
based on Brindisi. The pollutants that would be regulated for this technology option are
arsenic, mercury, selenium, and total dissolved solids.

IV

-------
The following bullet summarizes the plants from which data were used to calculate the effluent
limitations for gasification wastewater together with the pollutants that would be regulated:

¦ The effluent limitations for vapor-compression evaporation technology option are
based on Polk and Wabash River. The pollutants that would be regulated for this
technology option are arsenic, mercury, selenium, and total dissolved solids.

In addition to the technology options for FGD and gasification wastewaters discussed in the
previous paragraphs, EPA is also establishing effluent limitations for combustion residuals leachate
(hereafter referred to as leachate) at certain facilities based on the chemical precipitation technology
option. The effluent limitations for leachate are based on transferring the effluent limitations
calculated for the chemical precipitation technology option for FGD wastewater because EPA does
not have available data for chemical precipitation treatment of leachate. This transfer of effluent
limitations is appropriate because the pollutants in leachate are similar to those in FGD wastewater
and would be removed by the chemical precipitation technology evaluated for FGD wastewater. See
Section 13 of the Technical Development Document for a detailed discussion of the transfer of
limitations.

Section 1 of this report provides an overview of the document. Section 2 provides a brief
description of the available data used to develop the effluent limitations for each technology option.
Section 3 describes the data corrections, exclusions, baseline substitutions, and aggregation made to
the data before calculating the effluent limitations. Section 4 describes the data editing procedures
used to select datasets for developing the effluent limitations for each pollutant at each plant.

Section 5 provides the statistical methodology used to calculate the effluent limitations together with
the long-term average and variability factors. Sections 6, 7, and 8 contain a detailed summary of the
data together with the effluent limitations for each pollutant in each treatment technology option for
FGD wastewater. Section 9 contains a detailed summary of the data together with the effluent
limitations for each pollutant for the gasification wastewater treatment technology option. Section
10 presents the overall summary of the long-term average, variability factors, and effluent limitations
for each pollutant in each technology option. Section 11 provides a discussion of the engineering
review of the effluent limitations to verify that the limitations are reasonable based upon the design
and expected operation of the control technologies.

In addition to the eleven sections described above, seven appendices are attached to the report.
Appendix 1 contains a list of the data that were corrected due to entry errors. Appendix 2 identifies
the data that were excluded, along with the reasons for the exclusions. Appendix 3 contains
longitudinal plots (along with smoothed curves superimposed over the data) for the plant self-

-------
monitoring data that EPA used to aid in determining the start-up or initial commissioning period for
the treatment systems at Allen and Belews Creek. Appendix 4 contains plots of all available data for
each plant used in calculating the limits for FGD wastewater based on the chemical precipitation or
biological treatment technology options, prior to any of the exclusions noted in Appendix 2.
Appendix 5 contains plots used to help identify outlier observations for each plant used in
calculating the limits for FGD wastewater based on the chemical precipitation or biological
treatment technology options. Appendix 6 contains a summary of the results of the data editing
criteria (i.e., long-term average test). Appendix 7 contains a summary of the results of the
engineering review of the limitations.

VI

-------
Table of Contents

Chapter	Page

Acknowledgements and Disclaimer		iii

Executive Summary		iv

1	Introduction		1

2	Description of Data Used to Calculate Long-term Averages,

Variability Factors, and Effluent Limits		2

2.1	Overview of the Data		2

2.2	Analytical Methods		3

2.3	Combining Data from Multiple Sources within a Plant		4

3	Data Corrections, Exclusions, Baseline Substitutions, and

Aggregation		5

3.1	Data Corrections		5

3.2	Data Exclusions and Rationale for the Exclusions		6

3.3	Baseline Substitutions		7

3.4	Data Aggregation		9

3.4.1	Aggregation of Field Duplicates		9

3.4.2	Aggregation of Overlapping Samples		10

4	Data Editing Criteria		11

5	Statistical Methodology		12

5.1	Definitions and Background		12

5.2	Statistical Model Selected for the Data		12

5.3	Autocorrelation Analysis		13

5.4	Dataset Requirement		15

5.5	Calculating Effluent Limits for Each Pollutant in Each
Technology Option		16

5.6	Special Case — Effluent Limits Based on Detection Limits		17

6	Treatment Technology Option for FGD Wastewater: Chemical
Precipitation		18

6.1 Chemical Precipitation Treatment for FGD Wastewater:

Arsenic		20

vii

-------
Table of Contents
(continued)

Chapter	Page

6.1.1	Longitudinal Plots of the Data for Arsenic

(Mg/L)		;	 20

6.1.2	Summary Statistics for Arsenic (ng/L)	 20

6.1.3	Plant-specific Long-term Averages and
Variability Factors, Option Long-term Average
and Variability Factors, and Effluent Limits for

Arsenic (ng/L)	 23

6.2	Chemical Precipitation Treatment for FGD Wastewater:

Mercury	 25

6.2.1	Longitudinal Plots of the Data for Mercury

(ng/L)	 25

6.2.2	Summary Statistics for Mercury (ng/L)	 27

6.2.3	Plant-specific Long-term Averages and
Variability Factors, Option Long-term Average
and Variability Factors, and Effluent Limits for

Mercury (ng/L)	 28

6.3	Chemical Precipitation Treatment for FGD Wastewater:

Summary of the Option Long-term Averages, Option

Variability Factors, and Effluent Limits	 29

7	Treatment Technology Option for FGD Wastewater: Biological

Treatment	 31

7.1	Biological Treatment for FGD Wastewater: Arsenic	 32

7.1.1	Longitudinal Plots of the Data for Arsenic

(Mg/L)	•••••			 33

7.1.2	Summary Statistics for Arsenic (ng/L)	 35

7.2	Biological Treatment for FGD Wastewater: Mercury	 36

7.2.1	Longitudinal Plots of the Data for Mercury

(ng/L)			 36

7.2.2	Summary Statistics for Mercury (ng/L)	 36

7.3	Biological Treatment for FGD Wastewater: Nitrate-nitrite

as N	 39

Vlll

-------
Table of Contents
(continued)

Chapter	Page

7.3.1	Longitudinal Plots of the Data for Nitrate-nitrite

as N (mg/L)	 39

7.3.2	Summary Statistics for Nitrate-nitrite as N

(mg/L)	 41

7.3.3	Plant-specific Long-term Averages and
Variability Factors, Option Long-term Averages
and Variability Factors, and Effluent Limits for
Nitrate-nitrite as N (mg/L)	 42

7.4	Biological Treatment for FGD Wastewater: Selenium	 44

7.4.1	Longitudinal Plots of the Data for Selenium

(Mg/L)	 44

7.4.2	Summary Statistics for Selenium (ng/L)	 46

7.4.3	Plant-specific Long-term Averages and
Variability Factors, Option Long-term Averages
and Variability Factors, and Effluent Limits for

Selenium (ng/L)	 47

7.5	Biological Treatment for FGD Wastewater: Summary of
the Option Long-term Averages, Variability Factors, and

Effluent Limits	 49

8	Treatment Technology Option for FGD Wastewater: Vapor-

compression Evaporation System	 50

8.1	Vapor-compression Evaporation for FGD Wastewater:

Arsenic	 51

8.1.1	Longitudinal Plots of the Data for Arsenic

(Mg/L)	•••••			 51

8.1.2	Summary Statistics for Arsenic (ng/L)	 51

8.1.3	Plant-specific Long-term Averages and
Variability Factors, Option Long-term Averages
and Variability Factors, and Effluent Limits for

Arsenic (ng/L)	 52

8.2	Vapor-compression Evaporation for FGD Wastewater:

Mercury	 53

IX

-------
Table of Contents
(continued)

Chapter	Page

8.2.1	Longitudinal Plots of the Data for Mercury

(ng/L)	•••••	 53

8.2.2	Summary Statistics for Mercury (ng/L)	 54

8.2.3	Plant-specific Long-term Averages and
Variability Factors, Option Long-term Averages
and Variability Factors, and Effluent Limits for

Mercury (ng/L)	 55

8.3	Vapor-compression Evaporation for FGD Wastewater:

Selenium	 55

8.3.1	Longitudinal Plots of the Data for Selenium

(Mg/L)	 55

8.3.2	Summary Statistics for Selenium (ng/L)	 56

8.3.3	Plant-specific Long-term Averages and
Variability Factors, Option Long-term Averages
and Variability Factors, and Effluent Limits for

Selenium (ng/L)	 58

8.4	Vapor-compression Evaporation for FGD Wastewater:

Total Dissolved Solids (TDS)	 59

8.4.1	Longitudinal Plots of the Data for TDS (mg/L)	 59

8.4.2	Summary Statistics for TDS (mg/L)	 60

8.4.3	Plant-specific Long-term Averages and
Variability Factors, Option Long-term Averages
and Variability Factors, and Effluent Limits for

TDS (mg/L)	 61

8.5	Vapor-compression Evaporation for FGD Wastewater:

Summary of the Option Long-term Averages, Option

Variability Factors, and Effluent Limits	 62

9	Treatment Technology Option for Gasification Wastewater:

Vapor-compression Evaporation	 65

9.1 Vapor-compression Evaporation for Gasification:

Arsenic	 66

x

-------
Table of Contents
(continued)

Chapter	Page

9.1.1	Longitudinal Plots of the Data for Arsenic

(Mg/L)	 66

9.1.2	Summary Statistics for Arsenic (ng/L)	 68

9.1.3	Plant-specific Long-term Average and
Variability Factors, Option Long-term Average
and Variability Factors, and Effluent Limits for

Arsenic (ng/L)	 69

9.2	Vapor-compression Evaporation for Gasification:

Mercury	 70

9.2.1	Longitudinal Plots of the Data for Mercury

(ng/L)			 70

9.2.2	Summary Statistics for Mercury (ng/L)	 72

9.2.3	Plant-specific Long-term Average and
Variability Factors, Option Long-term Average
and Variability Factors, and Effluent Limits for

Mercury (ng/L)	 73

9.3	Vapor-compression Evaporation for Gasification:

Selenium	 74

9.3.1	Longitudinal Plots of the Data for Selenium

(Mg/L)	•••••		 74

9.3.2	Summary Statistics for Selenium (ng/L)	 76

9.3.3	Plant-specific Long-term Averages and
Variability Factors, Option Long-term Average
and Variability Factors, and Effluent Limits for

Selenium (ng/L)	 77

9.4	Vapor-compression Evaporation for Gasification: Total

Dissolved Solids (TDS)	 78

9.4.1	Longitudinal Plots of the Data for TDS (mg/L)	 78

9.4.2	Summary Statistics for TDS (mg/L)	 80

9.4.3	Plant-specific Long-term Averages and
Variability Factors, Option Long-term Average
and Variability Factors, and Effluent Limits for

TDS (mg/L)	 81

XI

-------
Table of Contents
(continued)

Chapter	Page

9.5 Vapor-compression Evaporation for Gasification:

Summary of the Option Long-term Average, Option

Variability Factors, and Effluent Limits	 82

10	Summary of Effluent Limits for FGD Wastewater, Gasification
Wastewater, and Combustion Residual Leachate	 84

11	Engineering Review of the Effluent Limits	 87

11.1	Comparison of the Limits to Effluent Data Used as the

Basis for the Limits	 89

11.2	Comparison of the Effluent Limits to Influent Data	 102

Appendix

1	List of All Observations that Were Corrected Before Calculating

the Limits	 1-1

2	List of All Observations that Were Excluded Before Calculating

the Limits for the Final Rulemaking	 2-1

3	Plots of Plant Self-monitoring Data with Smoothed Curves to
Aid in Determining Initial Commissioning Period for the

Treatment System	 3-1

4	Plots of Available Data for Each Plant for Calculating Limits

Prior to Any Exclusions	 4-1

5	Plots to Identify Potential Outlier Observations Not Consistent

with BAT/NSPS Operation of the Treatment System	 5-1

6	Data Editing Criteria Results	 6-1

7	Summary of the Results for the Engineering Review of the Limits	 7-1

xii

-------
Table of Contents
(continued)

Tables	Page

1.	Summary of the analytical lab methods used for each pollutant by

plant self-monitoring	 3

2.	Summary of the analytical lab methods used for each pollutant by

EPA sampling and CWA 308 sampling	 3

3.	Summary of the baseline values used for each pollutant	 8

4.	Aggregation of field duplicates	 10

5.	Summary of estimated autocorrelation values by plant and

pollutant	 15

6.	Numbers of detected and non-detected observations and sample-
specific detection limits for arsenic (ng/L) by plant and sampling

location	 22

7.	Summary statistics of arsenic concentration (ng/L) for all
detected and non-detected samples combined, by plant and

sampling location	 23

8.	Plant-specific long-term averages and variability factors, option
long-term average and variability factors, and limits for arsenic

(ng/L) in chemical precipitation effluent	 24

9.	Numbers of detected and non-detected observations and sample-
specific detection limits for mercury (ng/L) by plant and

sampling location	 27

10.	Summary statistics of mercury concentration (ng/L) for all
detected and non-detected samples by plant and sampling

location	 28

11.	Plant-specific long-term average and variability factors, option
long-term average and variability factors, and limits for mercury

(ng/L) in chemical precipitation effluent	 29

12.	Summary of option long-term averages, option variability factors,
and effluent limits for chemical precipitation FGD wastewater

treatment technology	 30

13.	Numbers of detected and non-detected observations and sample-
specific detection limits for arsenic (ng/L) by plant and sampling

location	 35

14.	Summary statistics of arsenic concentration (ng/L) for all
detected and non-detected samples combined, by plant and

sampling location	 36

15.	Numbers of detected and non-detected observations and sample-
specific detection limits for mercury (ng/L) by plant and

sampling location	 38

Xlll

-------
16,

17,

18,

19,

20,

21,

22,

23,

24,

25,

26,

27,

28,

29,

Table of Contents
(continued)

Page

Summary statistics of mercury concentration (ng/L) for all
detected and non-detected samples combined by plant and

sampling location	 39

Number of detected, non-detected, and sample-specific detection
limits for nitrate-nitrate as N (mg/L) by plant and sampling

location	 41

Summary statistics of nitrate-nitrite as N concentration (mg/L)
for all detected and non-detected samples combined by plant and

sampling location	 42

Lag-1 autocorrelation values, plant-specific long-term averages
and variability factors, option long-term averages and variability
factors, and limits for nitrate-nitrite as N (mg/L) in bioreactor

effluent	 43

Number of detected, non-detected, and sample-specific detection

limits for selenium (ng/L) by plant and sampling location	 46

Summary statistics of selenium concentration (ng/L) for all
detected and non-detected samples combined by plant and

sampling location	 47

Lag-1 autocorrelation values, plant-specific long-term averages
and variability factors, option long-term averages and variability

factors, and limits for selenium (ng/L) in bioreactor effluent	 48

Summary of option long-term averages, variability factors, and
effluent limits for biological FGD wastewater treatment

technology	 49

Numbers of detected and non-detected observations and sample-

specific detection limits for arsenic (ng/L) by sampling location	 52

Summary statistics of arsenic concentration (ng/L) for all
detected and non-detected samples combined by sampling

location	 52

Plant-specific long-term averages and variability factors, option
long-term averages and variability factors, and limits for arsenic

(|J.g/L) in brine concentrator distillate and crystallizer condensate	 53

Summary statistics of mercury concentration (ng/L) for all
detected and non-detected samples combined by sampling

location	 54

Plant-specific long-term averages and variability factors, option
long-term averages and variability factors, and limits for mercury

(ng/L) in brine concentrator distillate and crystallizer condensate	 55

Number of detected, non-detected, and sample-specific detection

limits for selenium (ng/L) by sampling location	 57

xiv

-------
30,

31.

32,

33,

34,

35,

36,

37,

38,

39,

40,

41,

42,

Table of Contents
(continued)

Page

Summary statistics of selenium concentration (ng/L) for all
detected and non-detected samples combined by sampling

location	 57

Plant-specific long-term averages and variability factors, option
long-term averages and variability factors, and limits for selenium

(Mg/L) in brine concentrator distillate and crystallizer condensate	 59

Number of detected, non-detected, and sample-specific detection

limits for total dissolved solid (mg/L) by sampling location	 61

Summary statistics of total dissolved solids concentration (mg/L)
for all detected and non-detected samples combined by sampling

location	 61

Plant-specific long-term averages and variability factors, option
long-term averages and variability factors, and limits for TDS

(mg/L) in brine concentrator distillate and crystallizer condensate	 62

Summary of option long-term averages, variability factors, and
limits for vapor-compression evaporation technology for FGD

wastewater in brine concentrator distillate	 63

Summary of option long-term averages, option variability factors,
and limits for mechanical vapor-compression evaporation

technology for FGD wastewater in crystallizer condensate	 64

Numbers of detected and non-detected observations and sample-
specific detection limits for arsenic (ng/L) by plant and sampling

location	 68

Summary statistics of arsenic concentration (ng/L) for all
detected and non-detected samples combined, by plant and

sampling location	 69

Plant-specific long-term average and variability factors, option
long-term average and variability factors, and limits for arsenic

(ng/L) in vapor compression evaporator condensate	 70

Numbers of detected and non-detected observations and sample-
specific detection limits for mercury (ng/L) by plant and

sampling location	 72

Summary statistics of mercury concentration (ng/L) for all
detected and non-detected samples combined, by plant and

sampling location	 73

Plant-specific long-term average and variability factors, option

long-term average and variability factors, and limits for mercury

(ng/L) in vapor compression evaporator condensate	 74

xv

-------
Table of Contents
(continued)

Tables	Page

43.	Numbers of detected and non-detected observations and sample-
specific detection limits for selenium (|J.g/L) by plant and

sampling location	 76

44.	Summary statistics of selenium concentration (|J.g/L) for all
detected and non-detected samples combined, by plant and

sampling location	 77

45.	Plant-specific long-term averages and variability factors, option
long-term average and variability factors, and limits for selenium

(jag/L) in vapor compression evaporator condensate	 78

46.	Numbers of detected and non-detected observations and sample-
specific detection limits for TDS (mg/L) by plant and sampling

location	 80

47.	Summary statistics of TDS concentration (mg/L) for all detected
and non-detected samples combined, by plant and sampling

location	 81

48.	Plant-specific long-term averages and variability factors, option
long-term average and variability factors, and numeric limits for

TDS (mg/L) in vapor compression evaporator condensate	 82

49.	Summary of option long-term averages, option variability factors,
and limits for vapor-compression evaporation technology for

gasification wastewater	 83

50.	BAT/PSES limits for existing sources: Summary of the option
long-term averages, variability factors, and limits for FGD and

gasification wastewater	 85

51.	NSPS/PSNS limits for new sources: Summary of the option
long-term averages, variability factors, and limits for FGD
wastewater, gasification wastewater, and combustion residual

leachate	 86

xvi

-------
Table of Contents
(continued)

Figures	Page

1.	Plot of arsenic (ng/L) data on a logarithmic scale for Hatfield's

Ferry, Keystone, Miami Fort, and Pleasant Prairie	 21

2.	Plot of mercury (ng/L) data on a logarithmic scale for Hatfield's

Ferry, Keystone, Miami Fort, and Pleasant Prairie	 26

3.	Plot of arsenic (ng/L) data on a logarithmic scale for Allen and

Belews Creek	 34

4.	Plot of mercury (ng/L) data on a logarithmic scale for Allen and

Belews Creek	 37

5.	Plot of nitrate-nitrite as N (mg/L) data on a logarithmic scale for

Allen and Belews Creek	 40

6.	Plot of selenium (ng/L) data on a logarithmic scale for Allen and

Belews Creek	 45

7.	Plot of arsenic (ng/L) data on a logarithmic scale for Brindisi.

The Brine Concentrator Distillate and Crystallizer Condensate

data were jittered slightly to enhance visibility	 51

8.	Plot of mercury (ng/L) data on a logarithmic scale for Brindisi	 54

9.	Plot of selenium (ng/L) data on a logarithmic scale for Brindisi	 56

10.	Plot of TDS (mg/L) data on a logarithmic scale for Brindisi	 60

11.	Plot of arsenic (ng/L) data on a logarithmic scale for Polk and

Wabash River	 67

12.	Plot of mercury (ng/L) data on a logarithmic scale for Polk and

Wabash River	 71

13.	Plot of selenium (ng/L) data on a logarithmic scale for Polk and

Wabash River	 75

14.	Plot of TDS (mg/L) data on a logarithmic scale for Polk and

Wabash River	 79

XVll

-------
1 Introduction

This report presents the results of the statistical analyses performed on the Flue Gas Desulfurization
(FGD) wastewater and gasification wastewater treatment data and the resulting long-term average,
variability factors, and effluent limitations for each pollutant in each technology option.

This report is focused primarily on the summary of the data and the development of long-term
averages, variability factors, and effluent limitations. See the Technical Development Document for
a description of (i) FGD and gasification systems and the wastewater associated with each system,
(ii) the treatment technology options for each waste stream, (iii) the rationale for selecting plants
used to calculate the effluent limitations for each technology option, (iv) the rationale for
transferring limitations for different technology options, or (v) a discussion of the baseline values
used for adjusting the data before calculating the effluent limitations.

In this document, for simplicity, the final effluent limitations guidelines and standards are referred to
as "limits." The terms option long-term averages and option variability factors refer to the
technology options rather than the regulatory options described in Section 8 of the Technical
Development Document. The term "detected" refers to analytical results measured and reported
above the sample-specific quantitation limit (QL); and the term "non-detected" refers to values that
are below the method detection limit (MDL) and those values measured by the laboratory as being
between the MDL and the QL in the original data (before adjusting for baseline). The long-term
averages, variability factors, and effluent limits were calculated out to five decimal points for
accuracy. However, to simplify the presentation the values presented in this document have been
rounded to three decimal points. All exclusions, baseline substitutions, and aggregation of data were
made prior to performing the statistical analyses presented in Sections 6, 7, 8, and 9. Finally, the
Excel Spreadsheet titled "Sampling Data Used as the Basis for Effluent Limitations for the Steam
Electric Rulemaking" contains the listing of all the available data, the excluded data, the baseline
values for each pollutant, and the final concentration values that were used in the limit calculations,
among other relevant data items. This Excel Spreadsheet is in the record as DCN SE06277 (this
spreadsheet will hereafter be referred to as DCN SE06277).

1

-------
2 Description of Data Used to Calculate Long-

term Averages, Variability Factors, and Effluent
Limits

This section provides an overview of the available data at each plant selected as the basis for the
effluent limits for each technology option, the analytical methods used to analyze each of the
pollutants, and the rationale for combining sampling data from multiple sources at each plant into a
single dataset for the plant.

2.1 Overview of the Data

In developing the effluent limits for the FGD and gasification wastewater treatment technologies,
EPA used data from the following nine power plants: Allen, Belews Creek, Hatfield's Ferry,
Keystone, Miami Fort, Pleasant Prairie, Brindisi, Polk, and Wabash River. The following bullets
provide a brief overview of the data source available for each plant (see Sections 6, 7, 8, and 9 for a
more detailed description of the data at each of the plants):

¦	At Hatfield's Ferry, Keystone, Miami Fort, and Pleasant Prairie, sampling data were
available for FGD wastewater from three sources: (i) plant self-monitoring (hereafter
referred to as "plant self-monitoring") data; (ii) EPA sampling episodes (hereafter
referred to as "EPA sampling"); and (iii) EPA directed sampling through letters issued
under authority of CWA section 308 (hereafter referred to as "CWA 308 sampling").
The data used were collected at two sampling locations: FGD purge and chemical
precipitation effluent. In addition, We Energies provided self-monitoring data collected
at the secondary clarifier at Pleasant Prairie.

¦	At Allen and Belews Creek, sampling data were available for FGD wastewater from
three sources: (i) plant self-monitoring data for several years of operation provided by
Duke Energy, (ii) EPA sampling, and (iii) CWA 308 sampling. The data were collected
at three sampling locations: FGD purge, bioreactor influent, and bioreactor effluent.

¦	At Brindisi, the sampling data available were from EPA sampling. The data were
collected at three sampling locations: FGD purge, brine concentrator distillate, and
crystallizer condensate.

¦	At Polk and Wabash River, the sampling data were available for gasification wastewater
from CWA 308 sampling. For Polk, the data were collected at four sampling locations:
neutralized weak acid waste stream, vapor compression evaporator influent, vapor

2

-------
compression evaporator condensate, and forced circulation evaporator condensate. For
Wabash River, the sampling data were collected at four sampling locations: sour water
treatment influent, steam stripper effluent, vapor compression evaporator influent, and
vapor compression evaporator condensate.

2.2 Analytical Methods

Tables 1 and 2 summarize the analytical methods used to analyze each pollutant for the plant self-
monitoring data, EPA sampling data, and CWA 308 sampling data.

Table 1. Summary of the analytical lab methods used for each pollutant by plant self-
monitoring

Analytical method

Pollutant

200.7, 200.8, 6020A

Arsenic

1631E, 245.11 245.7, SM3112B2

Mercury

353.2

Nitrate-nitrite as N

200.8

Selenium

1 FGD purge and bioreactor influent only. Data for this method were excluded for the chemical precipitation effluent and the bloreactor
effluent, due to It being an Insufficiently sensitive method.

2 FGD purge only.

Table 2 summarizes the analytical methods used to analyze each pollutant for EPA sampling and
CWA 308 sampling.

Table 2. Summary of the analytical lab methods used for each pollutant by EPA sampling
and CWA 308 sampling

Analytical method

Pollutant

200.8, 6020

Arsenic 1

1631E, FGS - 069

Mercury 2

353.2

Nitrate-nitrite as N

2540C

Total Dissolved Solids

200.8

Selenium

1 Method 200.8 was used at each of the sampling locations except for FGD. At FGD purge, due to the high solids content of the
wastewater In certain samples, a combination of methods 200.8 and 6020 was used.

2 Method 1631E was used at each of the sampling locations except for FGD purge. At FGD purge, due to the high solids content of the
wastewater In certain samples, a combination of methods 1631E and FGS - 069 was used.

-------
2.3

Combining Data from Multiple Sources within a Plant

As described in Section 2.1, data for most of the model plants came from multiple sources such as
EPA sampling, CWA 308, or plant self-monitoring. For five plants (Allen, Belews Creek, Hatfield's
Ferry, Miami Fort, and Pleasant Prairie), the multiple sources of the data were collected during
overlapping time periods, thus, EPA combined these data into a single dataset for each plant. For
one plant (Keystone), the multiple sources of data were collected during non-overlapping time
periods. At Keystone, EPA and CWA 308 samples were collected from September 2010 through
January 2011 and arsenic self-monitoring data was available from January 2012 through April 2014.
Although the data collection periods for Keystone did not overlap, EPA has no information to
indicate that the data represent different operating conditions to the extent that would warrant
treating the non-overlapping periods as separate datasets. Thus, EPA also combined the multiple
sources of data for Keystone into a single dataset for the plant. This approach is consistent with
EPA's traditional approach for other effluent guidelines rulemakings.1 Three plants (Brindisi, Polk,
and Wabash River) had data from a single source, and for these plants it was not necessary to
combine data. For a listing of all the data and their sampling sources for each of the plants, see
DCN SE06277.

1 In some cases where the sampling data from a plant were collected over two or more distinct time periods, EPA may
analyze the data from each time period separately. In some past effluent guidelines rulemakings, EPA analyzed data as
if each time period represents a different plant when the data were considered to represent different operating
conditions due to changes in management, personnel, and procedures. On the other hand, when EPA obtains the data
(such as the EPA sampling and plant self-monitoring data in this rulemaking) from a plant during the same time
period, EPA usually combines the data from these sources into a single dataset for the plant for the statistical analyses.

-------
3 Data Corrections, Exclusions, Baseline
Substitutions, and Aggregation

This section describe the corrections EPA made to the data it obtained, why and how EPA either
excluded or substituted certain data, and how data were aggregated before developing the effluent
limits for each pollutant in each technology option. It should be noted that even though EPA did
not use the arsenic and mercury data from Allen and Belews Creek to calculate the limits for
biological treatment system (for the reason described above and in Section 7), EPA did perform data
corrections, exclusions, baseline substitutions, and aggregation for the arsenic and mercury data at
these two plants.2

3.1 Data Corrections

For the proposed rule, several corrections were made to the plant self-monitoring data at Belews
Creek due to data entry errors. In July 2011, EPA's Engineering and Analysis Division obtained the
plant self-monitoring data for Allen and Belews Creek from Duke Energy. After an initial review of
the data, some inconsistencies were found and EPA subsequently asked Duke Energy to address the
potential errors (an email from Ronald Jordan to Nathan Craig on 07/20/2011 summarized the
apparent inconsistencies). Duke Energy responded to EPA's request in an email from Nathan Craig
to Ronald Jordan (dated 07/21/2011). The data issues thatwere clarified dealtwith data entry errors
in certain detection indicators (see Table 1.1 in Appendix 1). Further, after continued interaction
between Ron Jordan and Nathan Craig, other data entry errors were also corrected (see Table 1.2 in
Appendix 1). Subsequent to EPA publishing the proposed rule, Duke Energy provided additional
plant self-monitoring data for Allen per EPA's request. After review and discussion with Duke
Energy, a correction was made to the December 20th, 2011 total Nitrate/Nitrite data from the
biological treatment effluent (See Table 1.2 in Appendix 1). Additional data entry errors in selenium
detection indicators and sample collection date from the proposed rule data at Belews Creek were
found and corrected (See Tables 1.1 and 1.3 in Appendix 1). Additional data at Keystone provided
by NRG were reviewed and errors in mercury detection indicators of and arsenic concentration were
corrected as shown in Table 1.1 and Table 1.2 in Appendix 1.

2 These arsenic and mercury data for Allen and Belews Creek were also used in the engineering review of the effluent
limits to demonstrate that the biological stage of the FGD wastewater treatment system provides additional removals
of these pollutants following the chemical precipitation treatment stage. See section 11.1.

5

-------
3.2

Data Exclusions and Rationale for the Exclusions

Prior to calculating the effluent limits, EPA thoroughly evaluated the available data for each plant.
EPA reviewed the analytical methods used for each reported observation to determine whether they
were approved in 40 CFR 136 for NPDES purposes and that the methods were sufficiently sensitive
to quantify pollutant concentrations. EPA also reviewed the laboratory reports and other available
information to evaluate whether the laboratory analyses were performed in a manner to minimize
analytical interferences and achieve detection levels low enough to quantify the pollutants present in
the wastewater samples. EPA conducted additional reviews of the data to identify results that
appeared to reflect quality control issues (either associated with sample collection or analysis) or
associated with conditions that do not represent proper operation of BAT or NSPS treatment
technology. Based on these evaluations, EPA identified certain data that warranted exclusion from
the calculations of the limits because:

(i)	samples were analyzed using an analytical method that is not approved in 40 CFR 136
for NPDES purposes;

(ii)	samples were analyzed using a method that was not a sufficiently-sensitive analytical
method (e.g., EPA Method 245.1 for mercury in effluent samples);

(iii)	samples were analyzed in a manner which resulted in an unacceptable level of analytical
interferences;

(iv)	the samples were collected during initial commissioning period for the treatment system
or plant decommissioning period and do not represent BAT/NSPS level of
performance;

(v)	the analytical results were identified as questionable due to quality control issues,
abnormal conditions or treatment upsets, or were analytical anomalies;

(vi)	the samples were collected from a location that is not representative of treated effluent
(e.g., secondary clarifier sample location instead of final effluent sample location); or

(vii)	the treatment system was operating in a manner that does not represent BAT/NSPS
level of performance (including time periods not associated with the initial
commissioning period for the treatment system or the plant decommissioning period).

Appendix 2 of this document provides a listing of each of the data points that were excluded, along
with the reasons for their exclusion. Also see the memorandum "Review of Analytical Methods
Used for Industry Self-Monitoring Data Considered for FGD Wastewater Effluent Limits," USEPA
Office of Water, dated December 8, 2014. This memorandum is in the record as DCN SE06278

6

-------
(this memo will hereafter be referred to as the "Analytical Methods Review Memo" or DCN
SE06278.

3.3 Baseline Substitutions

In general, EPA used detected values or the sample-specific detection limits (i.e., sample-specific
quantitation limit) for non-detected values in calculating the effluent limits. However, there were
some instances in which EPA substituted a baseline value for a detected value or a sample-specific
detection limit (i.e., non-detect) that were lower than the baseline value. Baseline substitution
accounts for the possibility that certain detected or non-detected results in the dataset may be at a
lower concentration than generally can be reliably quantified by well-operated laboratories. This
approach is consistent with the way EPA has calculated effluent limits in previous effluent guidelines
rulemakings and is intended to avoid establishing an effluent limit that could be biased toward a
lower concentration than plants can reliably demonstrate compliance with.3 After excluding all the
necessary data as described in Section 3.2 above, EPA compared each reported result to a baseline
value. Whenever a detected value or sample-specific detection limit was lower than the baseline
value, EPA used the baseline value instead and classified the value as non-detected (even if the
actual reported result was a detected value). For example, if the baseline value was 5 ng/L and the
laboratory reported a detected value of 3 ng/L, EPA's calculations would treat the sample result as
being non-detected with a sample-specific detection limit of 5 ng/L.

Table 3 presents the baseline values that were used for each pollutant for this rulemaking. It should
be noted that in cases when all the concentration values are above the baseline value, then the
baseline value would have no effect on the concentration values and subsequent limits. Effluent data
for mercury and total dissolved solids were all above the baseline values, thus, no baseline
substitution was performed when calculating limitations for these parameters. DCN SE06277

3 For example, if a daily maximum limit were established at a concentration lower than the baseline value, although some
laboratories might be able to achieve sufficiently low quantitation levels, it is possible that typical well-operated
laboratories could not reliably measure down to that level. In such cases, a plant would not be able to demonstrate
compliance with the limit. A similar situation might arise with monthly average limits, particularly if the limit is at a
concentration near the baseline value. EPA does not intend to suggest that the baseline value should be established at a
level that every laboratory in the country can measure to, nor that limits established for the ELGs must be established
sufficiently high that every laboratory in the country must be able to measure to that concentration; however, it is
appropriate to use baseline values that generally can be reliably achieved by well-operated laboratories. This approach
achieves a reasonable balance in establishing limits that are representative of treatment system performance and
protective of the environment, while at the same time ensuring that plants have adequate access to laboratories with
the analytical capabilities necessary to reliably demonstrate compliance with the limits.

7

-------
provides a listing of all the data, the indicator of whether a baseline substitution was performed, and
the value after the baseline substitution was performed.

See Section 13 of Technical Development Document for a detailed discussion of the baseline values
for each pollutant.

Table 3. Summary of the baseline values used for each pollutant

Pollutant

Baseline Value (unit)

Arsenic

2Mg/L

Mercury

0.5 ng/L

Nitrate-nitrite as N

0.05 mg/L

Selenium

5Mg/L

Total Dissolved Solids

10 mg/L

In addition to calculating the limits for each pollutant for each technology option adjusting for the
baseline values shown above, EPA also calculated the effluent limits using all valid reported results
(i.e., without substituting baseline values and/or changing the classification of the result). As noted
above, the reason for using baseline substitution is generally to prevent establishing an effluent that
could be biased toward a lower concentration than facilities can reliably demonstrate compliance
with. When incorporating baseline substitution, the effluent limit is often unchanged but sometimes
results in adjusting the effluent limit upward slightly. For the steam electric ELGs, using baseline
substitution results in such an increase in the limits in a few instances.4 However, EPA found that
using baseline substitution has the opposite effect in certain other instances, having the unexpected
effect of suppressing the limits.5 This downward effect on the effluent limit occurs because,
although the baseline substitution increased the long-term average value slightly, it also reduces the
variability of the dataset and the resulting net effect is a lower effluent limit. This means that using
baseline substitution in such instances would result in a lower effluent limit than EPA would
otherwise calculate directly from the unadjusted dataset. Because EPA wants to ensure that the
effluent limits established by the ELGs can be achieved by facilities, EPA evaluated both the
baseline-adjusted and unadjusted limits for each technology option and used the higher result for the
final ELGs.

4	This occurs for the selenium limit for gasification wastewater and for the selenium limit for FGD wastewater based on
evaporation technology.

5	This occurs for selenium and nitrate-nitrite as N in FGD wastewater (based on biological treatment technology).

8

-------
3.4 Data Aggregation

EPA used daily values in developing the effluent limits. In cases with two or more samples per day,
EPA mathematically aggregated these samples to obtain a single value for that day (the procedure to
aggregate the samples is described in the subsections below). For the sampling data used in this
rulemaking, there are instances when multiple sample results are available for a given day. This
occurred with field duplicates, overlaps between plant self-monitoring and EPA sampling, or
overlaps between plant self-monitoring and CWA 308 sampling.

When aggregating the data, EPA took into account whether each value was detected (D) or non-
detected (ND). Measurements reported as being less than the sample-specific detection limit (or
baseline values, as appropriate) are designated as non-detected (ND) for the purpose of statistical
analyses to calculate the effluent limits. In the tables and data listings in this document and in the
rulemaking record, EPA uses the indicators D and ND denote the censoring type for detected and
non-detected values, respectively.

The subsections below describe each of the different aggregation procedures. They are presented in
the order that the aggregation was performed; i.e., field duplicates were aggregated first, and then
any overlaps between plant self-monitoring and EPA sampling data or CWA 308 sampling were
aggregated.

3.4.1 Aggregation of Field Duplicates

During the EPA sampling episodes, EPA collected field duplicate samples as part of the quality
assurance/quality control activities. Field duplicates are two samples collected for the same sampling
point at approximately the same time. The duplicates are assigned different sample numbers, and
they are flagged as duplicates for a single sampling point at a plant. Because the analytical data from
a duplicate pair are intended to characterize the same conditions at a given time at a single sampling
point, EPA averaged the data to obtain one value for each duplicate pair.

In most cases, both duplicates in a pair had the same censoring type, so the censoring type of the
aggregated value was the same as that of the duplicates. In some instances, one duplicate was a
detected (D) value but the other duplicate was a non-detected (ND) value. When this occurred, EPA
determined that the aggregated value should be treated as detected (D) because the pollutant is

-------
confirmed to be present at a level above the sample-specified detection limit in one of the
duplicates. DCN SE06277 lists the data before the aggregation as well as after the aggregation.

Table 4 below summarizes the procedure for aggregating the sample measurements from the field
duplicates.

Table 4. Aggregation of field duplicates

If the field
duplicates are:

Censoring type
of average is:

Value of the aggregate is:

Formulas for aggregate
values of duplicates

Both detected

Arithmetic average of measured values

(Di + D2)/2

Both non-detected

Arithmetic average of sample-specific
detected limits (or baseline)

(DLi + DL2)/2

One detected and
one non-detected

Arithmetic average of measured value
and sample-specific detection limit (or
baseline)

(D + DL)/2

D: detected.

ND: non-detected.

DL: sample-specific detection limit.

3.4.2 Aggregation of Overlapping Samples

At the Allen, Belews Creek, Hatfield's Ferry, Miami Fort, and Pleasant Prairie plants, sampling data
were available from EPA sampling, CWA 308 sampling, and plant self-monitoring. As explained in
Section 2.3 above, there was some overlap between the data from these sources. On some days at a
given plant, samples were available from two sources, such as plant self-monitoring and either EPA
sampling or CWA 308 sampling. When these overlaps occurred, EPA aggregated the measurements
from the available samples to obtain one value for that day. DCN SE06277 lists the data before the
aggregation as well as after the aggregations.

The procedure averaged the measurements to obtain a single value for that day. When both
measurements had the same censoring type, then the censoring type of the aggregate was the same
as that of the overlapping values. Wien one or more measurements were detected (D), EPA
determined that the appropriate censoring type of the aggregate was detected because the pollutant
is confirmed to be present at a level above the sample-specific detection limit in one of the samples.
The procedure for obtaining the aggregated value and censoring type is similar to the procedure
shown in Table 4 above.

-------
4 Data Editing Criteria

After excluding and aggregating the data, EPA applied data editing criteria on a pollutant-by-
pollutant basis to select the datasets to be used for developing the effluent limits for each technology
option. These criteria are referred to as the long-term average test (or LTA test). EPA often uses the
LTA test to ensure that the pollutants for which limits are being set were present in the influent at
sufficient concentrations to evaluate treatment effectiveness at the plant. For each pollutant for
which EPA calculated a limit, the influent first had to pass a basic requirement that 50% of the
influent measurements for the pollutants had to be detected at any concentration. If the dataset at a
plant passed the basic requirement, then the data had to pass one of the following two criteria to
pass the LTA test:

Criterion 1. At least 50% of the influent measurements in a dataset at a plant were detected at
levels equal to or greater than 10 times the baseline value (shown in Section 3.3).

Criterion 2. At least 50% of the influent measurements in a dataset at a plant were detected at
any concentration and the influent arithmetic average was equal to or greater than 10 times the
baseline value (shown in Section 3.3).

If the dataset at a plant failed the basic requirement, then EPA automatically set both Criteria 1 and
2 to "fail." If the dataset for a plant failed the basic requirement, or passed the basic requirement but
failed both criteria, EPA would exclude the plant's effluent data for that pollutant when calculating
limits. Through the application of the LTA test, EPA ensures that the limits result from treatment of
the wastewater and not simply the absence or substantial dilution of that pollutant in the waste
stream.

After performing the LTA test for the pollutants at each plant that were selected as the basis for the
limits for this rulemaking, all the datasets passed the LTA test except for arsenic and mercury data at
Wabash River. Thus, data for arsenic and mercury at Wabash River were excluded from the
calculation of the limits. See Appendix 3 for the results of the LTA test for each pollutant at each
plant.

-------
5 Statistical Methodology

The sections below provide the following: (i) general definitions and background, (ii) statistical
model selected for the data, (iii) analysis of autocorrelation, (iv) dataset requirement, (v) the
calculation of the effluent limits, and (vi) description of how the limit was set in a special case where
most of the observations were non-detects.

5.1 Definitions and Background

In developing the effluent limits, a statistical procedure is used that involves fitting effluent data to a
distribution and estimating specified upper percentiles of the fitted distribution. The bullets below
describe in detail the important quantities that are typically estimated based on the effluent data.

¦ The long-term average (LTA) is the average concentration that is achieved over a period
of time. The long-term average is the mean of the underlying statistical distribution of
the daily effluent values.

¦ The daily maximum limitation is the highest allowable discharge in any one day. The
daily maximum limitation is the estimate of the 99th percentile (denoted as P99) of the
distribution of the historical daily effluent values.

¦ The daily variability factor is the ratio of the daily maximum limitation to the LTA. This
ratio represents the relationship between the large values and the average treatment
performance level that a well-designed and well-operated treatment system should be
capable of achieving at all times.

¦ The monthly average limitation is the highest allowable average of discharges calculated
from the effluent data collected over a calendar month (or period of time specified in
the permit). The monthly average limitation is the estimate of the 95th percentile
(denoted as P95) of the distribution of the monthly averages of the historical daily
effluent values.

¦ The monthly variability factor is the ratio of the monthly average limitation to the LTA.

5.2 Statistical Model Selected for the Data

In calculating the long-term average, variability factors, and effluent limits, a statistical model is
assumed for the distribution of the effluent data. For this rulemaking, EPA selected the modified

-------
delta-lognormal distribution to model pollutant effluent concentration to develop the long-term
average, variability factors, and limits. A typical effluent dataset from EPA sampling, CWA 308
sampling, or from a plant's self-monitoring consists of a mixture of detected and non-detected
values. The modified delta-lognormal distribution is appropriate for such datasets because it models
the data as a mixture of detected measurements that follow a lognormal distribution and non-detect
measurements that occur with a certain probability. The model also allows for the possibility that
non-detected measurements occur at multiple sample-specific detection limits. Because the data
appeared to fit the modified delta-lognormal model reasonably well, EPA has determined that this
model is appropriate for these data. See Appendix B of the Technical Development Document for
an overview of the statistical model and a description of the procedures EPA used to estimate these
parameters.

5.3 Autocorrelation Analysis

Effluent concentrations that are collected over time may be autocorrelated. The data are positively
autocorrelated when measurements taken at specific time intervals, such as one or two days apart,
are similar. For example, positive autocorrelation would occur if the effluent concentration was
relatively high one day and was likely to remain high on the next and possibly succeeding days.
Because the autocorrelated data affect the true variability of treatment performance, EPA typically
adjusts the variance estimates for the autocorrelated data, when appropriate.

For this rulemaking, whenever there was sufficient data for a pollutant at a plant to evaluate the
autocorrelation reliably, EPA estimated the autocorrelation and incorporated it into the calculation
of the limits. For a plant without enough data to reliably evaluate and obtain a reliable estimate of
the autocorrelation, when there was a correlation of a pollutant available from a similar technology
and waste stream, EPA transferred the autocorrelation estimates from that treatment technology.
Otherwise, EPA set the autocorrelation to zero in the calculation of the effluent limits. EPA did so
because it is reasonable to set the autocorrelation to zero whenever there is not a sufficient amount
of data to reliably evaluate the autocorrelation and when there is no correlation available that can be
transferred from a similar technology and waste stream. See the memorandum titled "Serial
correlations for Steam Electric with and without adjustment for baseline values" from John Rogers
to Ron Jordan for details of the statistical methods and procedures used to determine the
autocorrelation values, as well as a detailed discussion of the minimum number of observations
needed to obtain a reliable estimate of the autocorrelation. This memo is in the record as DCN

-------
SE06279 (this memo will hereafter be referred to as DCN SE06279). The following paragraphs
describe the instances where EPA was able to obtain an estimated autocorrelation and the
assumptions made about the autocorrelation when there were too few observations to estimate the
autocorrelation.

For the biological treatment technology for FGD wastewater (represented by Allen and Belews
Creek plants), EPA was able to perform a statistical evaluation and obtain a reliable estimate of the
autocorrelation for selenium because several years of data were available for these plants. Because of
the similarities between the removal processes for pollutants, EPA determined that it would be
appropriate to also use the value estimated for selenium as the autocorrelation estimate for nitrate-
nitrite as N. The arsenic and mercury limits for the biological technology option are based on (i.e.,
transferred from) the chemical precipitation technology option, thus, the data from the chemical
precipitation plants were used to estimate autocorrelation for arsenic and mercury. (See Section 13
of the Technical Development Document for a detailed discussion of the transfer of the limits.)
Table 5 below lists the autocorrelation values used in the limits calculation for nitrate-nitrite as N
and selenium for the biological treatment option.

For the chemical precipitation treatment option for FGD wastewater (represented by Hatfield's
Ferry, Keystone, Miami Fort, and Pleasant Prairie), EPA was able to perform a statistical evaluation
of the autocorrelation and obtain a reliable estimate of the autocorrelation because several years of
data were available for these plants. Table 5 below lists the autocorrelation values used in the limits
calculation for arsenic and mercury for the chemical precipitation option.

14

-------
Table 5. Summary of estimated autocorrelation values by plant and pollutant

Treatment
technology

Plants representing
the treatment
technology

Pollutant

Baseline4

Correlation value
used for limit
calculation

FGD Biological
treatment

Allen, Belews Creek

Selenium

0

0.69

5

0.67

Nitrate-nitrite as N1

0

0.69

0.05

0.67

FGD Chemical

precipitation

treatment

Hatfield's Ferry,
Keystone, Miami Fort,
Pleasant Prairie

Arsenic2

0

0.86

2

Mercury3

0

0.89

0.5

1: Although there were too few detected values for nitrate-nitrite as N for EPA to obtain a reliable estimate of autocorrelation (12
detected values at Allen; 3 baseline adjusted detected values and 4 non baseline adjusted values at Belews Creek), EPA transferred
the autocorrelation from selenium since these two chemicals behave similarly in the biological treatment system.

2: The correlation was estimated using data from Hatfield's Ferry.

3: The correlation was estimated using data from Hatfield's Ferry, Miami Fort, and Pleasant Prairie.

4: Baseline value of 0 indicates no adjustment for baseline.

For the vapor-compression evaporation treatment technology option for FGD wastewater
(represented by Brindisi), and for the vapor-compression treatment technology option for
gasification wastewater (represented by Polk and Wabash River), EPA was unable to perform an
evaluation and obtain a reliable estimate of the autocorrelation because there were too few
observations available at the plants. Thus, for these plants, EPA set the autocorrelation to zero in
the calculation of the limits. EPA did so because there were not sufficient data to reliably evaluate
the autocorrelation, nor did EPA have a valid correlation estimate available that could be transferred
from a similar technology and waste stream.

5.4 Dataset Requirement

The statistical model requires at least two distinct detected values in order to estimate the variance of
the distribution (i.e., to allow the variability factor to be calculated). Generally, EPA has been
reluctant to estimate a variability factor from plants with a small number of detected observations.
The minimum number of observations required to calculate the variability factor has been evaluated
on a case-by-case basis. For example, for the Organic Chemicals, Plastics and Synthetic Fibers
(OCPSF) rulemaking, a variability factor was estimated for a plant if it had at least 7 daily
observations with at least 3 observations above the detection limit. In the Iron and Steel rulemaking,
a variability factor was estimated for a plant if it had at least three observations, at least two of which

15

-------
were above the detection limit. For the steam electric rulemaking, a variability factor was estimated
for a plant if it had at least three observations, at least two of which were above the detection limit.

5.5 Calculating Effluent Limits for Each Pollutant in Each
Technology Option

EPA calculated the percentiles that are used as a basis for the limits by multiplying the long-term
average (the option long-term average discussed below) by the appropriate variability factors (the
option variability factor discussed below). The paragraphs below describe the calculation of the
option long-term average, option variability factors, and effluent limits. As mentioned in Section 1,
the option long-term averages and variability factors in this document refer to the long-term average
and variability factors for that particular technology option, not the regulatory option described in
Section 8 of the Technical Development Document.

The option long-term average is calculated for a pollutant using two steps. First, EPA calculated the
plant-specific long-term average for each pollutant that had enough distinct detected values by
fitting a statistical model to the daily concentration values. In cases when a dataset for a specific
pollutant did not have enough distinct detected values, the plant-specific long-term average was
calculated as the arithmetic mean of the available daily concentration values. Second, the option
long-term average for each pollutant is calculated as the median of the plant-specific long-term
averages for that pollutant.

The following describes the calculations performed to obtain the option variability factors. First,
EPA calculated the plant-specific variability factors for each pollutant that had enough distinct
detected values by fitting a statistical model to the daily concentration values. The plant-specific daily
variability factor for each pollutant is the estimated 99th percentile of the distribution of the daily
concentration values divided by the plant-specific long-term average. The plant-specific monthly
variability factor for each pollutant is the estimated 95th percentile of the distribution of the 4-day
average concentration divided by the plant-specific long-term average. The calculation of the plant-
specific monthly variability factor assumes that the monthly averages are based on the pollutant
being monitored weekly (approximately four times each month). The option variability factor for
each pollutant is calculated as the mean of the plant-specific long-term averages for that pollutant. In
cases when there were not enough distinct detected values for a specific pollutant at a plant, the data
for the pollutant at the plant was excluded from the calculation of the option variability factors.

-------
Finally, the daily maximum limit for each pollutant for each technology option is the product of the
option long-term average and option daily variability factors. The monthly average limit for each
pollutant for each technology option is the product of the option long-term average and option
monthly variability factors.

5.6 Special Case - Effluent Limits Based on Detection Limits

Although the percentile estimates described in the section above play an important role in
determining daily maximum and monthly average effluent limits, they are not a requirement under
the Clean Water Act and are not always used as the basis for effluent limits. In situations where
there are too few detected results, the statistical models are not appropriate for use in obtaining the
effluent limits since reliable estimates could not be calculated from the model. In such instances,
EPA has instead established the daily effluent limits based on a detection limit. Also, the monthly
average limit is not established when the daily limit is based on the detection limit. The purpose of a
monthly average limit is to require dischargers to provide better control, on a monthly basis, than
required by the daily maximum limit. However, for these pollutants, current analytical methods
cannot measure below the detection limit specified for the daily maximum limit. Thus, even if a
permitting or pretreatment authority requires more frequent monitoring for these pollutants than
once a month, monthly average limits would still be expressed as less than detection limits.

-------
6 Treatment Technology Option for FGD
Wastewater: Chemical Precipitation

EPA used data from four coal-fired power plants (Hatfield's Ferry, Keystone, Miami Fort, and
Pleasant Prairie) to develop the long-term averages, variability factors, and effluent limits for the
chemical precipitation treatment technology option for FGD wastewater. The pollutants for which
the limits are calculated are arsenic and mercury.

The following provides a detailed summary of all the available data for each plant. The available data
from each of the plants came from the following three sources: plant self-monitoring, EPA
sampling, and CWA 308 sampling. As explained in Section 2.3, EPA combined these three sources
of data at each plant into a single dataset for the plant. The data described below include all the
available data prior to any of the exclusions.

Hatfield's Ferry Sampling Data

The sampling data were collected at two sampling locations: FGD Purge and chemical precipitation
effluent. The bullets below provide a brief summary of the data:

¦ Plant self-monitoring data collected by the plant between: 06/30/2009 and 12/10/2013.

¦ EPA sampling data collected by EPA between 12/06/2010 and 12/10/2010.

¦ CWA 308 sampling data collected by the plant one day per month for four consecutive
months. These data were collected on 10/5/2010, 11/10/2010, 1/12/2011, and
02/09/2011.

Keystone Sampling Data

The sampling data were collected at two sampling locations: FGD Purge and chemical precipitation
effluent. The bullets below provide a brief summary of the data:

¦ Plant self-monitoring data collected by the plant between: 01/03/2012 and 04/28/2014.

¦ EPA sampling data collected by EPA between 09/13/2010 and 09/17/2010.

-------
¦ CWA 308 sampling data collected by the plant one day per month for four consecutive
months. These data were collected on 10/7/2010, 11/4/2010, 12/9/2010, and
01/13/2011.

Miami Fort Sampling Data

The sampling data were collected at two sampling locations: FGD Purge and chemical precipitation
effluent. The bullets below provide a brief summary of the data for each of the data sources:

¦ Plant self-monitoring data collected by the plant between: 07/01/2009 and 12/03/2013.

¦ EPA sampling data collected by EPA between 07/12/2010 and 07/16/2010.

¦ CWA 308 sampling data collected by the plant one day per month for four consecutive
months. These data were collected on 09/28/2010, 11/2/2010, 12/7/2010, and
01/14/2011.

Pleasant Prairie Data

The sampling data were collected at three sampling locations: FGD Purge, chemical precipitation
effluent, and secondary clarifier effluent. The bullets below provide a brief summary of the data for
each of the data sources:

¦ Plant self-monitoring data collected by the plant between: 10/04/2007 and 12/18/2013.

¦ EPA sampling data collected by EPA between 06/21/2010 and 06/25/2010.

¦ CWA 308 sampling data collected by the plant one day per month for four consecutive
months. These data were collected on 09/30/2010, 11/3/2010, 12/8/2010, and
01/26/2011.

EPA was able to estimate the autocorrelations for use in the calculation of the limits at Hatfield's
Ferry, Keystone, Miami Fort, and Pleasant Prairie (at the chemical precipitation effluent sampling
location) because several years of data were available for these plants. See DCN SE06279 for the
statistical method and results of the autocorrelation values for each pollutant. Thus, the estimated
autocorrelation values were used in developing limits for each pollutant for this technology option.
In addition, an autocorrelation value of zero (0) was also used to assess the sensitivity of the limits to
the correlation values used. The effluent limits calculated using both approaches are presented
below. The limits for this technology option are based on the calculations that incorporated the
estimated autocorrelations.

-------
The sections below provide the following for each of the pollutants at each of the plants:
longitudinal plots of the data (baseline adjusted data only), summary statistics, plant-specific long-
term average and plant-specific variability factors (see Appendix B of the Technical Development
Document for an overview of the statistical model and the procedures used to estimate the plant-
specific long term average and variability factors). Also provided in the sections below are the option
long-term average, option variability factors, and effluent limits for each pollutant. All exclusions,
baseline substitutions, and aggregation of data were made prior to conducting the analyses presented
below.

6.1 Chemical Precipitation Treatment for FGD Wastewater
Arsenic

6.1.1	Longitudinal Plots of the Data for Arsenic (pg/L)

Below are the longitudinal plots of the arsenic concentrations (on a logarithmic scale) for Hatfield's
Ferry, Keystone, Miami Fort, and Pleasant Prairie.

6.1.2	Summary Statistics for Arsenic (pg/L)

Table 6 provides summary statistics for the numbers of detected and non-detected observations
together with the sample-specific detection limits by sample location and plant. For example, for
Hatfield's Ferry at the FGD Purge sampling location, all 8 observations were detected; at the
chemical precipitation effluent sampling location, of the 130 observations, 129 observations were
detected and 1 was non-detected. The one non-detect had a detection limit of 4 ng/L (shown at the
top of the column).

20

-------
Figure 1.

Plot of arsenic (pg/L) data on a logarithmic scale for Hatfield's Ferry, Keystone,
Miami Fort, and Pleasant Prairie

Hatfields Ferry

5000

500 -

50

Jun 2009

5000

500

50

Sep 2010

5000 -

500

50 -

5 -

Jul 2010

5000

500 -

50 -

Feb 2010

Aug 2011

Jan 2008
~ FGD Purge

° Chemical Precipitation Effluent
* Not Detected

May 2009

Sep 2010

Keystone

Apr 2011

Jul 2012

Miami Fort

Jun 2013

Oct 2010	Jan 2011

Pleasant Prairie

Apr 2011

Sep 2010

Date

Dec 2011

Nov 2011

Apr 2014

Jun 2011

Apr 2013

21

-------
Table 6.	Numbers of detected and non-detected observations and sample-specific detection

limits for arsenic (Mg/L) by plant and sampling location

Plant
Name

Sampling Location

Baseline3

Indicator

(n)4

Sample Specific Detection Limits for
Arsenic (pg/L)

(Total Observations2)

2

4

5

10.656

200

Hatfield's
Ferry1

FGD Purge (N = 8)

0 or 2

D (n=8)











Chemical Precip.
Effluent (N = 130)

0 or 2

D (n=129)











ND (n=l)



1







Keystone1

FGD Purge
(N = 130)

0 or 2

D (n=123)











ND (n=7)









7

Chemical Precip.
Effluent (N = 24)

0

D (n=14)











ND (n=10)

1

7

2





2

D (n=13)











ND (n=ll)

2

7

2





Miami
Fort1

FGD Purge (N = 8)

0 or 2

D (n=8)











Chemical Precip.
Effluent (N = 9)

0 or 2

D (n=9)











Pleasant
Prairie1

FGD Purge
(N = 16)

0 or 2

D (n=15)











ND (n=l)







1



Chemical Precip.
Effluent (N = 20)

0 or 2

D (n=20)











1: Combination of EPA sampling, CWA 308 sampling, and plant self-monitoring data.

2: Detected and non-detected observations combined.

3: Baseline value of 0 indicates no adjustment for baseline.

4: D = detected and ND = non-detected.

Table 7 provides the summary statistics for all observations (detected and non-detected combined)
for FGD Purge and chemical precipitation effluent at each of the plants.

22

-------
Table 7. Summary statistics of arsenic concentration (ng/L) for all detected and
non-detected samples combined, by plant and sampling location

Plant Name

Sampling Location

Baseline2

Summary Statistics for Arsenic ([jg/L)

N3

Minimum

Mean

Median

Maximum

Std

Hatfield's
Ferry1

FGD Purge

0 or 2

8

300.0

1,796.3

1,575.0

4,610.0

1,471.2

Chemical Precip. Effluent

0 or 2

130

3.0

9.1

9.0

22.0

3.1

Keystone1

FGD Purge

0 or 2

130

200.0

1,569.9

1,580.0

5,250.0

962.1

Chemical Precip. Effluent

0

24

1.0

3.1

3.0

5.0

1.2

2

24

2.0

3.1

3.0

5.0

1.1

Miami Fort1

FGD Purge

0 or 2

8

320.0

744.8

729.8

1,330.0

323.5

Chemical Precip. Effluent

0 or 2

9

3.8

4.3

4.3

4.7

0.3

Pleasant
Prairie1

FGD Purge

0 or 2

16

10.7

117.0

145.0

187.0

55.3

Chemical Precip. Effluent

0 or 2

20

3.0

7.6

7.4

12.0

2.4

1: Combination of EPA sampling (4 daily samples), CWA 308 sampling (4 monthly samples), and plant self-monitoring data.
2: Baseline value of 0 indicates no adjustment for baseline.

3: Detected and non-detected observations combined.

6.1.3 Plant-specific Long-term Averages and Variability Factors, Option
Long-term Average and Variability Factors, and Effluent Limits for
Arsenic (Mg/L)

Table 8 provides the plant-specific LTA, plant-specific variability factors, option LTA, option
variability factors, and effluent limits for arsenic in the chemical precipitation effluent.

23

-------
Table 8.	Plant-specific long-term averages and variability factors, option long-term average

and variability factors, and limits for arsenic (Mg/L) in chemical precipitation
effluent

Baseline
Adjusted1

Autocorrelation
Value2

TYPE

N3

Plant
Name

LTA

Daily
Variability
Factor

Monthly
Variability
Factor

Limits

Daily

Monthly

0

0

Plant-
Specific

130

(D=129,
ND=1)

Hatfield's
Ferry4

9.130

2.091

1.315





24

(D=14,
ND=10)

Keystone4

3.059

1.634

1.342





9

(D=9,
ND=0)

Miami
Fort4

4.296

1.196

1.066





20

(D=20,
ND=0)

Pleasant
Prairie4

7.634

2.068

1.307





Option





5.965

1.747

1.258

10.424

7.501

0

0.86

Plant-
Specific

130

(D=129,
ND=1)

Hatfield's
Ferry4

9.135

2.098

1.412





24

(D=14,
ND=10)

Keystone4

3.064

1.632

1.431





9

(D=9,
ND=0)

Miami
Fort4

4.298

1.211

1.092





20

(D=20,
ND=0)

Pleasant
Prairie4

7.655

2.101

1.410





Option





5.976

1.760

1.336

10.521

7.985

2

0

Plant-
Specific

130

(D=129,
ND=1)

Hatfield's
Ferry4

9.130

2.091

1.315





24

(D=13,
ND=11)

Keystone4

3.093

1.616

1.301





9

(D=9,
ND=0)

Miami
Fort4

4.296

1.196

1.066





20

(D=20,
ND=0)

Pleasant
Prairie4

7.634

2.068

1.307





Option





5.965

1.743

1.247

10.397

7.441

24

-------
Table 8.	Plant-specific long-term averages and variability factors, option long-term average

and variability factors, and limits for arsenic (Mg/L) in chemical precipitation
effluent (continued)

Baseline
Adjusted1

Autocorrelation
Value2

TYPE

N3

Plant
Name

LTA

Daily
Variability
Factor

Monthly
Variability
Factor

Limits

Daily

Monthly

2

0.86

Plant-
Specific

130

(D=129,
ND=1)

Hatfield's
Ferry4

9.135

2.098

1.412





24

(D=13,
ND=11)

Keystone4

3.096

1.615

1.383





9

(D=9,
ND=0)

Miami
Fort4

4.298

1.211

1.092





20

(D=20,
ND=0)

Pleasant
Prairie4

7.655

2.101

1.410





Option





5.976

1.756

1.324

10.496

7.914

1: Baseline value of 0 Indicates no adjustment for baseline.

2: Correlation ranges from -1 to 1, with a value of 0 Indicating that EPA assumed no correlation In the data.
3: D = detected and ND = non-detected.

4: Combination of EPA sampling, CWA 308 sampling, and plant self-monltorlng data.

6.2 Chemical Precipitation Treatment for FGD Wastewater:
Mercury

6.2.1 Longitudinal Plots of the Data for Mercury (ng/L)

Below are the longitudinal plots of the mercury concentrations (on a logarithmic scale) for Hatfield's
Ferry, Keystone, and Miami Fort, and Pleasant Prairie.

25

-------
Figure 2.

Plot of mercury (ng/L) data on a logarithmic scale for Hatfield's Ferry, Keystone,
Miami Fort, and Pleasant Prairie

Hatfields Ferry

Keystone

E	Miami Fort

Pleasant Prairie

o Chemical Precipitation Effluent
* Not Detected

Date

26

-------
6.2.2 Summary Statistics for Mercury (ng/L)

Table 9 provides summary statistics for the numbers of detected and non-detected observations
together with the sample-specific detection limits by sample location and plant. For example, for
Hatfield's Ferry at the FGD Purge sampling location, all 7 observations were detected; at the
chemical precipitation effluent sampling location, of the 219 observations, 181 observations were
detected and 38 were non-detected. Of the 38 non-detects, all had a detection limit of 30 ng/L
(shown at the top of the column).

Table 9.	Numbers of detected and non-detected observations and sample-specific detection

limits for mercury (ng/L) by plant and sampling location

Plant
Name

Sampling Location

Baseline3

Indicator

(n)4

Sample Specific Detection Limits for Mercury (ng/L)

(Total Observations2)

30

37

47

60

63

200

100,000

Hatfield's
Ferry1

FGD Purge (N = 7)

0 or 0.5

D (n=8)















Chemical Precip.
Effluent (N = 219)

0 or 0.5

D (n=181)















ND (n=38)

38













Keystone1

FGD Purge (N = 130)

0 or 0.5

D (n=121)















ND (n=9)











1

8

Chemical Precip.
Effluent (N = 8)

0 or 0.5

D (n=8)















Miami
Fort1

FGD Purge (N = 62)

0 or 0.5

D (n=62)















Chemical Precip.
Effluent (N = 68)

0 or 0.5

D (n=68)















Pleasant
Prairie1

FGD Purge (N = 166)

0 or 0.5

D (n=166)















Chemical Precip.
Effluent (N = 375)

0 or 0.5

D (n=365)















ND (n=10)



1

2

5

2





1: Combination of EPA sampling, CWA 308 sampling, and plant self-monitoring data.

2: Detected and non-detected observations combined.

3: Baseline value of 0 indicates no adjustment for baseline.

4: D = detected and ND = non-detected.

Table 10 provides summary statistics for all observations for FGD Purge and chemical precipitation
effluent at each of the plants.

27

-------
Table 10. Summary statistics of mercury concentration (ng/L) for all detected and non-
detected samples by plant and sampling location

Plant
Name

Sampling
Location

Baseline2

Summary Statistics for Mercury (ng/L)

N3

Minimum

Mean

Median

Maximum

Std

Hatfield's
Ferry1

FGD Purge

0 or 0.5

8

184,000.0

465,125.0

467,000.0

789,000.0

216,886.7

Chemical
Precip. Effluent

0 or 0.5

219

30.0

122.2

67.3

909.0

140.0

Keystone1

FGD Purge

0 or 0.5

130

200.0

438,659.2

434,000.0

950,000.0

253,603.8

Chemical
Precip. Effluent

0 or 0.5

8

26.4

62.8

52.6

119.0

35.9

Miami
Fort1

FGD Purge

0 or 0.5

62

22,000.00

286,137.90

277,500.00

1,065,000.00

181,137.86

Chemical
Precip. Effluent

0 or 0.5

68

2.4

168.4

111.4

770.0

155.2

Pleasant
Prairie 1

FGD Purge

0 or 0.5

166

120,000.0

1,382,747.0

1,400,000.0

4,200,000.0

666,941.0

Chemical
Precip. Effluent

0 or 0.5

375

37.0

214.8

170.0

905.0

157.7

1: Combination of EPA sampling, CWA 308 sampling, and plant self-monitoring data.
2: Baseline value of 0 indicates no adjustment for baseline.

3: Detected and non-detected observations combined.

6.2.3 Plant-specific Long-term Averages and Variability Factors, Option
Long-term Average and Variability Factors, and Effluent Limits for
Mercury (ng/L)

Table 11 provides the plant-specific LTA, plant-specific variability factors, option LTA, option
variability factors, and effluent limits for mercury in chemical precipitation effluent.

28

-------
Table 11. Plant-specific long-term average and variability factors, option long-term average
and variability factors, and limits for mercury (ng/L) in chemical precipitation
effluent

Baseline
Adjusted1

Autocorrelation
Value2





Plant
Name



Daily
Variability
Factor

Monthly
Variability
Factor

Limits

TYPE

N3

LTA

Daily

Monthly







219

(D = 181,
ND = 38)

Hatfield's
Ferry4

118.294

5.289

2.035









Plant-

00

II

Q
00

Keystone4

64.260

3.257

1.584





0 or 0.5

0

Specific

68 (D =
68)

Miami
Fort4

199.080

7.007

2.418











375

(D = 365,
ND = 10)

Pleasant
Prairie4

214.161

3.732

1.693









Option





158.687

4.821

1.932

765.069

306.652







219

(D = 181,
ND = 38)

Hatfield's
Ferry4

118.683

5.321

2.352









Plant-

00

II

Q
00

Keystone4

66.878

3.642

1.907





0 or 0.5

0.89

Specific

68 (D =
68)

Miami
Fort4

200.007

7.044

2.732











375

(D = 365,
ND = 10)

Pleasant
Prairie 4

214.609

3.752

1.943









Option





159.345

4.940

2.233

787.119

355.873

1: Baseline value of 0 Indicates no adjustment for baseline.

2: Correlation ranges from -1 to 1, with a value of 0 Indicating that EPA assumed no correlation In the data.

3: D = detected and ND = non-detected.

4: Combination of EPA sampling, CWA 308 sampling, and plant self-monltorlng data.

6.3 Chemical Precipitation Treatment for FGD Wastewater:
Summary of the Option Long-term Averages, Option
Variability Factors, and Effluent Limits

Table 12 provides a summary of the option long-term averages, option variability factors, and
effluent limits for arsenic and mercury.

29

-------
Table 12. Summary of option long-term averages, option variability factors, and effluent limits
for chemical precipitation FGD wastewater treatment technology

Pollutant

Baseline1

Autocorrelation
Value2

Option LTA

Option Daily
Variability
Factor

Option
Monthly
Variability
Factor

Daily
Limit3

Monthly
Average
Limit3

Arsenic

(Ug/L)

0

0.86

5.976

1.760

1.336

11

8

2

5.976

1.756

1.324

11

8

Mercury

(ng/L)

0 or 0.5

0.89

159.345

4.940

2.233

788

356

1: Baseline value of 0 Indicates no adjustment for baseline.

2: Correlation ranges from -1 to 1, with a value of 0 Indicating that EPA assumed no correlation In the data.
3: Effluent limitations have been rounded upward to the next highest Integer.

30

-------
7 Treatment Technology Option for FGD
Wastewater: Biological Treatment

EPA used data from two coal-fired power plants (Allen and Belews Creek) to develop the long-term
averages, variability factors, and effluent limits based on the chemical precipitation followed by
biological treatment technology option for FGD wastewater. Summary statistics for arsenic,
mercury, nitrate-nitrite as N, and selenium are presented for these two plants. The data from these
two plants were used to calculate the limits for nitrate-nitrite as N and selenium. As noted above,
while these two plants operate the biological treatment system used as the basis for the technology
option, neither of these plants includes sulfide precipitation in the upstream chemical precipitation
system. For this reason, EPA is transferring the arsenic and mercury limits calculated based on the
chemical precipitation technology option to biological technology option for FGD wastewater (see
Section 13 of the Technical Development Document for a detailed discussion of the transfer of
limits).

The following provides a detailed summary of the available data for each plant. The available data
from each plant came from the following three sources: (i) plant self-monitoring, (ii) EPA sampling,
and (iii) CWA 308 sampling. As explained in Section 2.3, EPA combined the multiple sources of
data at each plant into a single dataset for the plant. The data described below includes all the
available data prior to any of the exclusions.

A lien Sampling Data

The sampling data were collected at three sampling locations: FGD purge, bioreactor influent, and
bioreactor effluent. The bullets below provide a brief summary of the data:

¦ Plant self-monitoring data collected by the plant over a period of several years. The data
provided were collected between 03/03/2009 and 10/22/2013.

¦ EPA sampling data collected by EPA between 08/02/2010 and 08/06/2010.

¦ CWA 308 sampling data collected by the plant one day per month for four consecutive
months. These data were collected on 10/5/2010, 11/1/2010, 12/6/2010, and
01/12/2011.

-------
Belews Creek Sampling Data

The sampling data were collected at three sampling locations: FGD purge, bioreactor influent, and
bioreactor effluent. The bullets below provide a brief summary of the data:

¦ Plant self-monitoring data collected by the plant over a period of several years. The data
provided were collected between 02/06/2008 and 11/28/2013.

¦ EPA sampling data collected by EPA between 06/07/2010 and 06/11/2011.

¦ CWA 308 sampling data collected by the plant one day per month for four consecutive
months. These data were collected on 10/6/2010, 11/3/2010, 12/8/2010, and
01/17/2011.

EPA was able to estimate the autocorrelations for use in the calculation of the limits at Allen and
Belews Creek (at the bioreactor effluent sampling location) because several years of data were
available for these plants. See DCN SE06279 for the statistical method and results of the
autocorrelation values for each pollutant. Thus, the estimated autocorrelation values were used in
developing limits for each pollutant for this technology option. In addition, an autocorrelation value
of zero (0) was also used to assess the sensitivity of the limits to the correlation values used. The
effluent limits calculated using both approaches are presented below. The limits for this technology
option are based on the calculations that incorporated the estimated autocorrelations.

The sections below provide the following for each of the pollutants at each of the plants:
longitudinal plots of the data (baseline adjusted data only) and summary statistics. Also provided in
the sections below are the plant-specific and option long-term averages, plant-specific and option
variability factors, and effluent limits for nitrate-nitrite as N and selenium. All exclusions, baseline
substitutions, and aggregation of data were made prior to conducting the analyses described below.

7.1 Biological Treatment for FGD Wastewater: Arsenic

Sections below provide the longitudinal plots and summary statistics for arsenic. Because these two
plants do not represent BAT/NSPS level of treatment for arsenic, EPA did not calculate the limits
for arsenic using the Allen and Belews Creek data. Instead, EPA is transferring the arsenic limits
calculated based on the chemical precipitation technology option to the biological technology option
for FGD wastewater (see Section 13 of the Technical Development Document for a detailed
discussion of the transfer of limits).

-------
7.1.1 Longitudinal Plots of the Data for Arsenic (pg/L)

Below are the longitudinal plots of the arsenic concentrations (on a logarithmic scale) for Allen and
Belews Creek. The plot for Allen shows the concentrations from 7/15/2009 to 10/22/2013, while
the plot for Belews Creek shows the concentrations for 6/12/2008 to 11/28/2013. Note that the
data collected in the first several months of operation for the treatment systems at Allen and Belews
Creek were excluded because they represent the initial commissioning period for the treatment
system.

33

-------
Figure 3. Plot of arsenic (pg/L) data on a logarithmic scale for Allen and Belews Creek

Allen

2000 -

2000

Jul 2009

Aug 2010

Sep 2011

Belews Creek

Sep 2012

Oct 2013

Jun 2008	Oct 2009	Mar 2011	Jul 2012	Nov 2013

~ FGD Purge

° Bioreactor Influent

a Bioreactor Effluent

* Not Detected

Date



34

-------
7.1.2 Summary Statistics for Arsenic (|jg/L)

Table 13 provides summary statistics for the numbers of detected and non-detected observations
together with the sample-specific detection limits by sample location and plant.

Table 13. Numbers of detected and non-detected observations and sample-specific detection
limits for arsenic (Mg/L) by plant and sampling location

Plant
Name

Sampling Location

Baseline3

Indicator

(n)4

Sample Specific Detection Limits for Arsenic (|Jg/L



(Total Observations2)

2

3

4

4.5

5

7

10

14.5

52

Allen1

FGD Purge (N = 173)

0 or 2

D (n = 184)



















Bioreactor Influent
(N = 173)

0 or 2

D(n = 2)



















0 or 2

ND

(n = 182)

1



5



18

3

155





Bioreactor Effluent
(N = 170)

0 or 2

D (n = 42)



















ND

(n = 139)





5



119



15





Belews
Creek1

FGD Purge (N = 201)

0 or 2

D (n = 211)



















Bioreactor Influent
(N = 202)

0 or 2

D (n = 15)



















ND

(n = 193)





5

1

22

1

163



1

Bioreactor Effluent
(N = 201)

0 or 2

D (n = 17)



















ND

(n = 194)

17

1

6

1

116



52

1



1: Combination of EPA sampling, CWA 308 sampling, and plant self-monitoring data.

2: Detected and non-detected observations combined.
3: Baseline value of 0 indicates no adjustment for baseline.
4: D = detected and ND = non-detected.

Table 14 provides summary statistics for all observations (detected and non-detected combined) for
arsenic concentrations by sampling location and plant.

35

-------
Table 14. Summary statistics of arsenic concentration (ng/L) for all detected and
non-detected samples combined, by plant and sampling location

Plant
Name

Sampling Location

Baseline2

Summary Statistics for Arsenic (ng/L)

Minimum

Mean

Median

Maximum

Std

Allen1

FGD Purge

0 or 2

184

68.0

230.3

203.0

725.0

104.9

Bioreactor Influent

0 or 2

184

2.0

9.2

10.0

2.0

Bioreactor Effluent

0 or 2

181

2.0

5.9

5.0

20.1

2.5

Belews
Creek1

FGD Purge

0 or 2

211

21.0

324.2

230.0

2820.0

313.6

Bioreactor Influent

0 or 2

208

4.0

9.8

10.0

52.0

3.8

Bioreactor Effluent

0 or 2

211

2.0

7.2

5.0

121.0

8.7

1: Combination of EPA sampling, CWA 308 sampling, and plant self-monitoring data.
2: Baseline value of 0 indicates no adjustment for baseline.

3: Detected and non-detected observations combined.

7.2 Biological Treatment for FGD Wastewater: Mercury

Sections below provide the longitudinal plots and summary statistics for mercury. Because these two
plants do not represent BAT/NSPS level of treatment for mercury, EPA did not calculate the limits
for mercury using the Allen and Belews Creek data. Instead, EPA is transferring the mercury limits
calculated based on the chemical precipitation technology option to the biological technology option
for FGD wastewater (see Section 13 of the Technical Development Document for a detailed
discussion of the transfer of limits).

7.2.1 Longitudinal Plots of the Data for Mercury (ng/L)

Below are the longitudinal plots of the mercury concentrations (on a logarithmic scale) for Allen and
Belews Creek.

7.2.2 Summary Statistics for Mercury (ng/L)

All observations at each of the sampling locations Belews Creek were detected. Table 15 provides
summary statistics for the numbers of detected and non-detected observations together with the
sample-specific detection limits for Allen by sample location.

-------
Figure 4. Plot of mercury (ng/L) data on a logarithmic scale for Allen and Belews Creek

Allen

1000000 H

10000 -

10000 -

Oct 2009

Mar 2011

Jul 2012

Nov 2013

Jun 2008

Aug 2010

Sep 2011

Belews Creek

Sep 2012

Oct 2013

1000000 H

Jul 2009

o FGD Purge
o Bioreactor Influent
£ Bioreactor Effluent

* Not Detected	Date

37

-------
Table 15. Numbers of detected and non-detected observations and sample-specific detection
limits for mercury (ng/L) by plant and sampling location

Plant
Name

Sampling Location

Baseline3

Indicator

(n)4

Sample Specific Detection Limits for Mercury
(ng/L)

(Total
Observations2)

0.15

0.5

1

1000

2500

Allen1

FGD Purge (N =
183)

0 or 0.5

D (n = 183)











Bioreactor Influent
(N = 153)

0

D (n = 146)











ND (n = 7)

1





6



0.5

D (n = 145)











ND (n = 8)



2



6



Bioreactor Effluent
(N = 143)

0 or 0.5

D (n = 143)











Belews
Creek1

FGD Purge (N =
210)

0 or 0.5

D (n = 209)











ND(n=l)









1

Bioreactor Influent
(N = 160)

0 or 0.5

D (n = 156)











ND (n = 4)







4



Bioreactor Effluent
(N = 153)

0 or 0.5

D (n = 152)











ND (n = 1)





1





1: Combination of EPA sampling, CWA 308 sampling, and plant self-monitoring data.

2: Detected and non-detected observations combined.
3: Baseline value of 0 indicates no adjustment for baseline.
4: D = detected and ND = non-detected.

Table 16 provides summary statistics for all observations for mercury concentrations by sampling
location and plant.

38

-------
Table 16. Summary statistics of mercury concentration (ng/L) for all detected and
non-detected samples combined by plant and sampling location

Plant
Name

Sampling
Location

Baseline 2

Summary Statistics for Mercury (ng/L)

N3

Minimum

Mean

Median

Maximum

Std

Allen1

FGD Purge

0 or0.5

183

9,050.0

51,127.6

44,400.0

180,000.0

29,732.4

Bioreactor
Influent

0

153

0.2

7,850.5

450.0

93,100.0

19,709.5

0.5

153

0.5

7,850.5

450.0

93,100.0

19,709.5

Bioreactor
Effluent

0 or0.5

143

1.9

38.8

22.0

234.0

39.8

Belews
Creek1

FGD Purge

0 or0.5

210

1,000.0

199,907.1

198,500.0

697,000.0

91,347.8

Bioreactor
Influent

0 or0.5

160

5.9

3,946.5

136.6

53,300.0

11,709.0

Bioreactor
Effluent

0 or0.5

153

1.0

66.4

9.8

746.0

136.4

1: Combination of EPA sampling, CWA 308 sampling, and plant self-monitoring data.
2: Baseline value of 0 indicates no adjustment for baseline.

3: Detected and non-detected observations combined (all observations were detected).

7.3 Biological Treatment for FGD Wastewater: Nitrate-nitrite as
N

7.3.1 Longitudinal Plots of the Data for Nitrate-nitrite as N (mg/L)

Below are the longitudinal plots of the nitrate-nitrite as N concentrations (on a logarithmic scale) for
Allen and Belews Creek.

39

-------
Figure 5. Plot of nitrate-nitrite as N (mg/L) data on a logarithmic scale for Allen and Belews
Creek

Allen

100

10 -

0.1 -

Aug 2010

May 2011

100

10 -

0.1 -

Mar 2012
Belews Creek

Dec 2012

Oct 2013

Jun 2010

Apr 2011

Mar 2012

Jan 2013

Nov 2013

~ FGD Purge
o Bioreactor Influent
a Bioreactor Effluent
* Not Detected

Date

40

-------
7.3.2 Summary Statistics for Nitrate-nitrite as N (mg/L)

Table 17 provides summary statistics for the numbers of detected and non-detected observations
together with the sample-specific detection limits by sample location and plant.

Table 17. Number of detected, non-detected, and sample-specific detection limits for
nitrate-nitrate as N (mg/L) by plant and sampling location

Plant
Name

Sampling Location

Baseline3

Indicator

(n)4

Sample Specific Detection Limit for
Nitrate-nitrite as N (mg/L)

(Total Observations2)

0.01

0.05

0.1

Allen1

FGD Purge (N = 32)

0 or 0.05

D (n = 32)







Bioreactor Influent
(N = 32)

0

D (n = 31)







ND (n=l)

1





0.05

D (n = 31)







ND (n=l)



1



Bioreactor Effluent
(N = 30)

0

D (n = 12)







ND (n=18)

11



7

0.05

D (n = 12)







ND (n=18)



11

7

Belews
Creek1

FGD Purge (N = 38)

0 or 0.05

D (n = 38)







Bioreactor Influent
(N = 41)

0

D (n = 40)







ND (n=l)

1





0.05

D (n = 40)







ND (n=l)



1



Bioreactor Effluent
(N = 40)

0

D (n = 4)







ND (n = 36)

31



5

0.05

D (n = 3)







ND (n = 37)



32

5

1: Combination of EPA sampling, CWA 308 sampling, and plant self-monitoring data.

2: Detected and non-detected observations combined.
3: Baseline value of 0 indicates no adjustment for baseline.
4: D = detected and ND = non-detected.

Table 18 provides summary statistics for all observations (detected and non-detected combined) for
nitrate-nitrite as N concentrations by sampling location and plant.

41

-------
Table 18. Summary statistics of nitrate-nitrite as N concentration (mg/L) for all detected and
non-detected samples combined by plant and sampling location

Plant
Name

Sampling
Location

Baseline2

Summary Statistics for Nitrate-nitrite as N (mg/L)

N3

Minimum

Mean

Median

Maximum

Std

Allen1

FGD Purge

0 or 0.05

32

2.70

50.53

42.50

130.00

37.99

Bioreactor
Influent

0

32

0.01

39.15

28.00

110.00

35.90

0.05

32

0.05

39.15

28.00

110.00

35.90

Bioreactor
Effluent

0

30

0.01

1.17

0.10

11.00

2.75

0.05

30

0.05

1.18

0.10

11.00

2.74

Belews
Creek1

FGD Purge

0 or 0.05

38

0.55

11.54

13.00

23.00

5.83

Bioreactor
Influent

0

41

0.10

11.90

12.00

21.00

5.86

0.05

41

0.05

11.90

12.00

21.00

5.86

Bioreactor
Effluent

0

40

0.01

0.03

0.01

0.15

0.04

0.05

40

0.05

0.06

0.05

0.15

0.03

1: Combination of EPA sampling, CWA 308 sampling, and plant self-monitoring data.
2: Baseline value of 0 indicates no adjustment for baseline.

3: Detected and non-detected observations combined.

7.3.3 Plant-specific Long-term Averages and Variability Factors, Option
Long-term Averages and Variability Factors, and Effluent Limits for
Nitrate-nitrite as N (mg/L)

Table 19 provides the plant-specific long-term averages, plant-specific variability factors, option
long-term averages, option variability factors, and effluent limits for nitrate-nitrite as N in bioreactor
effluent.

42

-------
Table 19. Lag-1 autocorrelation values, plant-specific long-term averages and variability
factors, option long-term averages and variability factors, and limits for nitrate-
nitrite as N (mg/L) in bioreactor effluent

Baseline
Adjusted1

Autocorrelation
Value2

TYPE

N3

Plant
Name

LTA

Daily
Variability
Factor

Monthly
Variability
Factor

Limits

Daily

Monthly

0

0

Plant-
Specific

30

(D=12,
ND=18)

Allen4

2.032

16.773

3.937





40

(D=4,
ND=36)

Belews
Creek4

0.035

9.360

2.879





Option





1.033

13.066

3.408

13.503

3.522

0

0.69

Plant-
Specific

30

(D=12,
ND=18)

Allen4

2.549

16.876

3.841





40

(D=4,
ND=36)

Belews
Creek4

0.035

9.360

2.892





Option





1.292

13.118

3.366

16.950

4.350

0.05

0

Plant-
Specific

30

(D=12,
ND=18)

Allen4

2.047

16.652

3.932





40

(D=3,
ND=37)

Belews
Creek4

0.063

2.402

1.417





Option





1.055

9.527

2.675

10.049

2.821

0.05

0.67

Plant-
Specific

30

(D=12,
ND=18)

Allen4

2.531

16.779

3.846





40

(D=3,
ND=37)

Belews
Creek4

0.063

2.402

1.454





Option





1.297

9.590

2.650

12.441

3.437

1: Baseline value of 0 Indicates no adjustment for baseline.

2: Correlation ranges from -1 to 1, with a value of 0 Indicating that EPA assumed no correlation In the data.
3: D = detected and ND = non-detected.

4: Combination of EPA sampling, CWA 308 sampling, and plant self-monltorlng data.

43

-------
7.4 Biological Treatment for FGD Wastewater: Selenium
7.4.1 Longitudinal Plots of the Data for Selenium (pg/L)

Below are the longitudinal plots of the selenium concentrations (on a logarithmic scale) for Allen
and Belews Creek.

44

-------
Figure 6.

Plot of selenium (jjg/L) data on a logarithmic scale for Ailen and Belews Creek

Allen

10000 -

1000 -

Jun 2008

Oct 2009

Mar 2011

Jul 2012

Nov 2013

~ FGD Purge
o Bioreactor Influent
A Bioreactor Effluent

* Not Detected	Dst©

Jul 2009

Aug 2010

Sep 2011

Belews Creek

Sep 2012

Oct 2013

10000 -

45

-------
7.4.2 Summary Statistics for Selenium (|jg/L)

Table 20 provides summary statistics for the numbers of detected and non-detected observations
together with the sample-specific detection limits by sample location and plant.

Table 20. Number of detected, non-detected, and sample-specific detection limits for
selenium (Mg/L) by plant and sampling location

Plant
Name

Sampling Location

Baseline3

Indicator

(n)4

Sample Specific Detection Limits for Selenium
(Ug/L)

(Total
Observations2)

1

2

4

4.5

5

10

14.5

15

Allen1

FGD Purge (N = 184)

0 or 5

D (n = 184)

















Bioreactor Influent
(N = 184)

0 or 5

D (n = 184)

















Bioreactor Effluent
(N = 182)

0

D (n = 82)

















ND (n =100)

1

6

5

2

65

21





5

D (n = 66)

















ND (n =116)









95

21





Belews
Creek1

FGD Purge (N =217)

0 or 5

D (n = 217)

















Bioreactor Influent
(N = 216)

0 or 5

D (n = 216)

















Bioreactor Effluent
(N = 216)

0

D (n = 96)

















ND (n = 120)



2

2



67

48

1



5

D (n = 86)

















ND (n = 130)









81

48



1

1: Combination of EPA sampling, CWA 308 sampling, and plant self-monitoring data.

2: Detected and non-detected observations combined.

3: Baseline value of 0 indicates no adjustment for baseline.

4: D = detected and ND = non-detected.

Table 21 provides summary statistics for all observations (detected and non-detected combined) for
selenium by sampling location and plant.

46

-------
Table 21. Summary statistics of selenium concentration (pg/L) for all detected and
non-detected samples combined by plant and sampling location

Plant
Name

Sampling
Location

Baseline2

Summary Statistics for Selenium (pg/L)

N3

Minimum

Mean

Median

Maximum

Std

Allen1

FGD Purge

0 or 5

184

572.0

2089.2

1655.0

10600.0

1484.7

Bioreactor
Influent

0 or 5

184

40.6

528.5

292.0

3980.0

634.2

Bioreactor
Effluent

0

182

1.0

7.1

5.0

27.6

4.3

5

182

5.0

7.4

5.0

27.6

4.1

Belews
Creek1

FGD Purge

0 or 5

217

274.0

5297.2

4820.0

26200.0

2675.9

Bioreactor
Influent

0 or 5

216

14.1

256.1

118.0

2940.0

382.8

Bioreactor
Effluent

0

216

2.0

7.9

6.2

29.4

4.1

5

216

5.0

8.0

6.2

29.4

4.0

1: Combination of EPA sampling, CWA 308 sampling, and plant self-monitoring data.
2: Baseline value of 0 indicates no adjustment for baseline.

3: Detected and non-detected observations combined.

7.4.3 Plant-specific Long-term Averages and Variability Factors, Option
Long-term Averages and Variability Factors, and Effluent Limits for
Selenium (|jg/L)

Table 22 provides the plant-specific long-term averages, plant-specific variability factors, option
long-term averages, option variability factors, and effluent limits for selenium in bioreactor effluent.

47

-------
Table 22. Lag-1 autocorrelation values, plant-specific long-term averages and variability

factors, option long-term averages and variability factors, and limits for selenium
(|jg/L) in bioreactor effluent

Baseline
Adjusted1

Autocorrelation
Value2

TYPE

N3

Plant
Name

LTA

Daily
Variability
Factor

Monthly
Variability
Factor

Limits

Daily

Monthly

0

0

Plant-
Specific

182

(D=82,

ND=100)

Allen4

7.126

3.265

1.556





Plant-
Specific

216

(D=96,

ND=120)

Belews
Creek4

7.920

2.771

1.453





Option





7.523

3.018

1.504

22.703

11.318

0

0.69

Plant-
Specific

182

(D=82,

ND=100)

Allen4

7.134

3.283

1.589





Plant-
Specific

216

(D=96,

ND=120)

Belews
Creek4

7.923

2.779

1.480





Option





7.528

3.031

1.535

22.819

11.554

5

0

Plant-
Specific

182

(D=66,

ND=116)

Allen4

7.386

2.856

1.466





Plant-
Specific

216

(D=86,

ND=130)

Belews
Creek4

8.010

2.610

1.420





Option





7.698

2.733

1.443

21.040

11.110

5

0.67

Plant-
Specific

182

(D=66,

ND=116)

Allen4

7.391

2.873

1.492





Plant-
Specific

216

(D=86,

ND=130)

Belews
Creek4

8.013

2.618

1.442





Option





7.702

2.746

1.467

21.146

11.299

1: Baseline value of 0 Indicates no adjustment for baseline.

2: Correlation ranges from -1 to 1, with a value of 0 Indicating that EPA assumed no correlation In the data.
3: D = detected and ND = non-detected.

4: Combination of EPA sampling, CWA 308 sampling, and plant self-monltorlng data.

48

-------
7.5 Biological Treatment for FGD Wastewater: Summary of the
Option Long-term Averages, Variability Factors, and Effluent
Limits

Table 23 summarizes the option long-term average, option variability factors, and effluent limits for
each pollutant for biological treatment technology option for FGD wastewater. It should be noted
that the option long-term averages and variability factors, and effluent limits presented in previous
tables also included values assuming zero autocorrelation. As described in 5.3, EPA was able to
evaluate and estimate an autocorrelation for selenium based on several years of data from Allen and
Belews Creek. The selenium limit set for this rulemaking therefore are based on the autocorrelation
obtained from the data. Limited data were available for nitrate-nitrite as N from these two plants.
However, since the removal process are similar for selenium and nitrate-nitrite as N, EPA
transferred selenium autocorrelation to nitrate-nitrite as N. It is also appropriate to use zero
autocorrelations when there is not enough data to reliably estimate the autocorrelation. The table
below presents the results for the option long-term averages, option variability factors, and daily
maximum and monthly average limits that incorporated the autocorrelations. As described above,
EPA is transferring the arsenic and mercury limits from the chemical precipitation technology
option to biological technology option. Thus, the table below presents the arsenic and mercury
limits based on the chemical precipitation technology option.

Table 23. Summary of option long-term averages, variability factors, and effluent limits for
biological FGD wastewater treatment technology

Pollutant

Baseline1

Autocorrelation
Value2

Option
LTA

Option Daily
Variability
Factor

Option
Monthly
Variability
Factor

Daily
Limit3

Monthly
Average
Limit3

Arsenic

(Ug/L)

0.86

5.976

1.760

1.336

5.976

1.756

1.324

Mercury

(ng/L)

0 or 0.5

0.89

159.345

4.940

2.233

788

356

Nitrate-
Nitrite as N

(mg/L)

0.69

1.292

13.118

3.366

17.0

4.4

0.05

0.67

1.297

9.590

2.650

12.5

3.5

Selenium

(Ug/L)

0.69

7.528

3.031

1.535

0.67

7.702

2.746

1.467

1: Baseline value of 0 Indicates no adjustment for baseline.

2: Correlation ranges from -1 to 1, with a value of 0 Indicating that EPA assumed no correlation In the data.

3: Effluent limitations have been rounded upward to the next highest Integer for all except nltrate-Nltrlte as N, which have been rounded
upward to the tenth decimal.

-------
8 Treatment Technology Option for FGD

Wastewater: Vapor-compression Evaporation
System

EPA used data from Brindisi power plant to calculate the effluent limits for the chemical
precipitation followed by vapor-compression evaporation treatment technology option for FGD
wastewater. The pollutants for which the effluent limits are calculated are arsenic, mercury, selenium,
and total dissolved solids.

The effluent limits were calculated for the following two sampling locations: (i) brine concentrator
distillate and (ii) crystallizer condensate. The effluent limits for this technology option were based on
the sampling location that produced the higher effluent limits. Setting the effluent limits on the
higher concentration stream is appropriate for this technology to ensure plants operating a well-
designed and well-operated evaporator system can meet the effluent limits prior to discharge or
reuse of the FGD wastewater in another plant process, regardless of whether they sample the
effluent streams separately or as a combined stream (see Section 13 of the Technical Development
Document for a detailed discussion).

Brindisi sampling data were collected by EPA between 04/05/2011 and 04/07/2011. Data were
available at the following sampling locations: FGD Purge, brine concentrator distillate, and
crystallizer condensate.

EPA attempted to use the available data to estimate the autocorrelation. However, for Brindisi, EPA
was not able to perform an evaluation of the autocorrelation because there were too few
observations available at the plant. Thus, EPA set the autocorrelation to zero in developing the
limits for the reasons described in section 5.3.

The sections below provide the following for each of the pollutants: longitudinal plots of the data
(baseline adjusted data only), summary statistics, and plant-specific long-term average and variability
factors (see Appendix B for an overview of the statistical model and the procedures used to estimate
the plant-specific long term average and variability factors). Since only one plant was used, the
option long-term average and option variability factors are the same as the plant-specific long-term
average and variability factors.

-------
8.1 Vapor-compression Evaporation for FGD Wastewater:
Arsenic

8.1.1 Longitudinal Plots of the Data for Arsenic (pg/L)

Below is die longitudinal plot of the arsenic concentrations (on a logarithmic, scale) for Brindisi.

Figure 7. Plot of arsenic (pg/L) data on a iogarithmic scale for Brindisi. The Brine

Concentrator Distillate and Crystallizer Condensate data were jittered slightly to
enhance visibility

Brindisi

O
'c

(L>
(/)

CD

jj

CD
O
(Ji

CT)
O

50

20

10 -

5 -

~ FGD Purge
A Crystallizer Condensate
O Brine Concentrator Distillate
* Not Detected

Apr 05 2011

Apr 06 2011
Date

Apr 07 2011

8.1.2 Summary Statistics for Arsenic (pg/L)

Table 24 provides summary statistics for die numbers of detected and non-detected observations
together with the sample-specific detection limits for each sample location.

51

-------
Table 24. Numbers of detected and non-detected observations and sample-specific detection
limits for arsenic (Mg/L) by sampling location

Plant
Name

Sampling Location

Baseline3

Indicator

(n)4

Sample Specific Detection Limits for
Arsenic (|Jg/L)

(Total Observations2)

Detection Limit = 4

Brindisi1

FGD Purge (N = 2)

0 or 2

D (n=2)



Brine Concentrator Distillate

(N = 3)

0 or 2

D (n=0)



ND (n=3)

3

Crystallizer Condensate
(N = 3)

0 or 2

D (n=0)



ND (n=3)

3

1: EPA sampling data.

2: Detected and non-detected observations combined.

3: Baseline value of 0 indicates no adjustment for baseline.

4: D = detected and ND = non-detected.

Table 25 provides summary statistics for all observations (detected and non-detected combined) for
each of sampling location.

Table 25. Summary statistics of arsenic concentration (Mg/L) for all detected and non-
detected samples combined by sampling location

Plant
Name

Sampling Location

Baseline2

Summary Statistics for Arsenic (jjg/L)

N3

Minimum

Mean

Median

Maximum

Std

Brindisi1

FGD Purge

0 or 2

2

53.0

54.0

54.0

55.0

1.4

Brine Concentrator Distillate

0 or 2

3

4.0

4.0

4.0

4.0

0.0

Crystallizer Condensate

0 or 2

3

4.0

4.0

4.0

4.0

0.0

1: EPA sampling data.

2: Baseline value of 0 indicates no adjustment for baseline.
3: Detected and non-detected observations combined.

8.1.3 Plant-specific Long-term Averages and Variability Factors, Option
Long-term Averages and Variability Factors, and Effluent Limits for
Arsenic (|Jg/L)

Table 26 provides the plant-specific long-term averages and variability factors, option long-term
averages and variability factors, and effluent limits for arsenic at two sampling locations: brine
concentrator distillate and crystallizer condensate.

52

-------
Since all observations were non-detected, the statistical model was not used to obtain the long-term
average and variability factors (see Sections 5.4, 5.5, and 5.6). The daily limits for both sampling
locations were set based on the detection limit. The monthly average limits are not established when
the daily limit is set equal to the detection limit.

Table 26. Plant-specific long-term averages and variability factors, option long-term averages
and variability factors, and limits for arsenic (Mg/L) in brine concentrator distillate
and crystallizer condensate

Baseline
adjusted1

Plant
Name

Sampling
Location

Type

LTA

Daily
Variability
Factor

Monthly
Variability
Factor

Limits

Daily

Monthly

0 or 2

Brindisi

Brine

Concentrator
Distillate

Plant -
Specific

(D=0,
ND=3)

4.03

NA4

Option

4.03

NA4

NA6

0 or 2

Crystallizer
Condensate

Plant -
Specific

(D=0,
ND=3)

4.03

NA4

Option

4.03

NA4

NA6

1: Baseline value of 0 Indicates no adjustment for baseline.

2: D = detected and ND = non-detected.

3: Long-term average is the arithmetic mean since all observations were non-detected (not able to estimate the variance of the
distribution).

4: All observations were non-detected, so the variability factors could not be calculated.

5: Limit is set equal to the detection limit.

6: Monthly average limit is not established when the daily limit is equal to detection limit.

8.2 Vapor-compression Evaporation for FGD Wastewater:
Mercury

8.2.1 Longitudinal Plots of the Data for Mercury (ng/L)

Below is the longitudinal plot of the mercury concentrations (on a logarithmic scale) for Brindisi.

-------
Figure 8.

Plot of mercury (ng/L) data on a logarithmic scale for Brindisi

Brindisi

10000

0)
c

3

o
CD

E

0
CD
O
cn

CD
O

1000 -

100

10 -

n	





¦	-0



~ FGD Purge



A Crystallizer Condensate



O Brine Concentrator Distillate



* Not Detected

A	A		



				



	¦©		—O

Apr 05 2011

Apr 06 201'
Date

Apr 07 2011

8.2.2 Summary Statistics for Mercury (ng/L)

All observations: at each of the sampling locations for Brindisi were detected. Table 27 provides
summary statistics for all observations for each sampling location.

Table 27. Summary statistics of mercury concentration (ng/L) for all detected and
non-detected samples combined by sampling location

Plant
Name

Sampling Location

Baseline2

Summary Statistics for Mercury (ng/L)

N3

Minimum

Mean

Median

Maximum

Std

Brindisi1

FGD Purge

0 or 0.5

2

21,100.0

24,000.0

24,000.0

26,900.0

4,101.2

Brine Concentrator
Distillate

0 or 0.5

3

2.9

3.3

3.1

4.0

0.6

Crystallizer Condensate

0 or 0.5

3

10.9

17.3

20.4

20.7

5.6

1: EPA sampling data.

2: Baseline value of 0 indicates no adjustment for baseline.
3: Detected and non-detected observations combined.

54

-------
8.2.3 Plant-specific Long-term Averages and Variability Factors, Option
Long-term Averages and Variability Factors, and Effluent Limits for
Mercury (ng/L)

Table 28 provides the plant-specific LTAs and plant-specific variability factors, option LTAs and
variability factors, and limits for mercury at two sampling locations: brine concentrator distillate and
crystallizer condensate.

Table 28. Plant-specific long-term averages and variability factors, option long-term averages
and variability factors, and limits for mercury (ng/L) in brine concentrator distillate
and crystallizer condensate

Baseline
adjusted1

Plant
Name

Sampling
Location

Type

N2

LTA

Daily
Variability
Factor

Monthly
Variability
Factor

Limits

Daily

Monthly

0 or 0.5

Brindisi

Brine

Concentrator
Distillate

Plant -
Specific

3

(D=3,
ND=0)

3.348

1.453

1.144





Option



3.348

1.453

1.144

4.865

3.829

0 or 0.5

Crystallizer
Condensate

Plant -
Specific

3

(D=3,
ND=0)

17.788

2.192

1.338





Option



17.788

2.192

1.338

38.989

23.800

1: Baseline value of 0 Indicates no adjustment for baseline.
2: D = detected and ND = non-detected.

8.3 Vapor-compression Evaporation for FGD Wastewater:
Selenium

8.3.1 Longitudinal Plots of the Data for Selenium (|jg/L)

Below is the longitudinal plot of the selenium concentrations (on a logarithmic scale) for Brindisi.

55

-------
Figure 9.

Plot of selenium (jjg/L) data on a logarithmic scale for Brindisi

Brindisi

E

"c
0)


0
03
O
(/)

O)
O

200 -

20 -







~ FGD Purge



A Crystallizer Condensate



O Brine Concentrator Distillate



* Not Detected









Apr 05 2011

Apr 06 201'
Date

Apr 07 2011

8.3.2 Summary Statistics for Selenium (Mg/L)

Table 29 provides summary statistics for the numbers of detected and non-detected observations
together with the sample-specific detection limits by sample location.

56

-------
Table 29. Number of detected, non-detected, and sample-specific detection limits for
selenium (Mg/L) by sampling location

Plant
Name

Sampling Location

Baseline3

Indicator

(n)4

Sample-Specific Detection Limits for Selenium
(Mg/L)

(Total Observations2)

Detection Limit = 4

Detection Limit = 5

Brindisi1

FGD Purge (N = 2)

0 or 5

D (n = 2)





Brine Concentrator
Distillate (N = 3)

0

D (n = 0)





ND (n = 3)

3



5

D (n = 0)





ND (n = 3)



3

Crystal lizer
Condensate (N = 3)

0

D (n = 0)





ND (n = 3)

3



5

D (n = 0)





ND (n = 3)



3

1: EPA sampling data.

2: Detected and non-detected observations combined.

3: Baseline value of 0 indicates no adjustment for baseline.

4: D = detected and ND = non-detected.

Table 30 provides summary statistics for all observations (detected and non-detected combined) for
each sampling location.

Table 30. Summary statistics of selenium concentration (Mg/L) for all detected and
non-detected samples combined by sampling location

Plant
Name

Sampling
Location

Baseline2

Summary Statistics for Selenium (Mg/L)

N3

Minimum

Mean

Median

Maximum

Std

Brindisi1

FGD Purge

0 or 5

2

220.0

255.0

255.0

290.0

49.5

Brine

Concentrator
Distillate

0

3

4.0

4.0

4.0

4.0

0.0

5

3

5.0

5.0

5.0

5.0

0.0

Crystal lizer
Condensate

0

3

4.0

4.0

4.0

4.0

0.0

5

3

5.0

5.0

5.0

5.0

0.0

1: EPA sampling data.

2: Baseline value of 0 indicates no adjustment for baseline.

3: Detected and non-detected observations combined.

57

-------
8.3.3 Plant-specific Long-term Averages and Variability Factors, Option
Long-term Averages and Variability Factors, and Effluent Limits for
Selenium (pg/L)

Table 31 provides the option plant-specific LTAs and variability factors, option LTAs and variability
factors, and effluent limits for selenium at two sampling locations: brine concentrator distillate and
crystallizer condensate.

All observations were non-detected, so the statistical model was not used to obtain the long-term
average and variability factors. The daily limits for both sampling locations were set based on the
detection limit. The monthly average is not set when the daily limit is based on the detection limit.

58

-------
Table 31.

Plant-specific long-term averages and variability factors, option long-term averages
and variability factors, and limits for selenium (Mg/L) in brine concentrator distillate
and crystallizer condensate

Baseline
adjusted1

Plant
Name

Sampling
Location







Daily
Variability
Factor

Monthly
Variability
Factor

Limits

Type

N2

LTA

Daily

Monthly





Brine

Concentrator
Distillate

Plant -
Specific

3

(D=0,
ND=3)

4.03

NA4

NA4





0

Brindisi

Option



4.03

NA4

NA4

45

NA6





















Crystallizer
Condensate

Plant -
Specific

3

(D=0,
ND=3)

4.03

NA4

NA4











Option



4.03

NA4

NA4

45

NA6





Brine

Concentrator
Distillate

Plant -
Specific

3

(D=0,
ND=3)

5.03

NA4

NA4





5

Brindisi

Option



5.03

NA4

NA4

55

NA6





















Crystallizer
Condensate

Plant -
Specific

3

(D=0,
ND=3)

5.03

NA4

NA4











Option



5.03

NA4

NA4

55

NA6

1: Baseline value of 0 Indicates no adjustment for baseline.

2: D = detected and ND = non-detected.

3: Long-term average is the arithmetic mean since all observations were non-detected (not able to estimate the variance of the
distribution).

4: All observations were non-detected, so the variability factors could not be calculated.

5: Limit is set equal to the detection limit.

6: Monthly average limit is not established when the daily limit is equal to detection limit.

8.4 Vapor-compression Evaporation for FGD Wastewater: Total
Dissolved Solids (TDS)

8.4.1 Longitudinal Plots of the Data for TDS (mg/L)

Below is the longitudinal plot of the TDS concentrations (on a logarithmic scale) for Brindisi.

59

-------
Figure 10

Plot of TDS (mg/L) data on a logarithmic scale for Brindisi

Brindisi

10000 -

O)

C/)
Q

cu
o


CD

o

1000 -

100 -

10

~ FGD Purge
A Crystallizer Condensate
O Brine Concentrator Distillate
* Not Detected

G—

A-

Apr 05 2011

Apr 06 201'
Date

Apr 07 2011

8.4.2 Summary Statistics for TDS (mg/L)

Table 32 provides summary statistics for die numbers of detected and non-detected observations
together with the sample-specific detection limits by sample location.

60

-------
Table 32. Number of detected, non-detected, and sample-specific detection limits for total
dissolved solid (mg/L) by sampling location

Plant
Name

Sampling Location

Baseline3

Indicator

(n)2

Sample-Specific Detection Limits for
TDS(ug/L)

(Total Observations1)

Detection Limit = 10

Brindisi1

FGD Purge (N = 2)

0 or 10

D (n = 2)



Brine Concentrator
Distillate (N = 3)

0 or 10

D (n = 2)



ND (n = 1)

1

Crystallizer Condensate
(N = 3)

0 or 10

D (n = 2)



ND (n = 1)

1

1: EPA sampling data.

2: Detected and non-detected observations combined.

3: Baseline value of 0 indicates no adjustment for baseline.

4: D = detected and ND = non-detected.

Table 33 provides summary statistics for all observations (detected and non-detected combined) for
each sampling location.

Table 33. Summary statistics of total dissolved solids concentration (mg/L) for all detected
and non-detected samples combined by sampling location

Plant
Name

Sampling
Location

Baseline2

Summary Statistics for Total Dissolved Solids (mg/L)

N3

Minimum

Mean

Median

Maximum

Std

Brindisi1

FGD Purge

0 or 10

2

13,000.0

14,000.0

14,000.0

15,000.0

1,414.2

Brine

Concentrator
Distillate

0 or 10

3

10.0

11.0

10.0

13.0

1.7

Crystallizer
Condensate

0 or 10

3

10.0

14.0

10.0

22.0

6.9

1: EPA sampling data.

2: Baseline value of 0 indicates no adjustment for baseline.
3: Detected and non-detected observations combined.

8.4.3 Plant-specific Long-term Averages and Variability Factors, Option
Long-term Averages and Variability Factors, and Effluent Limits for
TDS (mg/L)

Table 34 provides the plant-specific LTAs and variability factors, option LTAs and variability
factors, and effluent limits for TDS at two sampling locations: brine concentrator distillate and
crystallizer condensate.

61

-------
Table 34. Plant-specific long-term averages and variability factors, option long-term averages
and variability factors, and limits for TDS (mg/L) in brine concentrator distillate and
crystallizer condensate

Baseline
adjusted1

Plant
Name

Sampling
Location

Type

N2

LTA

Daily
Variability
Factor

Monthly
Variability
Factor

Limits

Daily

Monthly

0 or 10

Brindisi

Brine

Concentrator
Distillate

Plant -
Specific

3

(D=2,
ND=1)

11.066

1.541

1.150





Option



11.066

1.541

1.150

17.054

12.722

0 or 10

Crystallizer
Condensate

Plant -
Specific

3

(D=2,
ND=1)

14.884

3.341

1.572





Option



14.884

3.341

1.572

49.734

23.393

1: Baseline value of 0 Indicates no adjustment for baseline.
2: D = detected and ND = non-detected.

8.5 Vapor-compression Evaporation for FGD Wastewater:
Summary of the Option Long-term Averages, Option
Variability Factors, and Effluent Limits

Table 35 and Table 36 summarize the option long-term average, option variability factors, and limits
for each of the pollutants for the chemical precipitation followed by vapor compression evaporation
system for FGD wastewater. Table 35 provides the results for the brine concentrator distillate
sampling location while Table 36 provides the results at the crystallizer condensate sampling
location. Since only one plant was used, the option LTA and option variability factors are the same
as the plant-specific LTA and variability factors. Note that the effluent limits for regulation for this
technology option are based on the crystallizer condensate. See Section 13 of the Technical
Development Document for a discussion of EPA's reasons for basing the effluent limits for this
technology option on the crystallizer condensate rather than the effluent limits calculated at brine
concentrator distillate.

62

-------
Table 35. Summary of option long-term averages, variability factors, and limits for vapor-
compression evaporation technology for FGD wastewater in brine concentrator
distillate

Pollutant

Baseline1

Autocorrelation
Level2

Option
LTA

Option Daily
Variability
Factor

Option
Monthly
Variability
Factor

Daily
Limit3

Monthly
Average
Limit3

Arsenic

(Ug/L)

0 or 2

0

4.04

NA5

NA5

46

NA7

Mercury

(ng/L)

0 or 0.5

0

3.348

1.453

1.144

5

4

Selenium

(Ug/L)

0

0

4.0 4

NA5

NA5

46

NA7

5

0

5.0 4

NA5

NA5

56

NA7

TDS (mg/L)

0 or 10

0

11.066

1.541

1.150

18

13

1: Baseline value of 0 Indicates no adjustment for baseline.

2: Correlation ranges from -1 to 1, with a value of 0 Indicating that EPA assumed no correlation In the data.

3: Effluent limitations have been rounded upward to the next highest Integer.

4: Long-term average is the arithmetic mean since all observations were non-detected (not able to estimate the variance of the
distribution).

5: All observations were non-detected, so the variability factors could not be calculated.

6: Limit is set equal to the detection limit.

7: Monthly average limit is not established when the daily limit is equal to detection limit.

63

-------
Table 36. Summary of option long-term averages, option variability factors, and limits for
mechanical vapor-compression evaporation technology for FGD wastewater in
crystallizer condensate

Pollutant

Baseline1

Autocorrelation
Value2

Option
LTA

Option Daily
Variability
Factor

Option
Monthly
Variability
Factor

Daily
Limit3

Monthly
Limit3

Arsenic

(Ug/L)

0 or 2

0

4.04

NA5

NA5

46

NA7

Mercury

(ng/L)

0 or 0.5

0

17.788

2.192

1.338

39

24

Selenium

(Ug/L)

0

0

4.04

NA5

NA5

46

NA7

5

0

5.04

NA5

NA5

56

NA7

TDS (mg/L)

0 or 10

0

14.884

3.341

1.572

50

24

1: Baseline value of 0 Indicates no adjustment for baseline.

2: Correlation ranges from -1 to 1, with a value of 0 Indicating that EPA assumed no correlation In the data.

3: Effluent limitations have been rounded upward to the next highest Integer.

4: Long-term average is the arithmetic mean since all observations were non-detected (not able to estimate the variance of the
distribution).

5: All observations were non-detected, so the variability factors could not be calculated.

6: Limit is set equal to the detection limit.

7: Monthly average limit is not established when the daily limit is equal to detection limit.

64

-------
9 Treatment Technology Option for Gasification
Wastewater: Vapor-compression Evaporation

EPA used data from two power plants (Polk and Wabash River) to develop the effluent limits for
the vapor-compression evaporation treatment technology for gasification wastewater. The four
pollutants for which the limits are calculated are arsenic, mercury, selenium, and total dissolved
solids. Arsenic and mercury data at Wabash River failed the data editing criteria, thus, EPA excluded
the arsenic and mercury data from this plant in developing the limits for gasification wastewater
treatment technology. Even though the Wabash River data for arsenic and mercury were not used to
develop the limits, summary statistics and plots for these two pollutants at this plant are also
presented in this section.

For this technology option, EPA considered limits at two sampling locations: (i) forced circulation
evaporator condensate and (ii) vapor compression evaporator condensate. EPA is establishing the
limits based on vapor compression evaporator condensate data since, as discussed at proposal, EPA
determined that the data collected at forced circulation evaporator condensate do not demonstrate
typical removal rates for pollutants generally well-treated by evaporation and therefore are not
adequate to form the basis of the limits. Based on EPA's review of the treatment system, the data
indicate that the evaporator (or at a minimum the forced circulation evaporation stage) at Polk was
operating abnormally and allowing carryover of pollutants to the condensate effluent stream.6 For
this reason, EPA based the limits for this technology option on the limits calculated from the vapor
compression evaporator condensate data (see Section 13 of the Technical Development Document). If
EPA was to calculate limits using the data at the forced circulation evaporator condensate, it would
follow the same methodology used to calculate the limits for vapor compression evaporator
condensate data.

The following provides a detailed summary of the available data for calculating the numeric limits at
each plant.

6 Comments on the proposed effluent guidelines from Tampa Electric Company (the owner/operator of Polk Power
Station) did not dispute EPA's conclusion about abnormal operation of the evaporator, nor EPA's decision not to use
data for the forced circulation evaporation condensate for calculating the effluent limits for gasification wastewater.

65

-------
Polk Sampling Data

Polk sampling data were collected by the plant on the following dates: 10/18/2011, 10/19/2011,
10/26/2011, and 10/27/2011. Data were available at the following sampling locations: neutralized
weak acid, vapor compression evaporator influent, forced circulation evaporator condensate, and
vapor compression evaporator condensate.

Wabash River S'ambling Data

Wabash River sampling data were collected by the plant between 02/21/2011 to 02/24/2011. Data
were available at the following sampling locations: sour water treatment influent, steam stripper
effluent, vapor compression evaporator influent (RCC evaporator influent), and vapor compression
evaporator condensate (RCC evaporator condensate).

EPA attempted to use the available data to estimate the autocorrelation. However, for both of these
plants, EPA was not able to perform an evaluation of the autocorrelation because there were too
few observations available at each plant. Thus, EPA set that the autocorrelation value is zero in the
calculation of the limits as described in Section 5.3.

The sections below provide the following for each of the pollutants: longitudinal plots of the data
(baseline adjusted data only), summary statistics, and plant-specific long-term average and variability
factors (see Appendix B for an overview of the statistical model and the procedures used to estimate
the plant-specific long term average and variability factors). Also provided in the sections below are
the option long-term average, option variability factors, and effluent limits for each pollutant.

9.1 Vapor
-------
Figure 11

Plot of arsenic (pg/L) data on a logarithmic scale for Polk and Wabash River

Polk

100 -

20

CT>

3

o

'c

-------
9.1.2 Summary Statistics for Arsenic (|jg/L)

Table 37 provides summary statistics for the numbers of detected and non-detected observations
together with the sample-specific detection limits by sample location and plant.

Table 37. Numbers of detected and non-detected observations and sample-specific detection
limits for arsenic (Mg/L) by plant and sampling location

Plant
Name

Sampling Location

Baseline2

Indicator

(n)3

Sample Specific Detection Limits for
Arsenic (pg/L)

(Total Observations1)

4

4.1

4.2

Polk

Neu. Weak Acid Waste
Stream (N = 4)

0 or 2

D (n = 4)







Vapor Compr. Evap.
Influent (N = 4)

0 or 2

D (n = 4)







Forced Cir. Evap.
Condensate (N = 4)

0 or 2

D (n = 4)







Vapor Compr. Evap.
Condensate (N = 4)

0 or 2

ND (n = 4)

4





Wabash
River

Sour Water Trt Influent

(N = 4)

0 or 2

ND (n = 4)



4



Steam Stripper Effluent
(N = 4)

0 or 2

ND (n = 4)



2

2

Vapor Compr. Evap.
Influent (N = 4)

0 or 2

D (n = 2)







ND (n = 2)

2





Vapor Compr. Evap.
Condensate (N = 4)

0 or 2

ND (n = 4)

4





1: Detected and non-detected observations combined.
2: Baseline value of 0 indicates no adjustment for baseline.
3: D = detected and ND = non-detected.

Table 38 provides summary statistics for all observations (detected and non-detected combined) for
each sampling location at each of the plants.

68

-------
Table 38. Summary statistics of arsenic concentration (ng/L) for all detected and
non-detected samples combined, by plant and sampling location

Plant
Name





Summary Statistics for Arsenic (pg/L)

Sampling Location

Baseline1

N2

Minimum

Mean

Median

Maximum

Std



Neu. Weak Acid Waste
Stream

0 or 2

4

140.0

160.0

165.0

170.0

14.1

Polk

Vapor Compr. Evap.
Influent

0 or 2

4

220.0

280.0

280.0

340.0

49.7

Forced Cir. Evap.
Condensate

0 or 2

4

29.0

33.0

33.5

36.0

2.9



Vapor Compr. Evap.
Condensate

0 or 2

4

4.0

4.0

4.0

4.0

0.0



Sour Water Trt Influent

0 or 2

4

4.1

4.1

4.1

4.1

0.0

Wabas
h River

Steam Stripper Effluent

0 or 2

4

4.1

4.2

4.2

4.2

0.1

Vapor Compr. Evap.
Influent

0 or 2

4

4.0

4.5

4.5

5.0

0.6



Vapor Compr. Evap.
Condensate

0 or 2

4

4.0

4.0

4.0

4.0

0.0

1: Baseline value of 0 Indicates no adjustment for baseline.
2: Detected and non-detected observations combined.

9.1.3 Plant-specific Long-term Average and Variability Factors, Option
Long-term Average and Variability Factors, and Effluent Limits for
Arsenic (|Jg/L)

Table 39 provides the plant-specific LTA, plant-specific variability factors, option LTA and
variability factors, and numeric limits for arsenic at vapor compression evaporator condensate (using
Polk data only). As described above, the arsenic data for Wabash River failed the data editing
criteria, so, EPA excluded the arsenic dataset from Wabash River in developing the limits for this
technology option.

69

-------
Table 39. Plant-specific long-term average and variability factors, option long-term average

and variability factors, and limits for arsenic (Mg/L) in vapor compression evaporator
condensate

Baseline
Adjusted1

TYPE

Plant
Name

Sampling
Location

N2

LTA

Daily
Variability
Factor

Monthly
Variability
Factor

Limits

Daily

Monthly

0 or 2

Plant-
Specific

Polk

Vapor

Compr. Evap.
Condensate

4

(D=0,
ND=4)

4.003

NA4

NA4





Option



Vapor

Compr. Evap.
Condensate



4.003

NA4

NA4

45

NA6

1: Baseline value of 0 Indicates no adjustment for baseline.

2: D = detected and ND = non-detected.

3: Long-term average is the arithmetic mean since all observations were non-detected (not able to estimate the variance of the
distribution).

4: All observations were non-detected, so the variability factors could not be calculated.

5: Limit is set equal to the detection limit.

6: Monthly average limit is not established when the daily limit is equal to detection limit.

9.2 Vapor-compression Evaporation for Gasification: Mercury
9.2.1 Longitudinal Plots of the Data for Mercury (ng/L)

Below are the longitudinal plots of the mercury concentrations (on a logarithmic scale) for Polk
River.

70

-------
Figure 12

Plot of mercury (ng/L) data on a logarithmic scale for Polk and Wabash River

Polk

10000 -

100 -

D)
c

o

CD

E

0

CD
O
C/)

O)
O

1 -

~ Neutralized Weak Acid Waste Stream
O Vapor Compression Evaporator Influent
o Forced Circulation Evaporator Condensate
A Vapor Compression Evaporator Condensate
* Not Detected

A	A-

Oct 18 2011

Oct 22 2011

Wabash River

Oct 27 2011

10000

100 -

1 -

£ Sour Water Treatment Influent
~ Steam Stripper Effluent
O Vapor Compression Evaporator Influent
° Vapor Compression Evaporator Condensate
* Not Detected

Feb 21 2011

Feb 22 2011

Feb 24 2011

Date

71

-------
9.2.2 Summary Statistics for Mercury (ng/L)

Table 40 provides summary statistics for the numbers of detected and non-detected observations
together with the sample-specific detection limits by sample location and plant.

Table 40. Numbers of detected and non-detected observations and sample-specific detection
limits for mercury (ng/L) by plant and sampling location

Plant
Name

Sampling Location
(Total Observations1)

Baseline2

Indicator

(n)3

Sample Specific Detection Limits for Mercury (ng/L)

0.5

4.95

9.9

Polk

Neu. Weak Acid
Waste Stream (N =
4)

0 or 0.5

D (n = 4)







Vapor Compr. Evap.
Influent (N = 4)

0 or 0.5

D (n = 4)







Forced Cir. Evap.
Condensate (N = 4)

0 or 0.5

D (n = 4)







Vapor Compr. Evap.
Condensate (N = 4)

0 or 0.5

D (n = 4)







Wabash
River

Sour Water Trt
Influent (N = 4)

0 or 0.5

D (n = 3)







ND (n = 1)





1

Steam Stripper
Effluent (N = 4)

0 or 0.5

D (n = 2)







0 or 0.5

ND (n = 2)





2

Vapor Compr. Evap.
Influent (N = 4)

0 or 0.5

ND (n = 4)



1

3

Vapor Compr. Evap.
Condensate (N = 4)

0 or 0.5

ND (n = 4)

4





1: Detected and non-detected observations combined.
2: Baseline value of 0 indicates no adjustment for baseline.
3: D = detected and ND = non-detected.

Table 41 provides summary statistics for all observations (detected and non-detected combined) for
each sampling location at each of the plants.

72

-------
Table 41. Summary statistics of mercury concentration (ng/L) for all detected and
non-detected samples combined, by plant and sampling location

Plant
Name

Sampling Location

Baseline1

Summary Statistics for Mercury (ng/L)

N2

Minimum

Mean

Median

Maximum

Std

Polk

Neu. Weak Acid Waste Stream

0 or 0.5

4

2,030.0

4,392.5

2,270.0

11,000.0

4,408.9

Vapor Compr. Evap. Influent

0 or 0.5

4

17.0

70.4

85.9

92.7

36.2

Forced Cir. Evap. Condensate

0 or 0.5

4

5.0

6.1

5.7

8.0

1.3

Vapor Compr. Evap. Condensate

0 or 0.5

4

0.8

1.1

1.1

1..26

0.2

Wabash
River

Sour Water Trt Influent

0 or 0.5

4

9.9

1,872.3

29.6

7,420.0

3,698.5

Steam Stripper Effluent

0 or 0.5

4

9.9

22.6

17.2

46.3

17.2

Vapor Compr. Evap. Influent

0 or 0.5

4

5.0

8.7

9.9

9.9

2.5

Vapor Compr. Evap. Condensate

0 or 0.5

4

0.5

0.5

0.5

0.5

0.0

1: Baseline value of 0 Indicates no adjustment for baseline.
2: Detected and non-detected observations combined.

9.2.3 Plant-specific Long-term Average and Variability Factors, Option
Long-term Average and Variability Factors, and Effluent Limits for
Mercury (ng/L)

Table 42 provides the plant-specific LTA, plant-specific variability factors, option LTA and
variability factors, and effluent limits for mercury at vapor compression evaporator condensate
(using Polk data only). Because the mercury data at Wabash River failed the data editing criteria,
EPA excluded the data in developing the limits for this technology option. Thus, the mercury limits
are based on Polk data only.

73

-------
Table 42.

Plant-specific long-term average and variability factors, option long-term average
and variability factors, and limits for mercury (ng/L) in vapor compression
evaporator condensate

Baseline
Adjusted1

TYPE

Plant
Name

Sampling
Location

N2

LTA

Daily
Variability
Factor

Monthly
Variability
Factor

Limits

Daily

Monthly

0 or 0.5

Plant-
Specific

Polk

Vapor Compr
Eva p.

Condensate

4

(D=4,
ND=0)

1.075

1.632

1.194





Option



Vapor Compr.
Eva p.

Condensate



1.075

1.632

1.194

1.754

1.283

1: Baseline value of 0 Indicates no adjustment for baseline.
2: D = detected and ND = non-detected.

9.3 Vapor-compression Evaporation for Gasification: Selenium
9.3.1 Longitudinal Plots of the Data for Selenium (Mg/L)

Below are the longitudinal plots of the selenium concentrations (on a logarithmic scale) for Polk and
Wabash River.

74

-------
Figure 13

Plot of selenium (jjg/L) data on a logarithmic scale for Poik and Wabash River

Polk

5000 -

500 -

50 -

5 -

~ Neutralized Weak Acid Waste Stream
o Forced Circulation Evaporator Condensate
O Vapor Compression Evaporator Influent
A Vapor Compression Evaporator Condensate
* Not Detected

T

Oct 18 2011

Oct 22 2011

Oct 27 2011

Wabash River

5000 -

500

O Vapor Compression Evaporator Influent
A Sour Water Treatment Influent
~ Steam Stripper Effluent
o Vapor Compression Evaporator Condensate
* Not Detected

50 -

Feb 21 2011

Feb 22 2011

Feb 24 2011

Date

75

-------
9.3.2 Summary Statistics for Selenium (|jg/L)

Table 43 provides summary statistics for the numbers of detected and non-detected observations
together with the sample-specific detection limits by sample location and plant.

Table 43. Numbers of detected and non-detected observations and sample-specific detection
limits for selenium (Mg/L) by plant and sampling location

Plant
Name

Sampling Location

Baseline2

Indicator3

Sample-Specific Detection
Limits for Selenium (pg/L)

(Total Observations1)

Detection Limit = 5

Polk

Neu. Weak Acid Waste Stream (N = 4)

0 or 5

D (n = 4)



Vapor Compr. Evap. Influent (N = 4)

0 or 5

D (n = 4)



Forced Cir. Evap. Condensate (N = 4)

0 or 5

D (n = 4)



Vapor Compr. Evap. Condensate (N = 4)

0 or 5

D (n = 4)



Wabash
River

Sour Water Trt Influent (N = 4)

0 or 5

D (n = 4)



Steam Stripper Effluent (N = 4)

0 or 5

D (n = 4)



Vapor Compr. Evap. Influent (N = 4)

0 or 5

D (n = 4)



Vapor Compr. Evap. Condensate (N = 4)

0

D (n = 4)



5

D (n = 1)



ND (n = 3)

3

1: Detected and non-detected observations combined.
2: Baseline value of 0 indicates no adjustment for baseline.
3: D = detected and ND = non-detected.

Table 44 provides summary statistics for all observations for each sampling location at each of the
plants.

76

-------
Table 44. Summary statistics of selenium concentration (Mg/L) for all detected and
non-detected samples combined, by plant and sampling location

Plant
Name

Sampling Location

Baseline1

Summary Statistics for Selenium (|Jg/L)

N2

Minimum

Mean

Median

Maximum

Std

Polk

Neu. Weak Acid Waste Stream

0 or 5

4

6,000.0

7,550.0

7,350.0

9,500.0

1,515.5

Vapor Compr. Evap. Influent

0 or 5

4

720.0

1,277.5

1,295.0

1,800.0

551.5

Forced Cir. Evap. Condensate

0 or 5

4

2,500.0

2,675.0

2,650.0

2,900.0

170.8

Vapor Compr. Evap. Condensate

0 or 5

4

140.0

277.5

250.0

470.0

149.8

Wabash
River

Sour Water Trt Influent

0 or 5

4

420.0

480.0

485.0

530.0

46.9

Steam Stripper Effluent

0 or 5

4

49.0

152.3

165.0

230.0

75.5

Vapor Compr. Evap. Influent

0 or 5

4

800.0

920.0

890.0

1,100.0

128.3

Vapor Compr. Evap. Condensate

0

4

4.1

4.5

4.3

5.5

0.7

5

4

5.0

5.1

5.0

5.5

0.3

1: Baseline value of 0 Indicates no adjustment for baseline.

2: Detected and non-detected observations combined (all were detected).

9.3.3 Plant-specific Long-term Averages and Variability Factors, Option
Long-term Average and Variability Factors, and Effluent Limits for
Selenium (|jg/L)

Table 45 provides the plant-specific LTA, plant-specific variability factors, option LTA and
variability factors, and numeric limits for selenium at vapor compression evaporator condensate
(using both Polk and Wabash River data).

77

-------
Table 45. Plant-specific long-term averages and variability factors, option long-term average
and variability factors, and limits for selenium (Mg/L) in vapor compression
evaporator condensate

Baseline
Adjusted1

TYPE

Plant
Name

Sampling
Location

N3

LTA

Daily
Variability
Factor

Monthly
Variability
Factor

Limits

Daily

Monthly

0

Plant-
Specific

Polk

Vapor Compr
Evap.

Condensate

4

(D=4,
ND=0)

288.434

3.083

1.545





Wabash
River

Vapor Compr
Evap.

Condensate

4

(D=4,
ND=0)

4.534

1.360

1.116





Option



Vapor Compr.
Evap.

Condensate2



146.484

2.222

1.331

325.469

194.948

5

Plant-
Specific

Polk

Vapor Compr
Evap.

Condensate

4

(D=4,
ND=0)

288.434

3.083

1.545





Wabash
River

Vapor Compr
Evap.

Condensate

4

(D=l,
ND=3)

5.125

NA4

NA4





Option



Vapor Compr.
Evap.

Condensate2



146.780

3.083

1.545

452.560

226.808

1: Baseline value of 0 Indicates no adjustment for baseline.

2: Polk and Wabash River combined.

3: D = detected and ND = non-detected.

4: Nearly all observations were non-detected, so the variability factors could not be calculated.

9.4 Vapor-compression Evaporation for Gasification: Total
Dissolved Solids (TDS)

9.4.1 Longitudinal Plots of the Data for TDS (mg/L)

Below are the longitudinal plots of the TDS concentrations (on a logarithmic scale) for Polk and
Wabash River.

78

-------
Figure 14. Plot of TDS (mg/L) data on a logarithmic scale for Polk and Wabash River

Polk

O)

cn
Q
h-

0

ro
o

CO

CD
O

5000 -

500 -

50

Q~

~ Neutralized Weak Acid Waste Stream
<> Vapor Compression Evaporator Influent
o Forced Circulation Evaporator Condensate
A Vapor Compression Evaporator Condensate
* Not Detected

-0-

-O

Oct 18 2011

I

Oct 22 2011

Wabash River

Oct 27 2011

5000

500

50 -

O Vapor Compression Evaporator Influent
~ Steam Stripper Effluent
A Sour Water Treatment Influent
o Vapor Compression Evaporator Condensate
* Not Detected



3					

	—A

	



Feb 21 2011

Feb 22 2011

Feb 24 2011

Date

79

-------
9.4.2 Summary Statistics for TDS (mg/L)

Table 46 provides summary statistics for the numbers of detected and non-detected observations
together with the sample-specific detection limits by sample location and plant.

Table 46. Numbers of detected and non-detected observations and sample-specific detection
limits for TDS (mg/L) by plant and sampling location

Plant
Name

Sampling Location

Baseline2

Indicator

(n)3

Sample-Specific
Detection Limits for
TDS (mg/L)

(Total Observations1)

10

200

Polk

Neu. Weak Acid Waste Stream (N = 4)

0 or 10

D (n = 4)





Vapor Compr. Evap. Influent (N = 4)

0 or 10

D (n = 4)





Forced Cir. Evap. Condensate (N = 4)

0 or 10

D (n = 4)





Vapor Compr. Evap. Condensate (N = 4)

0 or 10

D (n = 4)





Wabash
River

Sour Water Trt Influent (N = 4)

0 or 10

D (n = 2)





ND (n = 2)



2

Steam Stripper Effluent (N = 4)

0 or 10

D (n = 4)





Vapor Compr. Evap. Influent (N = 4)

0 or 10

D (n = 4)





Vapor Compr. Evap. Condensate (N = 4)

0 or 10

D (n = 2)





ND (n = 2)

2



1: Detected and non-detected observations combined.

2: Baseline value of 0 indicates no adjustment for baseline.

3: D =detected and ND = non-detected.

Table 47 provides summary statistics for all observations (detected and non-detected combined) for
each sampling location at each of the plants.

80

-------
Table 47. Summary statistics of TDS concentration (mg/L) for all detected and non-detected
samples combined, by plant and sampling location

Plant
Name

Sampling Location

Baseline1

Summary Statistics for TDS (mg/L)

N2

Minimum

Mean

Median

Maximum

Std

Polk

Neu. Weak Acid Waste Stream

0 or 10

4

26,000.0

27,250.0

26,000.0

31,000.0

2,500.0

Vapor Compr. Evap. Influent

0 or 10

4

4,500.0

4,575.0

4,600.0

4,600.0

50.0

Forced Cir. Evap. Condensate

0 or 10

4

48.0

106.5

114.0

150.0

44.5

Vapor Compr. Evap. Condensate

0 or 10

4

11.0

16.3

15.5

23.0

5.7

Wabash
River

Sour Water Trt Influent

0 or 10

4

200.0

245.0

210.0

360.0

77.2

Steam Stripper Effluent

0 or 10

4

2,000.0

2,800.0

2,400.0

4,400.0

1,095.5

Vapor Compr. Evap. Influent

0 or 10

4

3,600.0

4,225.0

4,400.0

4,500.0

419.3

Vapor Compr. Evap. Condensate

0 or 10

4

10.0

13.5

11.0

22.0

5.7

1: Baseline value of 0 Indicates no adjustment for baseline.
2: Detected and non-detected observations combined.

9.4.3 Plant-specific Long-term Averages and Variability Factors, Option
Long-term Average and Variability Factors, and Effluent Limits for
TDS (mg/L)

Table 48 provides the plant-specific LTA, plant-specific variability factors, option LTA and
variability factors, and numeric limits for TDS at vapor compression evaporator condensate (using
both Polk and Wabash River data).

81

-------
Table 48. Plant-specific long-term averages and variability factors, option long-term average
and variability factors, and numeric limits for TDS (mg/L) in vapor compression
evaporator condensate

Baseline
Adjusted1

TYPE

Plant
Name

Sampling
Location

N3

LTA

Daily
Variability
Factor

Monthly
Variability
Factor

Limits

Daily

Monthly

0 or 10

Plant-
Specific

Polk

Vapor Compr
Evap.

Condensate

4

(D=4,
ND=0)

16.512

2.149

1.327





Wabash
River

Vapor Compr.
Evap.

Condensate

4

(D=2,
ND=2)

13.906

2.818

1.450





Option



Vapor Compr.
Evap.

Condensate2



15.209

2.483

1.389

37.767

21.122

1: Baseline value of 0 Indicates no adjustment for baseline.

2: Polk and Wabash River combined.
3: D = detected and ND = non-detected.

9.5 Vapor-compression Evaporation for Gasification: Summary
of the Option Long-term Average, Option Variability Factors,
and Effluent Limits

Table 49 summarizes the option long-term averages, option variability factors, and effluent limits for
vapor compression-evaporation technology option. Note that EPA is deciding that the daily and
monthly average limitations for mercury for gasification wastewater be rounded to two decimal
places instead of rounding to the next highest integer in order to avoid having the same value for the
daily and monthly average limitation.

82

-------
Table 49. Summary of option long-term averages, option variability factors, and limits for
vapor-compression evaporation technology for gasification wastewater

Pollutant

Baseline1

Autocorrelation
Value2

Option
LTA

Option
Daily
Variability
Factor

Option
Monthly
Variability
Factor

Daily
Limit3

Monthly
Limit3

Arsenic

(Ug/L)

Oor 2

0

4.04

NA5

NA5

4 6

NA7

Mercury

(ng/L)

0 or 0.5

0

1.075

1.632

1.194

1.8

1.3

Selenium

(Ug/L)

0

0

146.484

2.222

1.331

326

195

5

0

146.780

3.083

1.545

453

227

TDS (mg/L)

Oor 10

0

15.209

2.483

1.389

38

22

1: Baseline value of 0 Indicates no adjustment for baseline.

2: Correlation ranges from -1 to 1, with a value of 0 Indicating that EPA assumed no correlation In the data.

3: Effluent limitations have been rounded upward to the next highest integer, except for limits for mercury, which were rounded to the
next highest tenth decimal place.

4: Long-term average is the arithmetic mean since all observations were non-detected (not able to estimate the variance of the
distribution).

5: All observations were non-detected, so the variability factors could not be calculated.

6: Limit is set equal to the detection limit.

7: Monthly average limit is not established when the daily limit is equal to detection limit.

83

-------
10 Summary of Effluent Limits for FGD

Wastewater, Gasification Wastewater, and
Combustion Residual Leachate

Sections 6, 7, 8, and 9 above present detailed summary statistics of the data together with the long-
term averages, variability factors, and effluent limits for each treatment technology option for FGD
and gasification wastewaters and combustion residual leachate. This section provides an overall
summary of the option long-term averages, option variability factors, and limits for those technology
options selected as the basis for the limits in the final rule. In addition, this section also summarizes
the regulation for leachate.

The bullets below provide some important items that are discussed both in previous sections and
again in this section.

¦	For BAT for FGD wastewater, EPA is transferring the effluent limits for arsenic and
mercury calculated from the chemical precipitation technology option to the biological
technology option (see Section 13 of the TDD for the rationale for this transfer of
limits).

¦	For NSPS for FGD wastewater (and BAT for the voluntary incentive program), which
is based on the chemical precipitation followed by vapor-compression evaporation
treatment technology option, the effluent limits are based on the data from crystallizer
condensate sampling location. As explained in Section 8 above, EPA calculated limits
for this technology option at two separate sampling locations: (i) brine concentrator
distillate and (ii) crystallizer condensate. The limits selected for the rulemaking are based
on the stream (i.e., crystallizer condensate) with the higher pollutant concentrations.

¦	For NSPS for combustion residual leachate, which is based on the chemical
precipitation technology option, EPA is transferring the effluent limits from the
chemical precipitation technology option for FGD wastewater (see Section 13 of the
TDD for more information).

¦	In most cases, the limits were rounded upward to the next highest integer. Gasification
wastewater limits for mercury, and FGD wastewater limits for nitrate-nitrite as N, were
rounded to the nearest tenth decimal place.

Tables 50 and 51 provides the option long-term average, option variability factors, and limitations
for each of the FGD, gasification, and combustion residual leachate technology options selected as
the basis for the final rule.

84

-------
Table 50. BAT/PSES limits for existing sources: Summary of the option long-term averages,
variability factors, and limits for FGD and gasification wastewater

Treatment
Technology Option

Pollutant

Baseline

Option
LTA

Daily
Variability
Factor

Monthly
Variability
Factor

Daily
Limitation4

Monthly
Limitation4

FGD wastewater
(Based on
Chemical
Precipitation and
Biological
Treatment)

Arsenic

(Mg/L)1

0

5.976

1.760

1.336

11

8

Mercury

(ng/L)1

0 or 0.5

159.345

4.940

2.233

788

356

Nitrate-
nitrite

(mg/L)

0

1.292

13.118

3.366

17.0

4.4

Selenium

(Ug/L)

0

7.528

3.031

1.535

23

12

















Voluntary Incentive
Program BAT
Limits for FGD
wastewater
(Based on
Chemical
Precipitation and
Evaporation)

Arsenic

(Ug/L)

0 or 2

4.02

NA3

NA3

45

NA6

Mercury

(ng/L)

0 or 0.5

17.788

2.192

1.338

39

24

Selenium

(Ug/L)

5

5.02

NA3

NA3

5s

NA6

TDS

(mg/L)

0 orlO

14.884

3.341

1.572

50

24

















Gasification
wastewater (Based
on Vapor-
Compression
Evaporation for
Gasification

Arsenic

(Ug/L)

0 or 2

4.02

NA3

NA3

45

NA6

Mercury

(ng/L)

0 or 0.5

1.075

1.632

1.194

1.8

1.3

Selenium

(Ug/L)

5

146.78

3.083

1.545

453

227

TDS

(mg/L)

0 or 10

15.209

2.483

1.389

38

22

1: Option LTA, variability factors, and effluent limits were transferred from chemical precipitation technology option for FGD wastewater.
See sections 6 and 7 of this report and section 13 of the Technical Development Document.

2: Long-term average is the arithmetic mean since all observations were non-detected.

3: All observations were non-detected, so variability factors could not be calculated.

4: Effluent limits have been rounded upward to the next highest integer, except for effluent limits for nitrate-Nitrite as N based on
chemical precipitation and biological treatment technology option for FGD wastewater and mercury based on the vapor-compression
evaporation treatment technology option for gasification wastewater, which have been rounded upward to the tenth decimal.

5: Limit is set equal to the detection (quantitation) limit.

6: Monthly average limits are not established when the daily maximum limitation is based on the detection limit.

85

-------
Table 51. NSPS/PSNS limits for new sources: Summary of the option long-term averages,
variability factors, and limits for FGD wastewater, gasification wastewater, and
combustion residual leachate

Treatment
Technology
Option

Pollutant

Baseline

Option
LTA

Daily
Variability
Factor

Monthly
Variability
Factor

Daily
Limitation 4

Monthly
Limitation 4

FGD

wastewater
(Based on
Chemical
Precipitation
and

Evaporation)

Arsenic

(Ug/L)

0 or 2

4.02

NA3

NA3

45

NA6

Mercury

(ng/L)

0 or 0.5

17.788

2.192

1.338

39

24

Selenium

(Ug/L)

5

5.02

NA3

NA3

5s

NA6

TDS

(mg/L)

0 orlO

14.884

3.341

1.572

50

24

















Gasification

wastewater

(Based on

Vapor-

Compression

Evaporation)

Arsenic

(Ug/L)

0 or 2

4.0 2

NA3

NA3

45

NA6

Mercury

(ng/L)

0 or 0.5

1.075

1.632

1.194

1.8

1.3

Selenium

(Ug/L)

5

146.78

3.083

1.545

453

227

TDS

(mg/L)

0 or 10

15.209

2.483

1.389

38

22

















Leachate
(Based on
Chemical
Precipitation)

Arsenic

(Ug/L)1

0

5.976

1.760

1.336

11

8

Mercury

(ng/L)i

0 or 0.5

159.345

4.940

2.233

788

356

1: Option LTA, option variability factors, and effluent limits were transferred from chemical precipitation technology option for FGD
wastewater. See section 6 of this report and section 13 of the Technical Development Document.

2: Long-term average is the arithmetic mean since all observations were non-detected.

3: All observations were non-detected, so variability factors could not be calculated.

4: Effluent limits have been rounded upward to the next highest integer, except for effluent limits for mercury based on the vapor-
compression evaporation treatment technology option for gasification wastewater which have been rounded up to the next highest
tenth decimal place.

5: Limit is set equal to the detection (quantitation) limit.

6: Monthly average limits are not established when the daily maximum limitation is based on the detection limit.

86

-------
11 Engineering Review of the Effluent Limits

Plants that install treatment technologies to comply with the newly promulgated limitations will need
to design and operate the systems to meet the limitations at all times. In summary, this means:

1.	A treatment system that includes the necessary process equipment and chemical
additives that is sized to accommodate the wastewater flows, and that is designed to
target removing the regulated pollutants to meet the long-term average; and

2.	Proper monitoring and operation that targets chemical addition rates and other
operational conditions to the long-term average for the regulated pollutants, considers
fluctuations in influent wastewater flows and pollutant concentrations, and that
proactively monitors for and responds to fluctuations in effluent pollutant
concentrations due to abnormal conditions or treatment system upsets.

A properly designed and operated treatment system that represents best available technology or best
available demonstrated control technology includes characteristics such as proper chemical usage,
periodic inspection and repair of equipment, use of appropriate redundant equipment such as
backup pumps, sufficient staffing by trained operators, communications and coordination among
production and wastewater treatment personnel, close attention to treatment system operating
parameters and effluent quality. Properly designed and operated systems recognize and correct
periods of degraded or abnormal operation.

Proper design does not include inappropriately designed or inadequately sized treatment facilities,
such as systems targeted to meet the limitations themselves rather than the long-term averages. For
example, treatment systems that lack sufficient equalization tank capacity to mitigate fluctuations in
wastewater flow rates or pollutant concentrations. Proper design does not include treatment systems
that do not include key process equipment or chemical additives necessary to achieve effluent limits
such as the organosulfldes used to enhance precipitation of dissolved mercury.

As part of its review of the final limitations, EPA carefully considers the data from the model plants
to see if the data demonstrate that the plants can comply with the final limitations. It is not unusual
for EPA to find that one or all of the model plants may need to make treatment technology
upgrades or improvements to their operation in order to comply with the final limitations. Although
most observations in the datasets used to calculate the effluent limitations are below the limitations,
some observations typically will exist above the limitations. This is reasonable in datasets used to
calculate effluent limitations and does not mean that the calculated limits cannot be met. In such

87

-------
cases, there are specific steps that plants can take that would enable them to improve treatment
system performance so that effluent concentrations would be in compliance with the limitations at
all times. Although EPA selects model plants as representing the best available technology or best
available demonstrated control technology, and they provide the best available data for establishing
limitations that reflect BAT/NSPS level of treatment, it does not necessarily mean that the plants
have the systems fully optimized. For example, the NPDES permit for a model plant may not
include limitations for the regulated pollutant or the NPDES permit limitations may be well above
the final limitations and what other systems are achieving. Thus, these plants would be expected to
have some observations above the limitations which do not reflect BAT/NSPS level of control, not
because the systems are incapable of meeting the limitations but rather because the existing permit
limitations do not drive the plants to optimize the performance of the treatment system and increase
the pollutant removals. If EPA's review demonstrates that a model plant is not consistently
achieving the final limitations, EPA looks at the treatment system design and operation for the
model plant to determine if it currently meets EPA's expectations for proper design and operation.7
In this way, EPA confirms that the final limitations are reasonable and will be achieved by properly
designed and operated systems.

For this final rule, EPA performed an engineering review to verify that the effluent limits are
reasonable based upon the design and expected operation of the control technologies. As part of
this review, EPA performed two types of comparisons. First, EPA compared the effluent limits for
each pollutant against the effluent data from the model plants used to develop the limits. This type
of comparison helps to evaluate how reasonable the limits are from an engineering perspective.
Second, EPA compared the limits for each pollutant to the influent data at the model plants. This
second comparison helps evaluate whether the influent concentrations were generally well-
controlled by the treatment system.

Section 11.1 presents the results of the comparisons between the limits and all effluent data that
were used to calculate the limits for each technology option. Section 11.2 presents the results of the
comparisons between the effluent limits and the influent data values for each technology option. See
Appendix 7 for a listing of all daily effluent values that are above the daily limits. Appendix 7 also
presents a comparison of the effluent values to the monthly average limits, for those periods where
there were sufficient data to represent weekly monitoring, showing those periods where the average
of the daily values for the month were above the monthly average limit. Plots comparing the effluent
data to the daily maximum and monthly average limits are also presented in Appendix 7.

7 If they do not, EPA includes costs for the model plants to do so.

88

-------
11.1 Comparison of the Limits to Effluent Data Used as the Basis
for the Limits

First, EPA compared the daily effluent concentrations to the daily limits to identify any observations
that were above the daily limit. The plots prepared for this first comparison also provide insight on
how other data (i.e., daily values below the daily limit) compare to the limit. Next, EPA compared
the daily concentrations to the monthly average limits, for those periods where there are sufficient
data to represent weekly monitoring,8 and identified those months where the average of all daily
values for the month is above the monthly limit. As was the case for the comparison to the daily
limits, the prepared plots also provide insight to how monthly averages below the monthly limit
compare to the limit.

After thoroughly evaluating the results of the comparison between the limits and the effluent values
used to calculate the limits (see details below), EPA determined that the statistical distributional
assumptions used to develop the limits are appropriate for the data (that is, they provide a
reasonable "fit" to the actual effluent data) and the limits for each wastestream are reasonable and
achievable. (This conclusion is also true for the leachate limits based on the chemical precipitation
technology since the leachate limits were transferred from the FGD wastewater technology option.)
If a plant properly designs and operates its wastewater treatment system to achieve the long-term
average for the model technology (rather than targeting performance at the effluent limits
themselves), it will be able to comply with the limits.

EPA methodology for establishing effluent limits based on certain percentiles of the statistical
distributions, as well as the presentation of the analyses described below in section 11.1, may give
the impression that EPA expects occasional exceedances of the limitations. This conclusion is
incorrect. EPA promulgates limitations that facilities are capable of complying with at all times by
properly operating and maintaining their treatment technologies. These limitations are based upon
statistical modeling of the data and engineering review of the limitations and data.

8 This approach is consistent with the manner in which EPA calculated the limits and anticipates plants will monitor for
compliance with NPDES permits. It is also consistent with the monitoring frequency EPA has generally observed in
NPDES permits, for those pollutants for which the permit includes effluent limits. Additionally, it is consistent with
EPA's methodology for estimating compliance costs for the final rule, which includes estimated O&M costs for weekly
compliance monitoring. Furthermore, it is a reasonable approach for conducting the engineering review because an
assessment that uses data from less frequent monitoring may more closely reflect the daily variability than the monthly
variability and therefore would not accurately reflect whether the monthly limit would have been met. For example,
comparing with the limits calculated for arsenic, the 4 weekly samples collected from Flatfield's Ferry in February 2010
are all equal to or below the daily limit of 11 ng/L. One of these observations (11 |ig/L) is higher than the monthly
limit of 8 |^g/L. Ffowever the average of all daily values for the month is 7.5 ng/L, which is below the monthly limit.

89

-------
Statistical methodology is used as a framework to establish limitations based on percentiles of the
effluent data. Statistical methods provide a logical and consistent framework for analyzing a set of
effluent data and determining values from the data that form a reasonable basis for effluent limits. In
conjunction with the statistical methods, EPA's engineering review verifies that the limits are
reasonable based upon the design and expected operation of the control technologies and the facility
process conditions. As part of that review, EPA examines the range of performance by the facility
data sets used to calculate the limits. The facility data sets represent operation of the best
available/demonstrated technology. However, although these facilities were operating the best
available (or best demonstrated) technology, in some cases these data sets, or periods of time within
a data set, may not necessarily represent the optimized performance of the technology. As described
in Section 3.2 and Appendix 2, EPA excluded certain data from the data sets used to calculate the
effluent limits. At the same time, however, data used by EPA to calculate effluent limits still retain
some other data that might reflect less than optimal performance. By retaining these data in
developing the limits, EPA has chosen a more conservative approach because these data help to
fully characterize the variability in treatment system effluent.

To the extent that a facility's data indicated periods of less than optimal performance or the need for
changes to facilitate targeting effluent performance toward the long-term average, EPA evaluated
the degree to which the facility could upgrade its design, operating, and maintenance conditions to
improve effluent performance as necessary to meet the limits at all times, and included costs for
such upgrades (e.g., additional labor or chemicals) in its estimated costs for the rulemaking. As a
result of the combined statistical modeling and engineering review used to establish the limits, EPA
expects that facilities will be able to design and operate wastewater treatment systems in a manner
that ensures compliance with the limitations. EPA does not expect facilities to violate the limitations
at some pre-set rate merely because probability models are used to develop limitations.

EPA concludes that all facilities properly operating and maintaining the appropriate technology will
be capable of complying with the limitations, even though some values in the data used to develop
the limitations are higher than the limitations, for the following reasons. EPA included data from
facilities using the BAT/NSPS technology that in most cases do not have effluent limits for the
regulated pollutants in their NPDES permits. EPA reviewed the data and other information in the
record and determined that observations exceeding the limitations identified during the engineering
review were a result of quality-control problems or upsets resulting from "loosely" controlled
performance due to very high NPDES permit limits (when the permit includes a limit for the
pollutant) or that reflect operation of systems that optimize the removal of other pollutants (such as
TSS) but not the pollutants regulated by the limitations established by this rule.

90

-------
Tables A7.1 through A7.8 in Appendix 7 list the pollutants and the corresponding technology
options and plants for which EPA identified observations above the effluent limits. Observations
that are equal to or below the effluent limits can be determined by comparing the observations
shown in Appendix 7 to the data listed in DCN SE06277; these values are not listed in Appendix 7.

In comparing the effluent data to the limits, EPA noted there are instances where one or more daily
values in a month are higher than the monthly limit, but the average of all results in a month are
equal to or less than the monthly average limit and as such the facility would be in compliance with
the monthly limit. Instances such as these are normal and consistent with the way effluent limits are
calculated and implemented in NPDES permits. This is illustrated in the effluent arsenic
concentration data from Hatfield's Ferry during February 2010, described in section 11.1. Also, EPA
identified some cases where only one sample was taken during a month and the resulting
concentration value for that one sample is above the monthly limit. In such cases, additional
monitoring of the effluent (e.g., at weekly intervals) would likely result in a monthly average that
would fall below the monthly average limit.

Based on the results described in the sections below for the comparisons of effluent and influent
data to the limits, and information described elsewhere in the record for the ELGs, EPA determined
that the statistical distributional assumptions are appropriate for the effluent data and that the daily
maximum and monthly average limits for the rule are reasonable and achievable.

Arsenic and Mercury Limits for FGD Wastewater, Based on Terformance of Chemical Trecipitation Technology

EPA calculated effluent limits for arsenic and mercury in FGD wastewater, based on the
performance of chemical precipitation treatment technology. These limitations were transferred to
the biological treatment technology option and thus form the bases for the BAT/PSES arsenic and
mercury limitations applicable to FGD wastewater. EPA calculated the limits using data from four
plants: Hatfield's Ferry, Keystone, Miami Fort, and Pleasant Prairie.

Arsenic - Comparison of effluent data to the daily maximum limit of 11 ug/L:

All observations for two plants were equal to or below the daily maximum limit (24
observations at Keystone; 9 observations at Miami Fort). At Pleasant Prairie, all but one of the
20 observations were equal to or below the daily limit. At Hatfield's Ferry, 102 observations
were equal to or below the daily limit; 28 of the 130 total observations at Hatfield's Ferry were
above the daily limit.

91

-------
Arsenic — Comparison of effluent data to the monthly average limit of 8 ug/L:

Only Keystone and Hatfield's Ferry collected effluent samples with sufficient frequency within
a month to represent weekly sampling. For the time periods where there were sufficient data,
EPA calculated the average of the daily values collected within a month, and compared that
average value to the monthly average limit. For Keystone, all such monthly average values
were below the limit. In fact, every daily observation at Keystone was below the monthly limit.
For Hatfield's Ferry, there were 12 months when the average value was equal to or below the
limit; there were 15 months when the average value was above the monthly limit (most of
those were 1-2 ug/L above the monthly limit). Although Miami Fort did not collect samples
with sufficient frequency to calculate monthly averages, all of the daily observations for the
plant were equal to or below the monthly limit and, therefore, EPA did not identify any
periods of time when the effluent concentrations were higher than the limit.

Mercury - Comparison of effluent data to the daily maximum limit of 788 ng/L:

All observations at Keystone and Miami Fort were below the daily maximum limit (8
observations at Keystone; 68 observations at Miami Fort). At Hatfield's Ferry, 217
observations were below the daily limit; 2 of the 219 total observations at Hatfield's Ferry
were above the daily limit. At Pleasant Prairie, 370 observations were below the daily limit; 5
of the 375 total observations at Pleasant Prairie were above the daily maximum limit.

Mercury — Comparison of effluent data to the monthly average limit of356 ng/L:

Only Hatfield's Ferry and Pleasant Prairie collected effluent samples with sufficient frequency
within a month to represent weekly sampling. For the time periods where there were sufficient
data, EPA calculated the average of the daily values collected within a month, and compared
that average value to the monthly average limit. At Hatfield's Ferry, all but one of the monthly
average values were below the limit (39 of the 40 monthly values). For Pleasant Prairie, there
were 27 months when the average value was below the limit; there were 3 months when the
average value was above the monthly limit. Although Keystone did not collect samples with
sufficient frequency to calculate monthly averages, all of the daily observations for the plant
were equal to or below the monthly limit and, therefore, EPA did not identify any periods of
time when the effluent concentrations were higher than the limit. Miami Fort also did not
collect samples with sufficient frequency to calculate monthly averages; however, it is worth
noting that 61 of the 68 daily observations for the plant were below the monthly limit.

EPA determined that all power plants discharging FGD wastewater, including the plants discussed
here, are capable of meeting the effluent limits for arsenic and mercury. While there are certain
instances where the model plants' effluent data concentrations are higher than the limits, based on
its engineering judgment developed over years of evaluating wastewater treatment processes for
power plants and other industrial sectors, EPA determined that the combination of additional
monitoring, closer operator attention, and optimizing treatment system performance to target the
effluent concentrations at the technology option long-term averages will result in lower effluent
concentrations that would be in compliance with the effluent limits.

92

-------
Although most observations in the datasets used to calculate the effluent limits were below the
limits, there were some observations above the limits. As explained above, it is reasonable for this
situation to arise in datasets used to calculate limits for the rule and there are specific steps that the
plants can take that would enable them to improve treatment system performance so that effluent
concentrations would be in compliance with the limits at all times. Although EPA selected these
plants as representing the best available technology and they provide the best available data for
establishing arsenic and mercury effluent limits that reflect BAT level of treatment for FGD
wastewater, it does not necessarily mean that the plants' systems are fully optimized, especially since
the NPDES permits for these plants either do not include limits for arsenic or mercury in their
discharges of FGD wastewater or because their NPDES permit limits are well above what the
system is capable of achieving.9,10, u'12 This is supported by Duke Energy's comments on the
proposed effluent guidelines, which state that the "performance of the chemical precipitation

9	Pleasant Prairie's NPDES permit (WI-0043583-06-1) FGD wastewater monitoring requirements include both in-plant
requirements (Outfall 102, effluent from the FGD blowdown wastewater treatment system) and final outfall
requirements (Outfall 001, combined discharge to Lake Michigan for 5 internal outfalls, including cooling tower
blowdown and FGD wastewater). Outfall 102 includes no effluent limits for arsenic (requiring only that the plant
monitor for the pollutant monthly); mercury has a daily limit of 1,500 ng/L (monitored twice weekly), but no monthly
limit. Outfall 001 includes a daily maximum limit for mercury (and a monitoring requirement but no effluent limit for
arsenic), but the NPDES permits specifically states that "the FGD effluent may only discharge when sufficient flow
from other wastewater streams (cooling tower blowdown, low volume wastewater, coal pile runoff, or metal cleaning
waste basin) is available if necessary to comply with the water quality based effluent limits at Outfall 001."

10	Miami Fort's NPDES permit (OF10009873) FGD wastewater monitoring requirements include both in-plant
requirements (Outfall 608, FGD wastewater treatment system discharge prior to the ash pond) and final outfall
requirements (Outfall 002, ash pond discharge, including FGD wastewater, prior to the Ohio River). Outfall 608
includes no effluent limits for arsenic or mercury, requiring only that the plant monitor the concentrations of these
pollutants monthly. Outfall 002 similarly includes no effluent limits for arsenic or mercury, requiring only that the plant
monitor the concentrations of these pollutants quarterly.

11	Flatfield's Ferry Power Station's NPDES permit (PA0002941) FGD wastewater monitoring requirements include both
in-plant requirements (Outfall IMP 306, effluent from the FGD scrubber blowdown wastewater treatment plant) and
final outfall requirements (Outfall 006, which includes ash transport water, coal pile runoff, low volume waste and
FGD wastewater treatment system effluent). Outfall IMP 306 includes no effluent limit for arsenic, requiring only that
the plant monitor for the pollutant weekly; the permit includes a mercury daily limit of 10,000 ng/L and a monthly
average limit of 5,000 ng/L (monitored weekly). Outfall 006 includes no effluent limit or monitoring requirement for
arsenic; the permit includes a mercury daily limit of 4,000 ng/L and a monthly average limit of 2,000 ng/L (monitored
weekly).

12	Keystone's NPDES permit (PA0002062) FGD wastewater monitoring requirements include both in-plant
requirements (IMP 101, discharge from the FGD scrubber blowdown wastewater treatment plant) and final outfall
requirements (Outfall 001, which includes discharges from the pipeline pigging wastewater treatment facility and FGD
scrubber blowdown wastewater treatment plant). IMP 101 includes no effluent limit for arsenic, requiring only that the
plant monitor for the pollutant weekly; the permit includes a mercury daily limit of 8,000 ng/L and a monthly average
limit of 4,000 ng/L (monitored weekly). Outfall 001 includes similar effluent monitoring/limits, with monitoring only
required when pigging wastewater is discharged.

93

-------
treatment systems from which EPA's data relies were optimized to meet current facility NPDES
permit requirements and may not reflect the system's maximum performance."13

The two plants that have observations above the limits for arsenic do not have specific limits in their
NPDES permits for discharges of arsenic in FGD wastewater. Also, only three of the plants have
NPDES permit limits for mercury, and for two of these plants the permit limits are more than 10
times higher than the BAT effluent limits established by the final ELGs. For this reason, these
plants currently do not need to closely monitor the concentrations of arsenic and mercury in the
treatment system effluent, nor do they need to take steps to optimize the removal of these
pollutants. This is illustrated by the plants' operational practices for their FGD wastewater treatment
systems. Other than Pleasant Prairie targeting an effluent concentration that would allow discharges
containing mercury at double the concentration established for the final ELGs, none of the plants
reported having operational target parameters for the allowable level of mercury and arsenic in their
effluent.14'15'16

The data in the record supports EPA's determination that these plants would be able to meet the
limits established for the final ELGs, although certain plants may need to make some adjustments to

13	Duke Energy's comments state that "[t]he operation and performance of chemical precipitation systems for FGD
water treatment is continuing to evolve and improve. The industry's current chemical precipitation system performance
data does not accurately reflect optimized performance.."The performance of the chemical precipitation treatment
systems from which EPA's data relies were optimized to meet current facility NPDES permit requirements and may
not reflect the system's maximum performance. For example, operating these systems at higher pF[ levels can increase
the percentage of metal removal. Improved treatment chemicals, improved clarification and filtration can further
increase removal percentages. If necessary, modifications to chemical precipitation systems towards optimizing
removal of specific constituents, not currently permitted, can result in more effective treatment of FGD wastewater."
Duke Energy goes on to state that "Miami Fort has implemented several improvements to their chemical precipitation
process to improve mercury removal. Increasing pH from initial startup settings, adding coagulants, and organosulfide
metal precipitants have led in favorable results. Polymer delivery systems have been modified and the baffling inside
the clarifier have also been changed to improve solids settling and reduce the Total Suspended Solids (TSS)."
Comments of Duke Energy to the United States Environmental Protection Agency, p. 13. September 19, 2013.

14	When asked about the operational target parameters for the treatment system, Pleasant Prairie reported that they
target achieving an effluent concentration below 1,000 ng/L in the effluent from the secondary clarifier. If the
measured concentration is below 1,000 ng/L, discharge is continuous. If clarifier effluent concentration is in the range
1,000-1,500 ng/L, continuous discharge ceases and depending on samples collected from the effluent tank the
wastewater is either discharged (if <1,500 ng/L) or recirculated for further processing (if >1,500 ng/L). If the clarifier
effluent is above 1,500 ng/L, the wastewater is recirculated to the absorber or treatment system equalization tank for
further processing. DCN SE04328.

15	When asked about the operational target parameters for the treatment system, Miami Fort reported that they make
day-to-day adjustments based on pH, turbidity, and TSS to make sure there is good settling and clarification.

Depending on settling performance, they may adjust the organosulfide addition. Miami Fort reported that they do not
target specific metals concentrations for the treatment system effluent. DCN SE04331.

16	FirstEnergy stated that the plant did not have any specific operational targets for metals. The FGD wastewater
treatment system was operated based on maintaining the pH, floe in the clarifier, and TSS in the effluent from the sand
filter. DCN SE04316.

94

-------
their operating practices. The effluent concentrations for both Keystone and Miami Fort were below
the effluent limits for arsenic and mercury. Both Pleasant Prairie and Hatfield's Ferry had some
observations above the limits for arsenic or mercury, but these can be attributed to their current
operating practices.

As noted above, the NPDES permit for Hatfield's Ferry includes no arsenic limits for the effluent
from the FGD wastewater treatment system. The NPDES permit daily maximum limit for mercury
is 10,000 ng/L (5,000 ng/L monthly limit). Additionally, the plant did not operate its treatment
system to achieve any specific operational targets for arsenic, nor for any other metal. Instead, the
treatment system is operated to maintain pH within an operational range, ensure the clarifler
indicates good settling of the precipitated floe, and provide adequate removal of TSS in the effluent
from the sand filters. It is reasonable to expect that the chemical addition rates for sodium
hydroxide, organosulfide, and other additives are not optimized for arsenic removal because they are
optimized only to maintain pH and TSS. Pleasant Prairie's NPDES permit includes a daily maximum
limit for mercury at 1,500 ng/L in the treatment system effluent, but no monthly limit. The plant's
operational target of 1,000 ng/L mercury for the treatment system effluent is very close to the
permit limit but significantly higher than the limits established by the ELGs. Given the high permit
limit and operational target, it is not surprising that a small number of observations are above the
limits established by the ELGs.

Both plants could take steps to ensure compliance with the effluent limits by making adjustments to
the treatment system operation to target the long-term average performance that the effluent limits
are based upon (i.e., 5.98 ng/L for arsenic and 159 ng/L for mercury). Operator attention to
effluent quality and process control indicators (e.g., wastewater flowrates, conductivity, clarifier bed
levels, TSS) facilitate steady state operation of the treatment system, as well as alerting operators of
system abnormalities or fluctuations in influent quality or flowrate. EPA's review of chemical
precipitation systems for this industry noted that plants could benefit from using an in-house
mercury analyzer to monitor the performance of the system on a daily basis (this was included as
part of the cost basis for the technology option). Mercury analyzers have been effectively used at
Pleasant Prairie to alert operators when mercury concentrations begin trending upward so that they
may take steps (such as altering the dosage rates for chemical additives) to adjust treatment system
performance to meet their current permit limitations.

Optimizing the treatment system to target effluent concentrations at the long-term average is
important, and is relevant to the evaluation of the effluent data for Hatfield's Ferry and Pleasant
Prairie. EPA evaluated the results of testing of a chemical precipitation treatment system at a power

95

-------
plant that, although the treatment system had been in operation for more than a year and was
operating at a steady state condition, the plant was able to significantly improve the pollutant
removal performance merely by altering the dosage rates for the wastewater treatment chemical
additives.17 The results of this study, as well as other information in the record, supports EPA's
determination that these plants can improve treatment system performance and meet the ELG limits
at all times. EPA notes that its compliance cost estimates for the final rule include costs for mercury
analyzers, organosulfide addition, proper dosing of treatment system chemical additives, and
increased staffing to operate the treatment system.

It is important to note that although the BAT limits and PSES for arsenic and mercury in the final
ELGs are based on chemical precipitation technology,18 the selected BAT/PSES technology option
for FGD wastewater is actually comprised of the combination of chemical precipitation followed by
biological treatment, which is more effective than chemical precipitation treatment alone. The data
in the record for the final rule demonstrate that the biological treatment stage provides pollutant
reductions for arsenic and mercury (and other pollutants of concern with similar removal
mechanisms) in addition to the pollutant removals that occur in the chemical precipitation stage of
the biological treatment technology option (see, e.g., the data plots and tables in Sections 7.1 and 7.2,
showing an average of 31 percent removal of arsenic and 99 percent removal of mercury across the
biological treatment stage at Allen and Belews Creek).

These additional pollutant removals are corroborated by the results of pilot testing conducted at
Indianapolis Power and Light's Petersburg power plant, which showed that both arsenic and

17	AEP's Mountaineer plant operated a chemical precipitation system to treat FGD wastewater, with operation targeted
to meet only the BPT-based limitations for TSS, pH, and oil and grease. In 2008, one year after the start-up of the
FGD scrubbers and the FGD wastewater treatment system, the plant went through a permit renewal process whereby
the permitting authority proposed to add a water quality-based effluent limit for mercury. Based on the mercury
limitations in the draft permit, AEP conducted a pilot study evaluating three different technologies that could be
installed to further treat the effluent from the chemical precipitation system. AEP conducted the pilot study from July
through December 2008. During the first three months of the study, the mercury concentrations of the chemical
precipitation system effluent feeding the pilot tests averaged 1,300 ng/L. Since none of the three technologies were
achieving the targeted effluent concentrations for the pilot testing, AEP took steps to optimize the precipitation of
dissolved metals and the removal ofprecipitants and other suspended solids in the chemical precipitation system,
including adding additional polymers and organosulfide. Using these optimization steps, AEP noted that "[t]he
combination of supplemental coagulation and organosulfide addition consistently yielded approximately 80 percent of
additional mercury reduction..." within the chemical precipitation system. American Electric Power Mercury Removal
Effectiveness Report. (January 29, 2010). DCN SE02008.

18	See Section 13 of the Technical Development Document for a discussion of the transfer of effluent limits for arsenic
and mercury from chemical precipitation technology to the selected BAT technology option (chemical precipitation
followed by biological treatment).

96

-------
mercury were effectively removed by the biological treatment stage.19 Data for a full-scale FGD
wastewater treatment system at AEP's Mountaineer plant similarly shows both arsenic and mercury
are effectively removed by the biological treatment stage.20 Thus, plants employing and optimally
operating all components of the biological treatment technology option (including adding
organosulfide to achieve sulfide precipitation) should achieve pollutant removals for arsenic and
mercury (and other pollutants with similar removal mechanisms) that are even greater than the
removals based on chemical precipitation technology alone.

Selenium and Nitrate-nitrite as N Limits for FGD Wastewater, Based on Chemical Vrecipitation Followed by
Biological Treatment Technology Option for FGD Wastewater

The ELGs establish BAT limits and PSES for arsenic, mercury, nitrate-nitrite as N, and selenium in
FGD wastewater based on the biological treatment technology option. As mentioned above, the
limits for arsenic and mercury were transferred from the chemical precipitation technology option.
See the discussion above for the comparison of effluent data to the limits for these two pollutants.
The effluent limits for selenium and nitrate-nitrite as N were calculated using data from two plants:
Allen and Belews Creek.

Nitrate-nitrite as N - Comparison of effluent data to the daily maximum limit of 17.0 mglL:

For both Allen and Belews Creek, all observations were below the daily limit (30 observations
at Allen; 40 observations at Belews Creek).

Nitrate-nitrite as N — Comparison of effluent data to the monthly average limit of 4.4 mglL:

Only Belews Creek collected effluent samples with sufficient frequency within a month to
represent weekly sampling. For the time periods where there were sufficient data, EPA
calculated the average of the daily values collected within a month, and compared that average
value to the monthly average limit. All of the monthly average values for Belews Creek were
below the limit. Although Allen did not collect samples with sufficient frequency to calculate
monthly averages, all but two of the 30 daily observations for the plant were below the
monthly limit. The two daily observations that were above the monthly limit are associated
with a spike in effluent concentration that occurred in December 2011.

19	Higgins, T., et al. "Recent Applications of Meeting Compliance Challenges through Flue Gas Desulfurization (FGD)
Wastewater," Presented at the Power Plant Pollutant Control "MEGA" Symposium. August 19-21, 2014. Also see
June 6, 2015 email from Tom Fliggins (CF12M FEill), presenting data showing the biological treatment stage removed
87% of the mercury and 84% of the arsenic entering that stage.

20	The FGD wastewater treatment system at AEP's Mountaineer plant includes chemical precipitation (both hydroxide
and sulfide precipitation, as well as iron coprecipitation) followed by anoxic/anaerobic biological treatment designed to
remove selenium. DCN SE05664/SE05645.

97

-------
Selenium - Comparison of effluent data to the daily maximum limit of 23 ug/ L:

At Allen, 178 observations were below the daily limit; 4 of the 182 total observations at Allen
were above the daily limit. At Belews Creek, 214 observations were below the daily limit; 2 of
the 216 total observations at Belews Creek were above the daily maximum limit.

Selenium — Comparison of effluent data to the monthly average limit of 12 ug/ L:

All monthly averages for Allen were below the monthly average. At Belews Creek, there were
10 months when the monthly average was equal to or below the monthly limit; there were 2
months when the average was above the limit. Both of these months occurred shortly after
the end of the initial commissioning period for the treatment system (the initial
commissioning period ended 6/11/2008; the two monthly averages above the monthly limit
were in August and September 2008).

EPA determined that all power plants discharging FGD wastewater, including the plants discussed
here, are capable of meeting the effluent limits for selenium and nitrate-nitrite as N. While there
were certain instances where the plants' effluent data are higher than the limits, based on its
engineering judgment developed over years of evaluating wastewater treatment processes for power
plants and other industrial sectors, including both physical/chemical and biological treatment
technologies, EPA determined that the combination of additional monitoring, closer operator
attention, and optimizing treatment system performance to target the effluent concentrations at the
technology option long-term averages will result in lower effluent concentrations that would be in
compliance with the effluent limits.

Although most observations in the datasets used to calculate the effluent limits were below the
limits, there were some observations above the limits. It is reasonable for this situation to arise in
datasets used to calculate limits for ELGs, particularly when the data are from plants that do not
have NPDES limits for the pollutants, and when there are specific steps that the plants can take that
would enable them to improve treatment system performance so that effluent concentrations would
be in compliance with the limits at all times. Although EPA selected these plants as representing the
best available technology, neither plants' system is fully optimized for removal of selenium or
nitrate-nitrite because they are not targeting specific effluent levels for these pollutants, nor are they
operationally controlling the treatment system (e.g., adjusting dosages for chemical additives or
altering bioreactor bed contact time) to maintain effluent concentrations below specified
concentrations. Neither plants' NPDES permits include effluent limits or monitoring requirements
nitrate-nitrite. The NPDES permits for these plants also do not include effluent limits for selenium,

98

-------
although they do include require them to periodically monitor effluent concentrations of
selenium.21,22

For selenium at Allen and Belews Creek, EPA identified a small number of observations that are
above the daily limit, or where the monthly average is above the monthly limit. As EPA explained
above, there are steps plants can take to achieve better treatment system performance to ensure
compliance with the effluent limits. After evaluating all selenium data for the biological treatment
technology option (more than five years of data for Belews Creek and more than four years for
Allen, excluding the initial commissioning periods for the treatment systems), EPA concluded that
the observations above the daily limit are reflective of periods of less than optimum performance of
the treatment system and were due either to the inexperience of operators with this type of
treatment system, operators not targeting their treatment systems to attain a specific effluent
concentration, failure on the part of operators to either closely monitor treatment system
performance or to respond in a timely manner to restore the system to steady state condition, or a
combination thereof.23 For example, the monthly averages at Belews Creek that are above the limits
occurred only 2-3 months following the end of the initial commissioning period and are associated
with daily observations that are significantly elevated relative to other observations in the months
immediately following that time, as well as nearly every other daily observation for Belews Creek.
Likewise, the observations at Allen that are above the daily limit are associated with two periods
where the effluent selenium concentrations spike upward significantly, and a sudden upward spike in

21	Plant Allen Steam Station's NPDES permit (NC0004979) FGD wastewater monitoring requirements include both in-
plant requirements (Internal Outfall 005, effluent from the FGD wet scrubber wastewater treatment system) and final
outfall requirements (Outfall 002, ash pond effluent including FGD wastewater and other wastestreams). Internal
Outfall 005 includes no effluent limits or monitoring requirement for nitrate-nitrite as N; selenium also has no effluent
limit and is required to be monitored monthly. Outfall 002 also includes no effluent limits or monitoring requirement
for nitrate-nitrite as N; selenium has no effluent limit and is required to be monitored monthly.

22	Belews Creek's NPDES permit (NC0024406) FGD wastewater monitoring requirements include both in-plant
requirements (Internal Outfall 002, treated FGD wet scrubber wastewater to ash settling basin) and final outfall
requirements (Outfall 003, discharge to the Dan River from the ash settling pond, which contains treated FGD
wastewater and other wastes). Internal Outfall 002 includes no effluent limits or monitoring requirement for nitrate-
nitrite as N; selenium also has no effluent limit but must be monitored quarterly. Outfall 003 also includes no effluent
limits or monitoring requirement for nitrate-nitrite as N; selenium has no effluent limit but must be monitored
monthly.

23	Note that although the Belews Creek selenium data (and other pollutant as well) from mid-2008 may be influenced by
the initial commissioning period for the treatment system, EPA used these data when calculating the final effluent
limitations for the biological treatment technology option for FGD wastewater. EPA has used these data because
although EPA believes that, as a general rule, the initial commissioning period duration will be on the order of 3-4
months, and certainly no more than 6 months except in unique circumstances; EPA has not confirmed that the initial
commissioning for Belews Creek was of such exceptional duration. Comments on the proposed ELGs stated that the
treatment system operators at Belews Creek reported 6/11/2008 as the end of the initial commissioning period.
Without the information to confirm the commissioning period was still in progress, EPA concluded that the sampling
data should be used when calculating effluent limits.

99

-------
effluent concentrations of nitrate-nitrite also occurred during one of those periods. These effluent
spikes indicate that the nutrient feed may have been insufficient or the bed contact time may not
have been long enough to allow the microorganisms in the bioreactor to complete the biochemical
reduction processes for nitrate-nitrite and selenium.

Both plants could take steps to ensure compliance with the limits by making adjustments to the
treatment system operation to target the long-term average performance that the effluent limits are
based upon (i.e., 1.3 mg/L for nitrate-nitrite as N and 7.5 ng/L for selenium). Operator attention to
influent and effluent quality and process control indicators (e.g., wastewater flowrates, ORP, TSS,
turbidity, gas production rates) facilitate steady state operation of the treatment system, as well as
alerting operators of system abnormalities or fluctuations in wastewater flowrate or quality (influent
or effluent). Significant changes in wastewater flowrate, if not managed properly, can affect effluent
quality because it will affect the bed contact time, as well as potentially affecting the appropriate
nutrient dosage. The influent ORP provides insight to the amount of nutrient that should be added
to maintain the appropriate carbon to nitrogen (C:N) ratio to support the microbial activity needed
to reduce nitrate-nitrite and selenium. ORP within the bioreactor should be monitored to ensure
that the wastewater has sufficient bed contact time with the biomass at low (-300 to -400 mV) ORP
for sufficient time for selenium reduction. Properly operating the system includes being attentive to
changes in pressure drop across the system to ensure the routine flush/backwash cycles are
occurring with sufficient frequency.

EPA notes that the compliance cost estimates for the final rule include costs for ORP monitoring of
the wastewater influent and within the bioreactor, proper dosing of the nutrient additive, chemical
feed system to remove free oxidants prior to the bioreactor, and increased staffing to operate the
treatment system. Although not necessary to operate the treatment system properly, some plants
may find it desirable to implement additional process controls, such as frequent monitoring of
nitrates or selenium (e.g., daily or once per shift) with available test kits or analyzers which, although
some of these may not be approved in 40 CFR 136 for NPDES purposes, their analytical results can
be correlated to approved methods (such as ICP-MS analysis for selenium) and used as another real-
time indicator of treatment system influent or effluent characteristics for enhanced process control.24

24 Similar to the mercury analyzer included as part of the technology basis for the effluent limits, these additional process
control options can provide treatment system operators with additional information, such as the influent nitrate-nitrite
concentrations (for use in addition to ORP to confirm the appropriate nutrient dosage), between stage or effluent
nitrate-nitrite concentrations (further confirmation of the nutrient dosage to support microbial activity for pollutant
reduction), or effluent selenium concentrations.

100

-------
Additionally, inexpensive test kits for measuring for oxidants in the treatment system influent are
available and can be used as another source of information about wastewater characteristics.

Based on the results of this comparison for the biological treatment technology option, EPA
determined that the statistical distributional assumptions are appropriate for the effluent data and
that the limits are reasonable.

Limits for FGD Wastewater Based on Chemical Vrecipitation Followed By Vapor-Compression Evaporation
(Arsenic, Mercury, Selenium and TPS)

The final rule establishes NSPS/PSNS for arsenic, mercury, selenium and TDS in FGD wastewater,
based on the performance of chemical precipitation followed by vapor-compression evaporation
treatment using data from Brindisi plant.25 All daily concentration values are equal to or below the
daily limits for all parameters. (The data for this plant were not collected at sufficient frequency to
represent weekly sampling; thus, monthly averages could not be calculated for comparison to the
monthly limit.) After thoroughly reviewing the data, EPA determined that the statistical
distributional assumptions are appropriate for the effluent data and that the limits are reasonable and
achievable.

Limitsfor Gasification Wastewater Based on Vapor-Compression Evaporation (Arsenic, Mercuy, Selenium and
TPS)

The final rule establishes BAT/PSES and NSPS/PSNS for arsenic, mercury, selenium and TDS in
gasification wastewater based on vapor-compression evaporation treatment. The limits for selenium
and TDS were calculated using data from two plants: Polk and Wabash River plants. The limits for
arsenic and mercury were based only on data from Polk, because the arsenic and mercury data from
Wabash River failed the LTA test (see section 4 for a discussion of the LTA test).

For arsenic and mercury, the daily concentration values used to calculate the limits (i.e., from Polk)
are below the daily limits.26 For TDS, all observations for both Polk and Wabash River are below the
daily limit.

25	The final rule also establishes BAT limitations for FGD wastewater that are the same as the NSPS for FGD
wastewater for plants who opt into the voluntary incentives program.

26	Although the arsenic and mercury data for the vapor-compression evaporator condensate at Wabash River were not
used to calculate the limits, due to failing the LTA test, EPA nevertheless compared these data to the limits and found
that all observations were equal to or below both the daily and monthly limits.

101

-------
For selenium, all observations for Wabash River were below the daily limit. At Polk, there is one
observation above the daily limit. As discussed above in section 9, the data for the Polk treatment
system indicates that the evaporator (or at a minimum the forced circulation evaporation stage) was
operating abnormally and allowing carryover of pollutants to the condensate effluent stream.27 Based
upon its review of the data, EPA concluded if the plant designs and operates its treatment system to
achieve the option long-term average for the model technology, then the plant will be able to
comply with the limits. Furthermore, EPA notes that Polk reuses all treated gasification wastewater
(i.e., condensate) in the gasification process and does not discharge any gasification wastewater. As
such, the plant's treatment objective is to ensure the wastewater is of sufficient quality for reuse in
the process rather than to comply with a NPDES permit limit.

The data for these plants were not collected at sufficient frequency to represent weekly sampling;
thus, monthly averages could not be calculated for comparison to the monthly limits.

After thoroughly reviewing the data, EPA concluded that the statistical distributional assumption is
appropriate for the effluent data and that the limits are reasonable.

11.2 Comparison of the Effluent Limits to Influent Data

In addition to comparing the limits to the effluent data used to develop the limits, EPA also
compared the pollutant concentrations for the treatment system influent to the daily limits. This
comparison helps evaluate whether the limits are set at a level that ensures that treatment of the
wastewater and that the influent concentrations were generally well-controlled by the treatment
system See Appendix 7 for a detailed listing of the summary statistics for the influent data for each
pollutant in each treatment technology option (Tables A7.6 to A7.9).

For all treatment technology options for both FGD and gasification wastewater, the minimum,
average, and maximum influent concentration values were much higher than the long-term average
and limits. EPA found that influent concentrations were generally well-controlled by the treatment
plant for all plants with the model technology. In general, the treatment systems adequately treated
even the extreme influent values, and the high effluent values did not appear to be the result of high
influent discharges.

27 Polk's data for the forced circulation evaporator condensate were not used to calculate the limits due to that portion
of the treatment system being in an upset condition; therefore, these data were not compared to the effluent limits.

102

-------
Appendix 1. List of All Observations that Were

Corrected Before Calculating the Limits

1-1

-------
Table of Contents

Table	Page

Al.l	Summary of the corrected detection indicators for the Belews

Creek and Keystone self-monitoring data, arranged by plant,

sampling location, pollutant, and sampling date	 1-4

A1.2	Summary of the data corrections for the Allen, Belews Creek, and

Keystone self-monitoring data, arranged by plant, sampling

location, pollutant, and sampling date	 1-5

A1.3 Summary of the corrected dates for the Belews Creek plant self-
monitoring data, arranged by sampling location, pollutant, and
sampling date	 1-5

1-2

-------
This appendix contains the listing of all data corrections that were made to the self-monitoring data
provided by the industry for the proposed and final rulemaking. Since the data EPA used for the
proposed rule are also used for the final rule, the "pre-proposal corrections" are included here along
with the "post-proposal corrections." For details about the "pre-proposal corrections," see the
listing in Appendix 1 of "Steam Electric Proposed Effluent Limits_10_20_2012.docx" (DCN
SE01999).

Table Al.l lists the corrections made to the detection indicator for data at Belews Creek and
Keystone prior to calculating effluent limits (Duke Energy and NRG confirmed that the errors were
due to the data entry errors). The data are arranged by plant, sampling location, pollutant, and
sampling date.

1-3

-------
Table Al.l Summary of the corrected detection indicators for the Beiews Creek and Keystone
self-monitoring data, arranged by plant, sampling location, pollutant, and sampling
date

Plant

Sampling Location

Pollutant

Date

Original
Indicator1

Corrected
Indicator1

Beiews
Creek

Bio 1 In (Bioreactor
Influent)

Mercury

2/6/2008

D

ND

2/9/2008

D

ND

Selenium

1/17/2011

N

ND

Bio 2 Eff (Bioreactor
Effluent)

arsenic

2/7/2008

ND

D

2/8/2008

ND

D

2/10/2008

ND

D

2/12/2008

ND

D

2/13/2008

ND

D

2/14/2008

ND

D

2/15/2008

ND

D

2/16/2008

ND

D

2/17/2008

ND

D

2/21/2008

ND

D

Mercury

2/6/2008

D

ND

9/29/2008

ND

D

8/11/2010

ND

D

9/8/2010

ND

D

10/7/2010

ND

D

Selenium

3/3/2008

ND

D

Keystone

Chemical Precipitation
Effluent

Mercury

4/10/2012

D

ND

4/11/2012

D

ND

4/12/2012

D

ND

4/15/2013

D

ND

1D = detected and ND = non-detected.

Table A1.2 lists the corrections made to concentration values for data from Allen, Beiews Creek,
and Keystone prior to performing the statistical analyses, based on information provided by Duke
Energy and NRG.

1-4

-------
Table A1.2 Summary of the data corrections for the Allen, Belews Creek, and Keystone self-

monitoring data, arranged by plant, sampling location, pollutant, and sampling date

Plant

Sampling Location

Pollutant (Unit)

Date

Original
Concentration
Value
(Indicator)1

Corrected
Concentration
Value
(Indicator)1

Allen

Bio 2 Eff (Bioreactor
Effluent)

Nitrate-Nitrite
as N (mg/L)

12/20/2011

97(D)

1(D)

Belews
Creek

Bio 1 In (Bioreactor
Influent)

Mercury (ng/L)

6/9/2010

59.3(D)

59,300(D)

7/14/2010

49.9(D)

49,900(D)

8/11/2010

47.7(D)

47,700(D)

Bio 2 Eff (Bioreactor
Effluent)

Mercury (ng/L)

5/12/2010

1000(ND)

136(D)

5/26/2010

136(D)

-

Keystone

Chemical Precipitation
Effluent

Arsenic (pg/L)

7/24/2012

20(D)

21.2 (D)

8/12/2013

20(D)

23.7 (D)

iD = detected and ND = non-detected.
2No data available for this date.

Table A1.3 lists the corrections made to sample collection date for data from Belews Creek prior to
performing the statistical analyses, based on information provided by Duke Energy.

Table A1.3 Summary of the corrected dates for the Belews Creek plant self-monitoring data,
arranged by sampling location, pollutant, and sampling date

Sampling Location

Pollutant

Original Date

Corrected Date

Bio 1 In (Bioreactor Influent)

Arsenic

4/26/2011

4/27/2011

Selenium

4/26/2011

4/27/2011

1-5

-------
Appendix 2. List of All Observations that Were

Excluded Before Calculating the Limits
for the Final Rulemaking

2-1

-------
Table of Contents

Table	Page

A2.1	List of excluded plant self-monitoring data at Allen		2-4

A2.2	List of excluded data at Belews Creek		2-7

A2.3	List of excluded data at Hatfield's Ferry		2-17

A2.4	List of excluded data at Keystone		2-20

A2.5	List of excluded plant self-monitoring data at Miami Fort		2-21

A2.6	List of excluded plant self-monitoring data at Pleasant Prairie		2-24

A2.7	List of excluded CWA 308 sampling data at Wabash River		2-26

2-2

-------
This appendix contains listing of all observations that were excluded prior to calculating limits for
the final rulemaking. The data excluded are listed by plant and analyte. The reasons for exclusion are
listed below and the rationales for each of the exclusions are detailed in the tables:

¦	Analytical Interferences

¦	Analytical anomalies

¦	Associated with TMT 15

¦	Data Entry Error

¦	Decommissioning Period

¦	Failed data editing criteria

¦	Insufficiently-sensitive analytical method

¦	Not Approved NPDES Compliance Monitoring Method

¦	Problem associated with start-up

¦	Questionable results due to QA/QC issues

¦	Sampling or Analytical Error

¦	Secondary clarifier sampling location not representative of effluent discharge

¦	Start-up or commissioning period

¦	System upset due to pipe broke. Exclude all data

¦	Treatment system upset or abnormal operation

¦	Unable to Confirm the Validity of the Results

¦	Unable to validate the results due to contradictory information

2-3

-------
Table A2.1 List of excluded plant self-monitoring data at Allen

Pollutant
(unit)

Date

Sampling Location

Indicator1

Concentration

Exclusion Rationale

Arsenic,
Mercury,
Selenium

3/3/2009 to
7/13/2009

All Locations





Start-up or commissioning period

EPA excluded data associated with the initial start-up or
commissioning period for a new wastewater treatment system
because the data tend to exhibit relatively high variability not
representative of typical operation. During the initial commissioning
period, the treatment system undergoes a variety of testing to
demonstrate that hydraulic flows through the system and equipment
such as pumps, valves, and other equipment operate as designed.
During this period, it is common for changes in process operations to
occur as the treatment system undergoes performance verification
testing (sometimes referred to as acceptance testing) and the
operation is modified to identify approaches to optimize (i.e.,
minimize) operational costs or to make adjustments to improve
pollutant reductions. Refinements that may take place during the
commissioning period include adjustments to chemical feed locations
and/or feed rates; evaluating different pump cycles, filter backwash
cycles, clarifier overflow rates, or sludge removal cycles; and changes
to the chemicals used. Generally, this initial period of operation also
serves as a time when the operators gain familiarity with the
intricacies of the treatment system operation. During this acclimation
and optimization period the effluent concentration values may exhibit
higher variability than would typically be observed for a well-operated
treatment system, producing occasional extreme values (either high or
low concentrations). After this initial adjustment period, a well-
operated system should operate at steady state (or more precisely, a
"quasi^steady state") with relatively low variability around a long-term
average. This period of instability affects the pollutant removal efficacy
and is not reflective of BAT/NSPS level of performance and therefore
data associated with the initial commissioning period are excluded.
The end of the initial commissioning period was the date specified by
the treatment system operators (as reported in UWAG comments on
the proposed rule).

2-4

-------
Table A2.1

List of excluded plant self-monitoring data at Allen (continued)

Pollutant
(unit)

Date

Sampling Location

Indicator1

Concentration

Exclusion Rationale

Mercury
(ng/L)

12/29/2009

Bio Treatment Influent

D

710

Questionable results due to QA/QC issues

EPA excluded mercury data that Duke Energy identified as
questionable due to quality control issues. Duke Energy stated that
these data are questionable because they suspected that the matrix
spike/matrix spike duplicate (MS/MSD) were not conducted using the
FGD wastewater matrix or the MS/MSD indicated inadequate
recoveries. EPA lacked sufficient laboratory quality control data for
these results to make an independent assessment of whether these
mercury results should have been retained and included in the
statistical analyses.

12/29/2009

Bio Treatment Effluent

D

11

1/11/2010

Bio Treatment Influent

D

92

1/11/2010

Bio Treatment Effluent

D

14.2

Mercury
(ng/L)

7/15/2009

Bio Treatment Effluent

ND

1000

Insufficiently-sensitive analytical method (Method 245.1)
See Analytical Methods Review Memo

7/16/2009

Bio Treatment Effluent

ND

1000

7/21/2009

Bio Treatment Effluent

ND

1000

7/23/2009

Bio Treatment Effluent

ND

1000

7/27/2009

Bio Treatment Effluent

ND

1000

7/29/2009

Bio Treatment Effluent

ND

1000

7/30/2009

Bio Treatment Effluent

ND

1000

6/29/2010

Bio Treatment Effluent

ND

1000

8/23/2010

Bio Treatment Effluent

ND

1000

10/5/2010

Bio Treatment Effluent

ND

1000

10/6/2010

Bio Treatment Effluent

ND

1000

10/7/2010

Bio Treatment Effluent

ND

1000

10/9/2010

Bio Treatment Effluent

ND

1000

10/10/2010

Bio Treatment Effluent

ND

1000

10/12/2010

Bio Treatment Effluent

ND

1000

10/13/2010

Bio Treatment Effluent

ND

1000

12/1/2010

Bio Treatment Effluent

ND

1000

12/4/2010

Bio Treatment Effluent

ND

1000

12/7/2010

Bio Treatment Effluent

ND

1000

1/12/2011

Bio Treatment Effluent

ND

1000

1/24/2011

Bio Treatment Effluent

ND

1000

4/27/2011

Bio Treatment Effluent

ND

1000

2-5

-------
Table A2.1

List of excluded plant self-monitoring data at Allen (continued)

Pollutant
(unit)

Date

Sampling Location

Indicator1

Concentration

Exclusion Rationale

Mercury

(ng/L)

(cont'd)

12/27/2011

Bio Treatment Effluent

D

3010

Insufficiently-sensitive analytical method (Method 245.1)

See Analytical Methods Review Memo

(cont'd)

10/17/2012

Bio Treatment Effluent

ND

1000

11/13/2012

Bio Treatment Effluent

ND

1000

11/26/2012

Bio Treatment Effluent

ND

1000

12/16/2012

Bio Treatment Effluent

ND

1000

3/8/2013

Bio Treatment Effluent

ND

1000

Arsenic,
Mercury,
Selenium

3/8/2010 to
5/25/2010

All Locations





System upset due to pipe broke. Exclude all data

Piping within the treatment system broke, resulting in a treatment
system upset and degraded pollutant removal performance. EPA
excluded all of the data collected at Allen on dates 3/8/10, 3/22/10,
4/5/10, 4/26/10, 5/10/10, and 5/25/10 for this reason. When
reviewing the data, EPA observed periodic occurrences in early 2010
of unusually high concentrations of arsenic in the bioreactor effluent
relative to effluent from the bioreactor influent treatment stage.
According to Duke Energy personnel, the treatment system suffered an
upset condition beginning 3/2/10, that particularly impacted arsenic
concentrations exiting the bioreactor. The treatment system suffered
pipe failures in the bioreactor on 3/2/10 and 3/11/10, dislodging a
portion of the carbon within the bioreactor cells. It took approximately
four weeks for repairs to be completed, and additional time for the
treatment system performance to stabilize. According to Duke Energy,
elevated arsenic concentrations in the bioreactor effluent during the
period March through May 2010 are likely attributable to the pipe
breakages and related repair activities. Since the overall performance
of the treatment system was degraded during this period, EPA
excluded all data collected from date of the first pipe failure until the
end of May.

1 D=Detected, ND=Non-detected.

2-6

-------
Table A2.2 List of excluded data at Belews Creek

Data Source

Pollutant
(unit)

Date

Sampling Location

Indicator

Concentration

Exclusion Rationale

Plant Self-
Monitoring

Arsenic,
Mercury,
Selenium

2/6/2008 to
6/11/2008

All locations





Start-up or commissioning period

EPA excluded data associated with the initial start-up or
commissioning period for a new wastewater treatment
system because the data tend to exhibit relatively high
variability not representative of typical operation. During
the initial commissioning period, the treatment system
undergoes a variety of testing to demonstrate that
hydraulic flows through the system and equipment such
as pumps, valves, and other equipment operate as
designed. During this period, it is common for changes
in process operations to occur as the treatment system
undergoes performance verification testing (sometimes
referred to as acceptance testing) and the operation is
modified to identify approaches to optimize (i.e.,
minimize) operational costs or to make adjustments to
improve pollutant reductions. Refinements that may
take place during the commissioning period include
adjustments to chemical feed locations and/or feed
rates; evaluating different pump cycles, filter backwash
cycles, clarifier overflow rates, or sludge removal cycles;
and changes to the chemicals used. Generally, this
initial period of operation also serves as a time when
the operators gain familiarity with the intricacies of the
treatment system operation. During this acclimation
and optimization period the effluent concentration
values may exhibit higher variability than would typically
be observed for a well-operated treatment system,
producing occasional extreme values (either high or low
concentrations). After this initial adjustment period, a
well-operated system should operate at steady state (or
more precisely, a "quasi^steady state") with relatively
low variability around a long-term average. This period
of instability affects the pollutant removal efficacy and

2-7

-------
Table A2.2

List of excluded data at Belews Creek (continued)

Data Source

Pollutant
(unit)

Date

Sampling Location

Indicator

Concentration

Exclusion Rationale

Plant Self-
Monitoring
(cont'd)











is not reflective of BAT/NSPS level of performance and
therefore data associated with the initial commissioning
period are excluded. The end of the initial
commissioning period was the date specified by the
treatment system operators (as reported in UWAG
comments on the proposed rule).

Plant Self-
Monitoring

Mercury (ng/L)

8/4/2008

Bio Treatment Influent

D

4.9

Questionable results due to QA/QC issues

EPA excluded mercury data that Duke Energy identified
as questionable due to quality control issues. Duke
Energy stated that these data are questionable because
they suspected that the matrix spike/matrix spike
duplicate (MS/MSD) were not conducted using the FGD
wastewater matrix or the MS/MSD indicated inadequate
recoveries. EPA lacked sufficient laboratory quality
control data for these results to make an independent
assessment of whether these mercury results should
have been retained and included in the statistical
analyses.

8/4/2008

Bio Treatment Effluent

D

2.4

8/11/2008

Bio Treatment Influent

D

4.8

8/11/2008

Bio Treatment Effluent

D

2.7

8/18/2008

Bio Treatment Influent

D

24

8/18/2008

Bio Treatment Effluent

D

5.3

8/25/2008

Bio Treatment Influent

D

27

8/25/2008

Bio Treatment Effluent

D

6.4

9/2/2008

Bio Treatment Influent

D

3.1

9/2/2008

Bio Treatment Effluent

D

2.1

9/8/2008

Bio Treatment Influent

D

2.7

9/8/2008

Bio Treatment Effluent

D

4.1

9/15/2008

Bio Treatment Influent

D

12

9/15/2008

Bio Treatment Effluent

ND

1

9/22/2008

Bio Treatment Influent

D

8.1

9/22/2008

Bio Treatment Effluent

ND

1

9/29/2008

Bio Treatment Influent

D

3.4

9/29/2008

Bio Treatment Effluent

D

3.7

12/3/2008

FGD Purge

D

16000

12/3/2008

Bio Treatment Influent

D

19

12/3/2008

Bio Treatment Effluent

D

3.6

12/10/2008

FGD Purge

D

44000

12/10/2008

Bio Treatment Influent

D

15

12/10/2008

Bio Treatment Effluent

D

2

12/17/2008

Bio Treatment Influent

D

7.6

12/17/2008

Bio Treatment Effluent

D

2.9

12/22/2008

Bio Treatment Influent

D

20

2-8

-------
Table A2.2

List of excluded data at Belews Creek (continued)

Data Source

Pollutant
(unit)

Date

Sampling Location

Indicator

Concentration

Exclusion Rationale

Plant Self-
Monitoring
(cont'd)



12/22/2008

Bio Treatment Effluent

D

1.4

Questionable results due to QA/QC issues

EPA excluded mercury data that Duke Energy identified
as questionable due to quality control issues. Duke
Energy stated that these data are questionable because
they suspected that the matrix spike/matrix spike
duplicate (MS/MSD) were not conducted using the FGD
wastewater matrix or the MS/MSD indicated inadequate
recoveries. EPA lacked sufficient laboratory quality
control data for these results to make an independent
assessment of whether these mercury results should
have been retained and included in the statistical
analyses, (cont'd)

12/29/2008

Bio Treatment Influent

D

12

12/29/2008

Bio Treatment Effluent

D

2.4

1/7/2009

Bio Treatment Influent

D

5

1/7/2009

Bio Treatment Effluent

D

2.8

1/14/2009

Bio Treatment Influent

D

15

1/14/2009

Bio Treatment Effluent

D

2.3

1/22/2009

Bio Treatment Influent

D

4.7

1/22/2009

Bio Treatment Effluent

D

3.9

1/28/2009

Bio Treatment Influent

D

12

1/28/2009

Bio Treatment Effluent

D

2.6

Plant Self-
Monitoring

Mercury (ng/L)

6/12/2008-
6/15/2008

Bio Treatment Effluent

ND

1000

Insufficiently^sensitive analytical method (Method
245.1)

See Analytical Methods Review Memo

1/27/2010

Bio Treatment Effluent

ND

1000

2/24/2010

Bio Treatment Effluent

ND

1000

3/24/2010

Bio Treatment Effluent

ND

1000

4/7/2010

Bio Treatment Effluent

ND

1000

6/23/2010

Bio Treatment Effluent

ND

1000

7/14/2010

Bio Treatment Effluent

ND

1000

7/28/2010

Bio Treatment Effluent

ND

1000

8/25/2010

Bio Treatment Effluent

ND

1000

9/22/2010

Bio Treatment Effluent

ND

1000

10/27/2010

Bio Treatment Effluent

ND

1000

11/4/2010

Bio Treatment Effluent

ND

1000

11/23/2010

Bio Treatment Effluent

ND

1000

12/8/2010

Bio Treatment Effluent

ND

1000

1/17/2011

Bio Treatment Effluent

ND

1000

1/26/2011

Bio Treatment Effluent

ND

1000

2/16/2011

Bio Treatment Effluent

ND

1000

3/23/2011

Bio Treatment Effluent

ND

1000

4/27/2011

Bio Treatment Effluent

ND

1000

2-9

-------
Table A2.2

List of excluded data at Belews Creek (continued)

Data Source

Pollutant
(unit)

Date

Sampling Location

Indicator

Concentration

Exclusion Rationale

Plant Self-
Monitoring
(cont'd)

Mercury (ng/L)
(cont'd)

5/25/2011

Bio Treatment Effluent

ND

1000

Insufficiently^sensitive analytical method (Method
245.1)

See Analytical Methods Review Memo
(cont'd)

6/8/2011

Bio Treatment Effluent

ND

1000

6/22/2011

Bio Treatment Effluent

ND

1000

7/13/2011

Bio Treatment Effluent

ND

1000

7/27/2011

Bio Treatment Effluent

ND

1000

8/10/2011

Bio Treatment Effluent

ND

1000

8/24/2011

Bio Treatment Effluent

ND

1000

9/14/2011

Bio Treatment Effluent

ND

1000

9/28/2011

Bio Treatment Effluent

ND

1000

10/7/2011

Bio Treatment Effluent

ND

1000

10/8/2011

Bio Treatment Effluent

ND

1000

10/9/2011

Bio Treatment Effluent

ND

1000

10/10/2011

Bio Treatment Effluent

ND

1000

10/11/2011

Bio Treatment Effluent

ND

1000

10/12/2011

Bio Treatment Effluent

ND

1000

10/13/2011

Bio Treatment Effluent

ND

1000

10/14/2011

Bio Treatment Effluent

ND

1000

10/15/2011

Bio Treatment Effluent

ND

1000

10/16/2011

Bio Treatment Effluent

ND

1000

10/26/2011

Bio Treatment Effluent

ND

1000

11/4/2011

Bio Treatment Effluent

ND

1000

11/6/2011

Bio Treatment Effluent

ND

1000

11/7/2011

Bio Treatment Effluent

ND

1000

11/8/2011

Bio Treatment Effluent

ND

1000

11/9/2011

Bio Treatment Effluent

ND

1000

11/10/2011

Bio Treatment Effluent

ND

1000

11/11/2011

Bio Treatment Effluent

ND

1000

11/12/2011

Bio Treatment Effluent

ND

1000

11/13/2011

Bio Treatment Effluent

ND

1000

11/23/2011

Bio Treatment Effluent

ND

1000

12/14/2011

Bio Treatment Effluent

ND

1000

12/28/2011

Bio Treatment Effluent

ND

1000

2-10

-------
Table A2.2

List of excluded data at Belews Creek (continued)

Data Source

Pollutant
(unit)

Date

Sampling Location

Indicator

Concentration

Exclusion Rationale

Plant Self-
Monitoring
(cont'd)

Mercury (ng/L)
(cont'd)

1/11/2012

Bio Treatment Effluent

ND

1000

Insufficiently^sensitive analytical method (Method
245.1)

See Analytical Methods Review Memo
(cont'd)

1/25/2012

Bio Treatment Effluent

ND

1000

2/8/2012

Bio Treatment Effluent

ND

1000

2/15/2012

Bio Treatment Effluent

ND

2500

2/16/2012

Bio Treatment Effluent

ND

1000

2/17/2012

Bio Treatment Effluent

ND

1000

2/18/2012

Bio Treatment Effluent

ND

1000

2/19/2012

Bio Treatment Effluent

ND

1000

2/20/2012

Bio Treatment Effluent

ND

1000

2/21/2012

Bio Treatment Effluent

ND

1000

2/22/2012

Bio Treatment Effluent

ND

1000

2/23/2012

Bio Treatment Effluent

ND

1000

2/24/2012

Bio Treatment Effluent

ND

1000

3/14/2012

Bio Treatment Effluent

ND

1000

3/28/2012

Bio Treatment Effluent

ND

1000

4/25/2012

Bio Treatment Effluent

ND

1000

5/22/2012

Bio Treatment Effluent

ND

1000

5/23/2012

Bio Treatment Effluent

ND

1000

5/28/2012

Bio Treatment Effluent

ND

1000

5/30/2012

Bio Treatment Effluent

ND

1000

5/31/2012

Bio Treatment Effluent

ND

1000

6/1/2012

Bio Treatment Effluent

ND

1000

6/13/2012

Bio Treatment Effluent

ND

1000

6/27/2012

Bio Treatment Effluent

ND

1000

8/8/2012

Bio Treatment Effluent

ND

1000

8/29/2012

Bio Treatment Effluent

ND

1000

9/12/2012

Bio Treatment Effluent

ND

1000

9/26/2012

Bio Treatment Effluent

ND

1000

10/10/2012

Bio Treatment Effluent

ND

1000

1/9/2013

Bio Treatment Effluent

ND

1000

1/23/2013

Bio Treatment Effluent

ND

1000

2/13/2013

Bio Treatment Effluent

ND

1000

2/27/2013

Bio Treatment Effluent

ND

1000

2-11

-------
Table A2.2 List of excluded data at Belews Creek (continued)

Data Source

Pollutant
(unit)

Date

Sampling Location

Indicator

Concentration

Exclusion Rationale

Plant Self-
Monitoring



3/13/2013

Bio Treatment Effluent

ND

1000



Mercury (ng/L)

5/9/2012

Bio Treatment Effluent

D

2500

Data Entry Error

The reported value of 2500 ng/L is inconsistent with the
other reported observation for that day (22.1 ng/L).
Additionally, the 2500 ng/L observation is orders of
magnitude higher than other observations for Bio
Treatment Effluent in the preceding and following
months, while the 22.1 ng/L observation for 5/9/2012
is consistent with observations reported for the
preceding and following months. Therefore, the 2500
ng/L observation was excluded as an analytical anomaly
or data entry error.

EPA

Sampling

Mercury (ng/L)

6/9/2010

Bio Treatment Effluent
Dup

D

6440

Analytical anomalies

The reported value of 6440 ng/L mercury is inconsistent
with the other reported observations for that day (247
ng/L for EPA split sample; 333 ng/L for industry self-
monitoring split sample). Additionally, the 6440 ng/L
observation is an order of magnitude higher than other
observations for Bio Treatment Effluent in the preceding
months and following days and months, in contrast to
the other observations for 6/9/2010. Therefore, the
6440 ng/L observation was excluded as an analytical
anomaly.

2-12

-------
Table A2.2 List of excluded data at Belews Creek (continued)

Data Source

Pollutant
(unit)

Date

Sampling Location

Indicator

Concentration

Exclusion Rationale

Plant Self-
Monitoring

Selenium
(Mg/L)

7/14/2010

Bio Treatment Effluent

D

299

Upset condition or analytical anomaly

The reported value of 299 ug/L selenium is much higher
than any other observation in the remainder of dataset
for Belews Creek. EPA evaluated the selenium
contributions in the FGD purge and bioreactor influent
(total and dissolved selenium, as well as concentrations
of selenate and selenite), and the values for total
selenium, selenite, and selenate in the bioreactor
effluent (there are no dissolved selenium results for the
bioreactor effluent on that date). In addition, EPA
evaluated the concentrations for total selenium,
dissolved selenium, selenite, and selenate at all three
sampling locations for the months preceding and
following the July 14 sampling event. EPA also
evaluated electricity generation, absorber and
wastewater pH data, bioreactor influent ORP, and
bioreactor influent and effluent nitrate concentrations
for 7/14/2010 and the months preceding and following
that date for indications of changes or abnormalities.

There were no significant changes in ORP to suggest the
wastewater characteristics reflected the effects of high
oxidizing conditions in the scrubber, so the commenter's
suggestion in that regard is unlikely. EPA noted that the
bioreactor influent nitrate concentrations rose
somewhat in the days preceding the observation;
however, the increase was not large and the bioreactor
effluent nitrate concentrations were non-detect at <2.3
mg/L, showing that the bioreactor was effectively
denitrifying the wastewater.

2-13

-------
Table A2.2 List of excluded data at Belews Creek (continued)

Data Source

Pollutant
(unit)

Date

Sampling Location

Indicator

Concentration

Exclusion Rationale

Plant Self-
Monitoring
(cont'd)











The high effluent selenium observation also coincided
with elevated selenate in the effluent (217 ug/L). There
were also substantial increases in total selenium and
selenate in the bioreactor influent and the FGD purge,
showing that the bioreactor did experience an increased
pollutant load. However, even taking into account the
concentration of selenate present in the FGD purge and
bioreactor influent samples, the bioreactor effluent is
substantially higher than would be expected based on
the demonstrated performance of the Belews Creek
biological treatment system, as well as the performance
observed for the biological treatment system at Allen.
For example, the selenium in the bioreactor influent on
7/14/2010 was less than 3 times higher than the
influent concentration for the previous observation on
6/23/2010; yet when the bioreactor influent selenium
load increased by 10 times in May 2011 there was little
effect on effluent concentrations (rising from <5 ug/L to
7/7 ug/L). Similar examples can be found at Belews
Creek for selenate, where a large change in selenate
(relative to the change observed for 7/14/2010) was
not accompanied by increased effluent selenium or
selenate. See, e.g., data for October 2011 and
December 2012.

These observations for Belews Creek are supported by
the data for the biological treatment system at Allen.
There were multiple occasions at Allen when the
bioreactor influent selenium and selenate increased
sharply, with the concentrations rising 10-15 times
higher in a short period of time, yet with little if any
observable effect on the bioreactor effluent
concentrations. See, e.g., Allen data for August 2010,
March 2011, December 2011, September 2012, and

2-14

-------
Table A2.2 List of excluded data at Belews Creek (continued)

Data Source

Pollutant
(unit)

Date

Sampling Location

Indicator

Concentration

Exclusion Rationale

Plant Self-
Monitoring
(cont'd)











March 2013. In addition, the influent selenium
concentrations at Allen in September 2012 and March
2013 were even higher than observed at Belews Creek
on 7/14/2010.

The 299 ug/L observation is 10 times higher than any of
the other 398 observations for the treatment systems at
Belews Creek and Allen, which represent more than 5
years of operation at Belews Creek and more than 4
years of operation at Allen. Excluding the 7/14/2010
observation, there are 216 observations for selenium in
the bioreactor effluent; the maximum value is 29.4 ug/L
and the median is 6.2 ug/L (i.e., half of all data are
lower than 6.2 ug/L). There are 182 observations for
selenium in the bioreactor effluent at Allen; the
maximum value is 27.6 ug/L and the median is 5.0
ug/L.

Based upon a thorough review of the bioreactor
performance data for Allen and Belews Creek, EPA
concluded that the 7/14/2010 observation should be
excluded when calculating the effluent limitations for
the final rule because the unusually high value is the
result of a treatment system upset condition or other
abnormal operation, such as would result from
inadequate operator control of the biological process.
This is consistent with information EPA obtained during
a site visit to Allen Station on October 22, 2009:

"Allen stated that the major improvement of the Allen
biological treatment system over the Belews Creek
system is the inclusion of oxidation-reduction potential
(ORP) and pH probes on each individual cell (the Belews
Creek system has one ORP probe for each stage).

2-15

-------
Table A2.2 List of excluded data at Belews Creek (continued)

Data Source

Pollutant
(unit)

Date

Sampling Location

Indicator

Concentration

Exclusion Rationale

Plant Self-
Monitoring
(cont'd)











Additionally, the Allen system has sampling points on
the effluent from the second stage (Belews Creek has
sampling points on the effluent from the second stage
recycle). These improvements have resulted in greater
ease of operations for the biological treatment system
at Allen." DCN EPA-HQ-OW-2009-0819-0598

A properly operated treatment system includes close
operator attention to influent wastewater
characteristics, key process control parameters and
effluent quality. See DCN SE05846 (memorandum
Variability in Flue Gas Desulfurization Wastewater:
Monitoring and Response).

Since no plausible explanation validating the extreme
observation could be determined, EPA excluded the
bioreactor effluent selenium observation for July 14,
2010.

1 D=Detected, ND=Non-detected.

2-16

-------
Table A2.3

List of excluded data at Hatfield's Ferry

Data
Source

Pollutant
(unit)

Date

Sampling
Location

Indicator1

Concentration

Exclusion Rationale

Plant Self-
Monitoring

Arsenic
(Mg/L)

12/6/2011
to

10/1/2013

Chem

Precipitation
Effluent





Analytical Interferences

See Analytical Methods Review Memo

Plant Self-
Monitoring

Mercury
(ng/L)

2/17/2012

Chem

Precipitation
Effluent

D

35.4

Data Entry Error

EPA excluded this observation because conflicting information results
in uncertainty about the correct value. FirstEnergy reported conflicting
information in its 5/2/2014 and 5/13/2014 submittals.

Plant Self-
Monitoring

Arsenic,
Mercury

10/8/2013
to

12/10/2013

Chem

Precipitation
Effluent





Decommissioning Period
See Analytical Methods Review Memo

Plant Self-
Monitoring

Mercury
(ng/L)

11/29/2012

Chem

Precipitation
Effluent

ND

200

Insufficiently-sensitive analytical method (Method 245.1)
See Analytical Methods Review Memo

12/19/2012

D

270

12/21/2012

D

430

CWA 308
Sampling

Mercury
(ng/L)

10/5/2010

Chem

Precipitation
Effluent

D

978

Sampling or Analytical Error.

EPA excluded mercury data collected on 10/5/10 at the chemical
precipitation effluent at Hatfield's Ferry. This exclusion was due to an
unusually high analytical result inconsistent with proper operation of
the BAT/NSPS chemical precipitation treatment technology.
Specifically, based on information in the record, the sampling result
indicates that the sample may have been contaminated during sample
collection or was adversely affected by treatment system upset. The
decision to exclude the analytical result (978 ng/L) is supported by the
plant's NPDES compliance monitoring data for that time period. The
maximum daily value reported for FGD wastewater treatment system
effluent for the month of October 2010 was 127 ng/L. If the excluded
value accurately represented the mercury concentration in the plant's
FGD treatment system effluent, due to the hydraulic residence times
and mixing that occurs in the FGD absorber and the wastewater
treatment system, similarly high concentrations of mercury should
have been reflected in the plant's NPDES compliance monitoring
samples. Furthermore, the maximum daily values reported by the
plant for each monthly reporting period for August 2010 through

2-17

-------
Table A2.3 List of excluded data at Hatfield's Ferry (continued)

Data
Source

Pollutant
(unit)

Date

Sampling
Location

Indicator1

Concentration

Exclusion Rationale

CWA 308
Sampling
(cont'd)











February 2011 were also substantially lower than the excluded value,
ranging from 108 ng/Lto 268 ng/L. The excluded value may be the
result of sample contamination occurring during sample collection.
EPA used "clean sampling" techniques for its field sampling program,
including requiring plant staff to also use such procedures when
collecting the CWA 308 samples. These clean sampling procedures
are intended to minimize the potential for contaminating samples,
particularly in environments where dust and other contaminants are
present. For all plants in the field sampling program, except Hatfield's
Ferry, the CWA 308 sampling was conducted after EPA had first
conducted its on-site field sampling program. This approach enabled
plant staff to become familiar with the specialized sampling collection
protocols used for "clean sampling" prior to initiating their CWA 308
sampling. However, in the case of Hatfield's Ferry, the plant initiated
the CWA 308 sampling prior to the EPA field sampling and
unfamiliarity with the clean sampling protocols may have allowed the
October 2010 sample to become contaminated. If the excluded result
is not associated with sample contamination, then it indicates that the
plant did not exercise good control of the treatment system during
startup following a shutdown period. All three generating units had
been shutdown for at least a week preceding the December 2010
sampling event. Unit 3 then started up on October 3, generating
electricity at approximately one-third capacity for the unit. Unit 3 then
operated at nearly full load on October 4 and approximately two-thirds
load on October 5, but then shutdown again on October 6. Based on
the information in the record, the FGD wastewater treatment system
likely was not operating in a stable manner during this period. In their
comments on the proposed effluent guidelines, FirstEnergy (operator
for Hatfield's Ferry) did not dispute EPA's determination that the 978
ng/L observation should be excluded, nor the reasons for doing so.

2-18

-------
Table A2.3

List of excluded data at Hatfield's Ferry (continued)

Data
Source

Pollutant
(unit)

Date

Sampling
Location

Indicator1

Concentration

Exclusion Rationale

Plant Self-
Monitoring

Arsenic
(Mg/L)

1/5/2010

Chem

Precipitation
Effluent

D

8

Treatment system upset or abnormal operation
See Figure A5.1 in Appendix 5

Observations during the time period including 1/5/2010 to
1/19/2010 are an order of magnitude higher than the preceding and
following weeks. FirstEnergy stated that they do not know what caused
the large spikes in mercury concentrations. Since the treatment
system was undergoing an upset or other abnormal operation, all
results for the affected days were excluded.

Observations during the time period including 2/28/2012 to
3/27/2012 are substantially higher (hundreds or thousands ppt) than
observations for the preceding and following weeks. FirstEnergy stated
that they do not know what caused the large spikes in mercury
concentrations. FirstEnergy noted that they occasionally had problems
with the sand filters. FirstEnergy also noted that the wastewater
treatment operators were changing frequently during the time period
because of training and that it could have led to multiple operators
making various changes that may have impacted the performance of
the treatment system.

Observations for the time period including 3/19/2013 to 6/4/2013
are substantially higher than observations for the preceding and
following periods. FirstEnergy was unable to explain the reason for the
spikes in mercury concentration and variable performance of the
treatment system, particularly in relation to the system's typical
performance. There were no changes in power plant operation, such as
changes in coal source or operation of FGD or other systems,
coinciding with the time period.

1/12/2010

D

7

1/19/2010

D

9

Mercury
(ng/L)

1/5/2010

D

1440

1/12/2010

D

3310

1/19/2010

D

2030

2/28/2012

D

1520

3/6/2012

D

466

3/13/2012

D

1160

3/20/2012

D

1710

3/27/2012

D

1320

4/3/2012

D

3400

3/19/2013

D

1430

3/26/2013

D

194

4/3/2013

D

2110

5/14/2013

D

1410

5/21/2013

D

573

5/29/2013

D

518

6/4/2013

D

1900

D=Detected, ND=Non-detected.

2-19

-------
Table A2.4 List of excluded data at Keystone

Pollutant
(unit)

Date

Sampling Location

Indicator1

Concentration

Exclusion Rationale

Arsenic
(Mg/L)

1/3/2012 to
1/6/2014

Chem Precipitation
Effluent





Insufficiently^sensitive analytical method (Method 200.7)
See Analytical Methods Review Memo

Mercury
(ng/L)

1/3/2012 to
4/28/2014

Chem Precipitation
Effluent





Insufficiently^sensitive analytical method (Method SM 3112B)
See the Analytical Methods Review Memo

Arsenic
(Mg/L)

2/21/2013

Chem Precipitation
Effluent

ND

5

Not an Approved NPDES Compliance Monitoring Method
See the Analytical Methods Review Memo

1 D=Detected, ND=Non-detected.

2-20

-------
Table A2.5 List of excluded plant self-monitoring data at Miami Fort

Pollutant (unit)

Date

Sampling
Location

Indicator

Concentration

Exclusion Rationale

Arsenic (ng/L)

7/13/2009

Chem

Precipitation
Effluent

ND

50

insufficiently-sensitive analytical method (Method 200.8)
See the Analytical Methods Review Memo.

8/3/2009

ND

20

9/1/2009

ND

20

10/5/2009

ND

20

11/2/2009

ND

20

12/1/2009

ND

20

1/5/2010

ND

20

2/2/2010

ND

20

3/2/2010

ND

40

4/6/2010

ND

20

5/4/2010

ND

20

6/1/2010

ND

20

7/6/2010

ND

20

8/3/2010

ND

20

9/7/2010

ND

20

10/5/2010

ND

20

11/2/2010

ND

20

12/7/2010

ND

20

12/7/2010

ND

20

1/4/2011

ND

20

2/1/2011

ND

20

3/1/2011

ND

20

4/5/2011

ND

20

5/9/2011

ND

20

6/7/2011

ND

20

7/5/2011

ND

20

8/2/2011

ND

20

9/6/2011

ND

20

10/4/2011

ND

20

11/1/2011

ND

20

12/6/2011

ND

20

1/3/2012

ND

20

2/7/2012

ND

20

2-21

-------
Table A2.5

List of excluded plant self-monitoring data at Miami Fort (continued)

Pollutant (unit)

Date

Sampling
Location

Indicator

Concentration

Exclusion Rationale

Arsenic (ng/L)
(cont'd)

3/6/2012



ND

20

insufficiently-sensitive analytical method (Method 200.8)

See the Analytical Methods Review Memo.

(cont'd)

4/3/2012

ND

20

5/1/2012

ND

20

6/5/2012

ND

20

7/2/2012

ND

20

8/7/2012

ND

20

9/4/2012

ND

20

10/2/2012

ND

20

11/6/2012

ND

20

12/4/2012

ND

20

1/2/2013

ND

20

2/5/2013

ND

20

3/5/2013

ND

20

4/2/2013

ND

20

5/7/2013

ND

20

6/4/2013

ND

20

7/2/2013

ND

20

8/6/2013

ND

20

9/3/2013

ND

20

10/1/2013

ND

20

11/5/2013

ND

20

12/3/2013

ND

20

Arsenic (ng/L)

5/8/2013

Chem

Precipitation
Effluent

ND

5

Unable to Confirm the Validity of the Results
See the Analytical Methods Review Memo

5/8/2013

ND

5

5/8/2013

D

46

5/8/2013

ND

500

2-22

-------
Table A2.5

List of excluded plant self-monitoring data at Miami Fort (continued)

Pollutant (unit)

Date

Sampling
Location

Indicator

Concentration

Exclusion Rationale

Arsenic (ng/L)

7/31/2012

Chem

Precipitation
Effluent

D

5.44

Unable to validate the results due to contradictory information

Contradictory information about these observations make them unusable. Originally,
Duke reported all four results as ND, with first 7/31/2012 and 8/1/2012 at 20 ug/L
and 8/8/2012 and 8/9/2012 at 10 ug/L. UWAG reported all four results as ND at 10
ug/L. When Duke was asked to verify discrepancy for first two results, they instead
reported all four observations as detected with new values (5.44, 5.86, 7.29, and 6.49
ug/L). Furthermore, Duke's spreadsheet included text showing that the samples were
altered (diluted) in the field during collection. [["Special field sample prep. Sample was
diluted in field by URS and diluted in lab by Applied Speciation." Note that Duke did not
provide supporting documentation for dilutions and corrections to results.]] Based on
this conflicting information and concerns about field-altered samples, all results for the
four samples are considered invalid.

7/31/2012

ND

10

8/1/2012

D

5.86

8/1/2012

ND

10

8/8/2012

D

7.29

8/8/2012

ND

10

8/9/2012

D

6.49

8/9/2012

ND

10

1D=Detected, ND=Non-detected.

2-23

-------
Table A2.6 List of excluded plant self-monitoring data at Pleasant Prairie

Pollutant
(unit)

Date

Sampling
Location

Indicator1

Concentration

Exclusion Rationale

Mercury
(ng/L)

11/7/2007
to

4/15/2009

Chem

Precipitation
Effluent





Associated with TMT 15

Associated with TMT-15 organosulfide usage, which We Energies determined is less
effective at removing mercury from Pleasant Prairie wastewater than other organosulfide
formulations (i.e., Nalmet, Metclear). Pleasant Prairie switched from TMT-15 to Nalmet
during the week of March 30, 2009. Effluent mercury concentrations rapidly decreased
over an initial transition period, which based on review of the effluent data and
engineering judgment continued into mid-April. As a result, mercury data associated with
TMT-a5 use does not reflect BAT/NSPS operation of the treatment system at Pleasant
Prairie and the data were excluded.

Arsenic
(Mg/L)

11/13/2007

Chem

Precipitation
Effluent

D

21

Problem associated with start-up

Elevated concentration associated with initial start-up of the treatment system. The FGD
system was operated in recirculation mode for an extended period of time as chlorides
were allowed to build up in the scrubber and the wastewater treatment system was tuned
to improve mercury removals. Discharge did not begin until November 2007 (and even
then was irregular) and wastewater pollutant concentrations were elevated due to the
extended recirculation period. The extended recirculation of the FGD wastewater resulted
in fines building up in the system, which were then difficult for the plant to remove in the
clarifier, resulting in elevated effluent concentrations. Data from this period do not reflect
BAT/NSPS operation of the treatment system and were excluded.

Arsenic,
Mercury

10/4/2007
to 5/7/2013

Secondary

Clarifier

Effluent





Secondary clarifier sampling location not representative of effluent discharge. (Certain
mercury data 10/4/2007 to 4/15/2009 are also associated with TMT-15 additive, which
is not representative of BAT/NSPS operation.)

Much of the secondary clarifier data is associated with time periods when the plant was
not discharging FGD wastewater. This is evident because the plant is required to monitor
the discharge twice per week, and there are many time periods where there are purge and
secondary clarifier samples, but no chem precip effluent. Furthermore, having secondary
clarifier samples but no effluent samples is consistent with the plant's operational
practice of recirculating the FGD wastewater when it is not in spec. Since EPA is unable to
positively determine for each secondary clarifier sample whether it is associated with a
discharge event or recirculation, all secondary clarifier data were excluded. (Note that
numerous secondary clarifier samples exceed the permit effluent limits.)

2-24

-------
Table A2.6

List of excluded plant self-monitoring data at Pleasant Prairie (continued)

Pollutant
(unit)

Date

Sampling
Location

Indicator1

Concentration

Exclusion Rationale

Mercury

6/13/2012

Chem

Precipitation
Effluent

D

2900

Treatment system upset or abnormal operation

Permit exceedance (out of compliance), due to treatment system upset/abnormal
operation.

The time period for these observations, plus a 8300 ng/L secondary clarifier result,
includes an exceedance of the NPDES permit limit and other observations of high effluent
mercury concentrations. According to We Energies, the higher concentrations are
associated with startup following a dual-unit outage and a 4-week FGD system shutdown.
We Energies stated that storage of the FGD absorber slurry in a holding tank for 4 weeks
without aeration led to reduction and dissolution of some of the oxidized particulate
mercury, resulting in an abnormally high proportion of dissolved mercury in the FGD
blowdown. The treatment system ultimately was operated in recirculation mode to aid in
precipitating the mercury. (Note: Situation could have been avoided/mitigated by either
providing aeration to the holding tank or placing the treatment system in recirculation
until effluent concentrations were at typical levels.)

6/27/2012

D

595

6/28/2012

D

640

7/2/2012

D

490

7/3/2012

D

555

1 D=Detected, ND=Non-detected.

2-25

-------
Table A2.7

List of excluded CWA 308 sampling data at Wabash River

Pollutant
(unit)

Date

Sampling Location

Indicator

Concentration

Exclusion Rationale

Arsenic
(Mg/L)

2/21/2011

Sour Water Treatment Influent

ND

4.1

Failed data editing criteria

While performing the LTA test, EPA found that all the datasets
passed the LTA test except for arsenic and mercury data at Wabash
River. Thus, data for arsenic and mercury at Wabash River were
excluded from the calculation of the limits. See Section 4 for
discussion of the data editing criteria.

2/21/2011

Steam Stripper Effluent

ND

4.2

2/21/2011

Vapor Compression Evaporator
Influent

D

5

2/21/2011

Vapor Compression Evaporator
Condensate

ND

4

2/22/2011

Sour Water Treatment Influent

ND

4.1

2/22/2011

Steam Stripper Effluent

ND

4.2

2/22/2011

Vapor Compression Evaporator
Influent

D

5

2/22/2011

Vapor Compression Evaporator
Condensate

ND

4

2/23/2011

Sour Water Treatment Influent

ND

4.1

2/23/2011

Steam Stripper Effluent

ND

4.1

2/23/2011

Vapor Compression Evaporator
Influent

ND

4

2/23/2011

Vapor Compression Evaporator
Condensate

ND

4

2/24/2011

Sour Water Treatment Influent

ND

4.1

2/24/2011

Steam Stripper Effluent

ND

4.1

2/24/2011

Vapor Compression Evaporator
Influent

ND

4

2/24/2011

Vapor Compression Evaporator
Condensate

ND

4

Mercury
(ng/L)

2/21/2011

Sour Water Treatment Influent

D

7420

2/21/2011

Steam Stripper Effluent

D

46.3

2/21/2011

Vapor Compression Evaporator
Influent

ND

9.9

2/21/2011

Vapor Compression Evaporator
Condensate

ND

0.5

2/22/2011

Sour Water Treatment Influent

D

10.8

2/22/2011

Steam Stripper Effluent

D

24.4

2-26

-------
Table A2.7

List of excluded CWA 308 sampling data at Wabash River (continued)

Pollutant
(unit)

Date

Sampling Location

Indicator

Concentration

Exclusion Rationale

Mercury

(ng/L)

(cont'd)

2/22/2011

Vapor Compression Evaporator
Influent

ND

4.95



2/22/2011

Vapor Compression Evaporator
Condensate

ND

0.5

2/23/2011

Sour Water Treatment Influent

D

48.3

2/23/2011

Steam Stripper Effluent

ND

9.9

2/23/2011

Vapor Compression Evaporator
Influent

ND

9.9

2/23/2011

Vapor Compression Evaporator
Condensate

ND

0.5

2/24/2011

Sour Water Treatment Influent

ND

9.9

2/24/2011

Steam Stripper Effluent

ND

9.9

2/24/2011

Vapor Compression Evaporator
Influent

ND

9.9

2/24/2011

Vapor Compression Evaporator
Condensate

ND

0.5

1D=Detected, ND=Non-detected.

2-27

-------
Appendix 3. Plots of Plant Self-monitoring Data with

Smoothed Curves to Aid in Determining
Initial Commissioning Period for the
Treatment System

3-1

-------
Table of Contents

Figures	Page

A3.1.1 Plot of all self-monitoring data for arsenic (jag/L) for Allen. The

red vertical line indicates the end of the commissioning period	 3-7

A3.1.2 Plot of all self-monitoring data for arsenic (jag/L) for Allen
superimposed with smooth curves using LOWESS. The red

vertical line indicates the end of the commissioning period	 3-8

A3.1.3 Plot of self-monitoring data for arsenic (jag/L) for Allen after

exclusions other than commissioning period described in Section
3.2. The red vertical line indicates the end of the commissioning

period	 3-9

A3.1.4 Plot of self-monitoring data for arsenic (jag/L) for Allen after

exclusions other than commissioning period described in Section
3.2 and superimposed with smooth curves using LOWESS. The

red vertical line indicates the end of the commissioning period	 3-10

A3.1.5 Plot of all self-monitoring data for mercury (ng/L) for Allen. The

red vertical line indicates the end of the commissioning period	 3-11

A3.1.6 Plot of all self-monitoring data for mercury (ng/L) for Allen
superimposed with smooth curves using LOWESS. The red

vertical line indicates the end of the commissioning period	 3-12

A3.1.7 Plot of self-monitoring data for mercury (ng/L) for Allen after

exclusions other than commissioning period described in Section
3.2. The red vertical line indicates the end of the commissioning

period	 3-13

A3.1.8 Plot of self-monitoring data for mercury (ng/L) for Allen after

exclusions other than commissioning period described in Section
3.2 and superimposed with smooth curves using LOWESS. The

red vertical line indicates the end of the commissioning period	 3-14

A3.1.9 Plot of all self-monitoring data for Nitrate-Nitrite as N (mg/L)

for Allen	 3-15

A3.1.10 Plot of all self-monitoring data for Nitrate-Nitrite as N (mg/L)

for Allen superimposed with smooth curves using LOWESS	 3-16

A3.1.11 Plot of self-monitoring data for Nitrate-Nitrite as N (mg/L) for
Allen after exclusions other than commissioning period described

in Section 3.2	 3-17

A3.1.12 Plot of self-monitoring data for Nitrate-Nitrite as N (mg/L) for
Allen after exclusions other than commissioning period described
in Section 3.2 and superimposed with smooth curves using

LOWESS	 3-18

A3.1.13 Plot of all self-monitoring data for selenium (|j,g/L) for Allen.

The red vertical line indicates the end of the commissioning

period	 3-19

3-2

-------
Table of Contents
(continued)

Figures	Page

A3.1.14 Plot of all self-monitoring data for selenium (jag/L) for Allen
superimposed with smooth curves using LOWESS. The red

vertical line indicates the end of the commissioning period	 3-20

A3.1.15 Plot of self-monitoring data for selenium (jag/L) for Allen after
exclusions other than commissioning period described in Section
3.2. The red vertical line indicates the end of the commissioning

period	 3-21

A3.1.16 Plot of self-monitoring data for selenium (jag/L) for Allen after
exclusions other than commissioning period described in Section
3.2 and superimposed with smooth curves using LOWESS. The

red vertical line indicates the end of the commissioning period	 3-22

A3.1.17 Plot of all self-monitoring data for selenium (jag/L) for Allen,

bioreactor effluent only. The red vertical line indicates the end of

the commissioning period	 3-23

A3.1.18 Plot of all self-monitoring data for selenium (jag/L) for Allen
superimposed with smooth curves using LOWESS, bioreactor
effluent only. The red vertical line indicates the end of the

commissioning period	 3-24

A3.1.19 Plot of self-monitoring data for selenium (jag/L) for Allen after
exclusions other than commissioning period described in Section
3.2, bioreactor effluent only. The red vertical line indicates the

end of the commissioning period	 3-25

A3.1.20 Plot of self-monitoring data for selenium (jag/L) for Allen after
exclusions other than commissioning period described in Section
3.2 and superimposed with smooth curves using LOWESS,,
bioreactor effluent only. The red vertical line indicates the end of
the commissioning period	 3-26

A3.2.1 Plot of all self-monitoring data for arsenic (jJg/L) for Belews
Creek. The red vertical line indicates the end of the

commissioning period	 3-27

A3.2.2 Plot of all self-monitoring data for arsenic (jJg/L) for Belews
Creek superimposed with smooth curves using LOWESS. The

red vertical line indicates the end of the commissioning period	 3-28

A3.2.3 Plot of self-monitoring data for arsenic (jJg/L) for Belews Creek
after exclusions other than commissioning period described in
Section 3.2. The red vertical line indicates the end of the

commissioning period	 3-29

3-3

-------
Table of Contents
(continued)

Figures	Page

A3.2.4 Plot of self-monitoring data for arsenic (ng/L) for Belews Creek
after exclusions other than commissioning period described in
Section 3.2 and superimposed with smooth curves using
LOWESS. The red vertical line indicates the end of the

commissioning period	 3-30

A3.2.5 Plot of all self-monitoring data for mercury (ng/L) for Belews
Creek. The red vertical line indicates the end of the

commissioning period	 3-31

A3.2.6 Plot of all self-monitoring data for mercury (ng/L) for Belews
Creek superimposed with smooth curves using LOWESS. The

red vertical line indicates the end of the commissioning period	 3-32

A3.2.7 Plot of self-monitoring data for mercury (ng/L) for Belews Creek
after exclusions other than commissioning period described in
Section 3.2. The red vertical line indicates the end of the

commissioning period	 3-33

A3.2.8 Plot of self-monitoring data for mercury (ng/L) for Belews Creek
after exclusions other than commissioning period described in
Section 3.2 and superimposed with smooth curves using
LOWESS. The red vertical line indicates the end of the

commissioning period	 3-34

A3.2.9 Plot of all self-monitoring data for Nitrate-Nitrite as N (mg/L)

for Belews Creek	 3-35

A3.2.10 Plot of all self-monitoring data for Nitrate-Nitrite as N (mg/L)
for Belews Creek superimposed with smooth curves using

LOWESS	 3-36

A3.2.11 Plot of self-monitoring data for Nitrate-Nitrite as N (mg/L) for
Belews Creek after exclusions other than commissioning period

described in Section 3.2	 3-37

A3.2.12 Plot of self-monitoring data for Nitrate-Nitrite as N (mg/L) for
Belews Creek after exclusions other than commissioning period
described in Section 3.2 and superimposed with smooth curves

using LOWESS	 3-38

A3.2.13 Plot of all self-monitoring data for selenium (ng/L) for Belews
Creek. The red vertical line indicates the end of the

commissioning period	 3-39

A3.2.14 Plot of all self-monitoring data for selenium (ng/L) for Belews
Creek superimposed with smooth curves using LOWESS. The
red vertical line indicates the end of the commissioning period	 3-40

3-4

-------
Table of Contents
(continued)

Figures	Page

A3.2.15 Plot of self-monitoring data for selenium (jag/L) for Belews
Creek after exclusions other than commissioning period
described in Section 3.2. The red vertical line indicates the end of

the commissioning period	 3-41

A3.2.16 Plot of self-monitoring data for selenium (jag/L) for Belews
Creek after exclusions other than commissioning period
described in Section 3.2 and superimposed with smooth curves
using LOWESS. The red vertical line indicates the end of the

commissioning period	 3-42

A3.2.17 Plot of all self-monitoring data for selenium (jag/L) for Belews

Creek, bioreactor effluent only. The red vertical line indicates the

end of the commissioning period	 3-43

A3.2.18 Plot of all self-monitoring data for selenium (jag/L) for Belews
Creek superimposed with smooth curves using LOWESS,
bioreactor effluent only. The red vertical line indicates the end of

the commissioning period	 3-44

A3.2.19 Plot of self-monitoring data for selenium (jag/L) for Belews
Creek after exclusions other than commissioning period
described in Section 3.2, bioreactor effluent only. The red vertical

line indicates the end of the commissioning period	 3-45

A3.2.20 Plot of self-monitoring data for selenium (jag/L) for Belews
Creek after exclusions other than commissioning period
described in Section 3.2 and superimposed with smooth curves
using LOWESS, bioreactor effluent only. The red vertical line
indicates the end of the commissioning period	 3-46

3-5

-------
As explained in Section 3.2, EPA excluded some data collected during the initial commissioning
period of the treatment system from the plant self-monitoring data from Allen and Belews Creek.
The initial commissioning period was determined from an engineering perspective based on
extensive experience evaluating treatment systems at power plants and other industry sectors and
knowledge of the physical, chemical, and biological treatment processes employed and by looking at
the longitudinal plots of all the data. In addition to the longitudinal plots of the data, smooth curves
obtained by LOWESS (Cleveland, 1979 and 1981) superimposed on the individual data plots were
also examined to help determine when the treatment system appeared to stabilize. In examining the
longitudinal plots, EPA looked at two types of plots: plots of all the available plant self-monitoring
data and plots of data after the exclusions for plant self-monitoring data due to reasons other than
initial commissioning period described in Section 3.2. The conclusions from both of these
approaches were consistent. The initial six months of operation was determined to represent a
reasonable estimate of the commissioning period for the treatment systems.

This Appendix contains the longitudinal plots for arsenic, mercury, and selenium that engineers used
as aids in determining the commissioning period for Allen and Belews Creek. Four sets of plots are
given for each pollutant and each plant in this Appendix. The first set of plots shows all the available
data points for the plant self-monitoring (before excluding any data for Allen and Belews Creek).
The second set of plots is based on all available data (as in the first plot) with smooth curves
(obtained by LOWESS) superimposed on the individual data. The third set of plots show the self-
monitoring data after all necessary exclusions other than the initial commissioning period for the
wastewater treatment system (as described in Section 3.2). The fourth set of plots show the data
presented in the third set of plots, but with smooth curves (obtained by LOWESS) superimposed on
the individual data points. The red vertical lines indicate the end of the commissioning period, thus
the commissioning period is represented by the data to the left of the red vertical line.

For Allen, the two plots show concentrations from March 3, 2009 to October 22, 2013. For Belews
Creek, the plots show concentrations from February 6, 2008 to November 28, 2013. For each plant,
the upper plot shows the concentrations collected at all three sampling locations, while the lower
plot shows the concentrations only at the bioreactor influent and bioreactor effluent.

3-6

-------
Figure A3.1.1 Plot of all self-monitoring data for arsenic (jJg/L) for Alien. The red vertical line
indicates the end of the commissioning period

Allen

Mar 2009

Apr 2010

Jun 2011

Aug 2012

Oct 2013

200 -r

900 -

600 -

300 -

150 -

100 -

Mar 2009	Apr 2010	Jun 2011	Aug 2012	Oct 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

3-7

-------
Figure A3.1.2 Plot of all self-monitoring data for arsenic (jJg/L) for Alien superimposed with
smooth curves using LOWESS. The red vertical line indicates the end of the
commissioning period

Allen

900 -

600 -

300 -

? 0 -

O)

~	Mar 2009	Apr 2010	Jun2011	Aug 2012	Oct 2013

c



< 200 -

150 -

100 -

50 -

0 -

Mar 2009	Apr 2010	Jun2011	Aug 2012	Oct 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

3-8

-------
Figure A3.1.3 Plot of self-monitoring data for arsenic (pg/L) for Allen after exclusions other than
commissioning period described in Section 3.2. The red vertical line indicates the
end of the commissioning period

Allen

900 -

600 -

300 -

150 -

100 -

Mar 2009

Apr 2010

Jun 2011

Aug 2012

Oct 2013

200 -f

Mar 2009

Apr 2010

Jun 2011

Aug 2012

Oct 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

3-9

-------
Figure A3.1.4 Plot of self-monitoring data for arsenic (pg/L) for Allen after exclusions other than
commissioning period described in Section 3.2 and superimposed with smooth
curves using LOWESS. The red vertical line indicates the end of the commissioning
period

Allen

C

<1)
c/)

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

3-10

-------
Figure A3.1.5 Plot of all self-monitoring data for mercury (ng/L) for Alien. The red vertical line
indicates the end of the commissioning period

Allen

Mar 2009	Apr 2010	Jun2011	Aug 2012	Oct 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected	_ .

Date

3-11

-------
Figure A3.1.6 Plot of all self-monitoring data for mercury (ng/L) for Alien superimposed with
smooth curves using LOWESS. The red vertical line indicates the end of the
commissioning period

Allen

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected	_ .

Date

3-12

-------
Figure A3.1.7 Plot of self-monitoring data for mercury (ng/L) for Allen after exclusions other than
commissioning period described in Section 3.2. The red vertical line indicates the
end of the commissioning period

Allen

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected	_ .

Date

3-13

-------
Figure A3.1.8 Plot of self-monitoring data for mercury (ng/L) for Allen after exclusions other than
commissioning period described in Section 3.2 and superimposed with smooth
curves using LOWESS. The red vertical line indicates the end of the commissioning
period

Allen

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected	_ .

Date

3-14

-------
Figure A3.1.9 Plot of all self-monitoring data for Nitrate-Nitrite as N (mg/L) for Allen

Allen

130

65 -

Sep 2011

Apr 2012

Oct 2012

Apr 2013

Oct 2013

100 -

50 -

Sep 2011

~ FGD Purge
O Bioreactor Inf.
A Bioreactor Eff.
* Not Detected

Apr 2012

Oct 2012

Date

Apr 2013

Oct 2013

3-15

-------
Figure A3.1.10 Plot of all self-monitoring data for Nitrate-Nitrite as N (mg/L) for Allen superimposed
with smooth curves using LOWESS

Allen

130

65 -

0 ^

Sep 2011

Apr 2012

Oct 2012

Apr 2013

Oct 2013

100

50

0 ^

Sep 2011

Apr 2012

~ FGD Purge
O Bioreactor Inf.
£ Bioreactor Eff.
* Not Detected

Oct 2012

Date

Apr 2013

Oct 2013

3-16

-------
Figure A3.1.11 Plot of self-monitoring data for Nitrate-Nitrite as N (mg/L) for Allen after exclusions
other than commissioning period described in Section 3.2

Allen

130 H

65 -

o H

Sep 2011

Apr 2012

Oct 2012

Apr 2013

Oct 2013

100 -

50 -

o H

Sep 2011

Apr 2012

~ FGD Purge
O Bioreactor Inf.

Bioreactor Eff.
* Not Detected

Oct 2012

Date

Apr 2013

Oct 2013

3-17

-------
Figure A3.1.12 Plot of self-monitoring data for Nitrate-Nitrite as N (mg/L) for Allen after exclusions
other than commissioning period described in Section 3.2 and superimposed with
smooth curves using LOWESS

Allen

Sep 2011

Apr 2012

Oct 2012

Apr 2013

Oct 2013

100 -

50 -

Sep 2011

~	FGD Purge

O	Bioreactor Inf.
£ Bioreactor Eff.

*	Not Detected

Apr 2012

Oct 2012

Date

Apr 2013

Oct 2013

3-18

-------
Figure A3.1.13 Plot of all self-monitoring data for selenium (ng/L) for Allen. The red vertical line
indicates the end of the commissioning period

Allen

10000 -







7500 -













5000 -



1

s







2500 -





1

1 l\

J



o -

pi in i mi rr «« - - - - -^c, .... ...



I W V -

^	Mar 2009	Apr 2010	Jun2011	Aug 2012	Oct 2013

D
"c

CD

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

3-19

-------
Figure A3.1.14 Plot of all self-monitoring data for selenium (ng/L) for Allen superimposed with
smooth curves using LOWESS. The red vertical line indicates the end of the
commissioning period

Allen

10000 -
7500 -
5000 -
2500 -

d 0 -

O)

^	Mar 2009	Apr 2010	Jun2011	Aug 2012	Oct 2013

D
*C
Q)

$ 4000 -
3000 -

2000 -

1000 -

0 -

Mar 2009	Apr 2010	Jun2011	Aug 2012	Oct 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

3-20

-------
Figure A3.1.15 Plot of self-monitoring data for selenium (pg/L) for Ailen after exclusions other than
commissioning period described in Section 3.2. The red vertical line indicates the
end of the commissioning period

Allen

Mar 2009	Apr 2010	Jun2011	Aug 2012	Oct 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

3-21

-------
Figure A3.1.16 Plot of self-monitoring data for selenium (pg/L) for Ailen after exclusions other than
commissioning period described in Section 3.2 and superimposed with smooth
curves using LOWESS. The red vertical line indicates the end of the commissioning
period

Allen

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

3-22

-------
Figure A3.1.17 Plot of all self-monitoring data for selenium (ng/L) for Allen, bioreactor effluent
only. The red vertical line indicates the end of the commissioning period

Allen

120

90

60 -

30 -

JL

I—
Mar 2009

T a A.

	1	

Apr 2010

	1	

Jun 2011

	1	

Aug 2012

	r

Oct 2013

^ Bioreactor Effluent
* Not Detected

Date

3-23

-------
Figure A3.1.18 Plot of all self-monitoring data for selenium (ng/L) for Allen superimposed with
smooth curves using LOWESS, bioreactor effluent only. The red vertical line
indicates the end of the commissioning period

Allen

120

90 -

60 -

30

Mar 2009

Apr 2010

Jun 2011

Aug 2012

Oct 2013

A Bioreactor Effluent
* Not Detected

Date

3-24

-------
Figure A3.1.19 Plot of self-monitoring data for selenium (pg/L) for Ailen after exclusions other than
commissioning period described in Section 3.2, bioreactor effluent only. The red
vertical line indicates the end of the commissioning period

Allen

120 -

90

60

30 -

0 -



Mar 2009

Apr 2010

Jun 2011

Aug 2012

Oct 2013

A Bioreactor Effluent
» Not Detected

Date

3-25

-------
Figure A3.1.20 Plot of self-monitoring data for selenium (pg/L) for Ailen after exclusions other than
commissioning period described in Section 3.2 and superimposed with smooth
curves using LOWESS,, bioreactor effluent only. The red vertical line indicates the
end of the commissioning period

Allen

120 -

90 -

60

30 -

0 -

Mar 2009

Apr 2010

Jun 2011

Aug 2012

Oct 2013

A Bioreactor Effluent
* Not Detected

Date

3-26

-------
Figure A3.2.1 Plot of all self-monitoring data for arsenic (jJg/L) for Belews Creek. The red vertical
line indicates the end of the commissioning period

Belews Creek

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

~	FGD Purge

O	Bioreactor Inf.

£	Bioreactor Eff.

*	Not Detected

Date

3-27

-------
Figure A3.2.2 Plot of all self-monitoring data for arsenic (jJg/L) for Belews Creek superimposed
with smooth curves using LOWESS. The red vertical line indicates the end of the
commissioning period

Belews Creek

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

~ FGD Purge
O Bioreactor Inf.
A Bioreactor Eff.
* Not Detected

Date

3-28

-------
Figure A3.2.3 Plot of self-monitoring data for arsenic (pg/L) for Belews Creek after exclusions
other than commissioning period described in Section 3.2. The red vertical line
indicates the end of the commissioning period

Belews Creek

2500 -

1875 -

1250 -

625 -

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

~	FGD Purge

O	Bioreactor Inf.

£	Bioreactor Eff.

*	Not Detected

Feb 2008	Jul 2009	Jan 2011	Jun 2012	Nov 2013

3-29

-------
Figure A3.2.4 Plot of self-monitoring data for arsenic (pg/L) for Belews Creek after exclusions

other than commissioning period described in Section 3.2 and superimposed with
smooth curves using LOWESS. The red vertical line indicates the end of the
commissioning period

o
'c
a)
<0

2500 -

Belews Creek

Feb 2008

Jul 2009

Feb 2008

Jul 2009

Jan 2011

Jan 2011

Jun 2012

Nov 2013

Jun 2012

Nov 2013

~	FGD Purge

O	Bioreactor Inf.

£	Bioreactor Eff.

*	Not Detected

Date

3-30

-------
Figure A3.2.5 Plot of all self-monitoring data for mercury (ng/L) for Belews Creek. The red vertical
line indicates the end of the commissioning period

Belews Creek

600000
400000
200000

25000 -

Feb 2008	Jul 2009	Jan 2011	Jun2012	Nov 2013

~ FGD Purge
O Bioreactor Inf.

Bioreactor Eff.

* Not Detected	Date

Feb 2008	Jul 2009	Jan 2011	Jun2012	Nov 2013

50000 -

3-31

-------
Figure A3.2.6 Plot of all self-monitoring data for mercury (ng/L) for Belews Creek superimposed
with smooth curves using LOWESS. The red vertical line indicates the end of the
commissioning period

Belews Creek

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

50000 -

600000

200000 -

25000 -

400000 -

Feb 2008	Jul 2009	Jan 2011	Jun 2012	Nov 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected	Date

3-32

-------
Figure A3.2.7 Plot of self-monitoring data for mercury (ng/L) for Belews Creek after exclusions
other than commissioning period described in Section 3.2. The red vertical line
indicates the end of the commissioning period

Belews Creek

400000 -

200000 -

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

25000 -

Feb 2008	Jul 2009	Jan 2011	Jun 2012	Nov 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected	Date

50000 -

3-33

-------
Figure A3.2.8 Plot of self-monitoring data for mercury (ng/L) for Belews Creek after exclusions
other than commissioning period described in Section 3.2 and superimposed with
smooth curves using LOWESS. The red vertical line indicates the end of the
commissioning period

Belews Creek

~ FGD Purge
O Bioreactor Inf.

Bioreactor Eff.

* Not Detected	Date

3-34

-------
Figure A3.2.9 Plot of all self-monitoring data for Nitrate-Nitrite as N (mg/L) for Beiews Creek

Belews Creek

May 2013	Jul 2013	Aug 2013	Oct 2013	Nov 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

May 2013	Jul 2013	Aug 2013	Oct 2013	Nov 2013

3-35

-------
Figure A3.2.10 Plot of all self-monitoring data for Nitrate-Nitrite as N (mg/L) for Beiews Creek
superimposed with smooth curves using LOWESS

Beiews Creek

May 2013	Jul 2013	Aug 2013	Oct 2013	Nov 2013

May 2013	Jul 2013	Aug 2013	Oct 2013	Nov 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

3-36

-------
Figure A3.2.11 Plot of self-monitoring data for Nitrate-Nitrite as N (mg/L) for Belews Creek after
exclusions other than commissioning period described in Section 3.2

Belews Creek

May 2013

Jul 2013

Aug 2013

Oct 2013

Nov 2013

May 2013	Jul 2013	Aug 2013	Oct 2013	Nov 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

3-37

-------
Figure A3.2.12 Plot of self-monitoring data for Nitrate-Nitrite as N (mg/L) for Belews Creek after
exclusions other than commissioning period described in Section 3.2 and
superimposed with smooth curves using LOWESS

Belews Creek

May 2013	Jul 2013	Aug 2013	Oct 2013	Nov 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

May 2013	Jul 2013	Aug 2013	Oct 2013	Nov 2013

3-38

-------
Figure A3.2.13 Plot of all self-monitoring data for selenium (ng/L) for Belews Creek. The red
vertical line indicates the end of the commissioning period

Belews Creek

25000 -
20000 -
15000 -
10000 -
5000 -

5" 0 -
a

^	Feb 2008

g
'c
0)

0		

CO

Feb 2008	Jul 2009	Jan 2011	Jun2012	Nov 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected	Date

Jul 2009	Jan 2011	Jun2012	Nov 2013

2500 -
2000 -
1500 -
1000 -
500 -
0 -

3-39

-------
Figure A3.2.14 Plot of all self-monitoring data for selenium (ng/L) for Belews Creek superimposed
with smooth curves using LOWESS. The red vertical line indicates the end of the
commissioning period.

Belews Creek

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

~ FGD Purge
O Bioreactor Inf.
A Bioreactor Eff.
* Not Detected

Date

3-40

-------
Figure A3.2.15 Plot of self-monitoring data for selenium (pg/L) for Belews Creek after exclusions
other than commissioning period described in Section 3.2. The red vertical line
indicates the end of the commissioning period

Belews Creek

25000 -
20000 -
15000
10000 -
5000 -
0



1

]















-re-/Vr Ortfrp- oE



Feb 2008

"T

Jul 2009

Jan 2011

r

Jun 2012

T

Nov 2013

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

~ FGD Purge
O Bioreactor Inf.
A Bioreactor Eff.
* Not Detected

Date

3-41

-------
Figure A3.2.16 Plot of self-monitoring data for selenium (jjg/L) for Beiews Creek after exclusions
other than commissioning period described in Section 3.2 and superimposed with
smooth curves using LOWESS. The red vertical line indicates the end of the
commissioning period

E
g
'c
m

-------
Figure A3.2.17 Plot of all self-monitoring data for selenium (ng/L) for Belews Creek, bioreactor
effluent only. The red vertical line indicates the end of the commissioning period

Belews Creek

A Bioreactor Effluent
* Not Detected

Date

3-43

-------
Figure A3.2.18 Plot of all self-monitoring data for selenium (ng/L) for Belews Creek superimposed
with smooth curves using LOWESS, bioreactor effluent only. The red vertical line
indicates the end of the commissioning period

Belews Creek

300 -

240

180 -

120

60 -

A A

	1	

Jul 2009

	r

Nov 2013

Feb 2008

Jan 2011

Jun 2012

A Bioreactor Effluent
* Not Detected

Date

3-44

-------
Figure A3.2.19 Plot of self-monitoring data for selenium (pg/L) for Belews Creek after exclusions
other than commissioning period described in Section 3.2, bioreactor effluent only.
The red vertical line indicates the end of the commissioning period

Belews Creek

30 -

24

18 -

12 -

0 -

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

A Bioreactor Effluent
* Not Detected

Date

3-45

-------
Figure A3.2.20 Plot of self-monitoring data for selenium (pg/L) for Beiews Creek after exclusions
other than commissioning period described in Section 3.2 and superimposed with
smooth curves using LOWESS, bioreactor effluent only. The red vertical line
indicates the end of the commissioning period

Beiews Creek

A Bioreactor Effluent
* Not Detected

Date

3-46

-------
References:

Cleveland, W. S. (1979) Robust locally weighted regression and smoothing scatterplots./. Amer.
Statist. Assoc. 74, 829—836.

Cleveland, W. S. (1981) LOWESS: A program for smoothing scatterplots by robust locally weighted
regression. The American Statistician, 35, 54.

3-47

-------
Appendix 4. Plots of Available Data for Each Plant

for Calculating Limits Prior to Any
Exclusions

4-1

-------
Table of Contents

Figures	Page

A4.1.1 FGD Purge and Chemical precipitation effluent data of arsenic
for Hatfield's Ferry, prior to excluding any data for reasons
described in Section 3.2. Pollutant concentrations are shown on a
linear scale in the upper plot, log scale in the lower plot. The
purple vertical line indicates the point when the plant changed
analytical laboratories, and the red vertical line indicates the start

of decommissioning period	 4-8

A4.1.2 Chemical precipitation effluent data of arsenic for Hatfield's

Ferry, prior to excluding any data for reasons described in Section
3.2. Pollutant concentrations are shown on a linear scale in the
upper plot, log scale in the lower plot. The purple vertical line
indicates the point when the plant changed analytical laboratories,
and the red vertical line indicates the start of decommissioning

period	 4-9

A4.1.3 FGD Purge and Chemical precipitation effluent data of mercury
for Hatfield's Ferry, prior to excluding any data for reasons
described in Section 3.2. Pollutant concentrations are shown on a
linear scale in the upper plot, log scale in the lower plot. The
purple vertical line indicates the point when the plant changed
analytical laboratories, and the red vertical line indicates the start

of decommissioning period	 4-10

A4.1.4 Chemical precipitation effluent data of mercury for Hatfield's

Ferry, prior to excluding any data for reasons described in Section

3.2. Pollutant concentrations are shown on a linear scale in the

upper plot, log scale in the lower plot. The purple vertical line

indicates the point when the plant changed analytical laboratories,

and the red vertical line indicates the start of decommissioning

period	 4-11

A4.2.1 FGD Purge and Chemical precipitation effluent data of arsenic

for Keystone, prior to excluding any data for reasons described in
Section 3.2. Pollutant concentrations are shown on a linear scale

in the upper plot, log scale in the lower plot	 4-12

A4.2.2 Chemical precipitation effluent data of arsenic for Keystone,

prior to excluding any data for reasons described in Section 3.2.

Pollutant concentrations are shown on a linear scale in the upper

plot, log scale in the lower plot	 4-13

A4.2.3 FGD Purge and Chemical precipitation effluent data of mercury
for Keystone, prior to excluding any data for reasons described in
Section 3.2. Pollutant concentrations are shown on a linear scale
in the upper plot, log scale in the lower plot	 4-14

4-2

-------
Table of Contents
(continued)

Figures	Page

A4.2.4 Chemical precipitation effluent data of mercury for Keystone,
prior to excluding any data for reasons described in Section 3.2.

Pollutant concentrations are shown on a linear scale in the upper
plot, log scale in the lower plot	 4-15

A4.3.1 FGD Purge and Chemical precipitation effluent data of arsenic
for Miami Fort, prior to excluding any data for reasons described
in Section 3.2. Pollutant concentrations are shown on a linear

scale in the upper plot, log scale in the lower plot	 4-16

A4.3.2 Chemical precipitation effluent data of arsenic for Miami Fort,
prior to excluding any data for reasons described in Section 3.2.

Pollutant concentrations are shown on a linear scale in the upper

plot, log scale in the lower plot	 4-17

A4.3.3 FGD Purge and Chemical precipitation effluent data of mercury
for Miami Fort, prior to excluding any data for reasons described
in Section 3.2. Pollutant concentrations are shown on a linear

scale in the upper plot, log scale in the lower plot	 4-18

A4.3.4 Chemical precipitation effluent data of mercury for Miami Fort,
prior to excluding any data for reasons described in Section 3.2.

Pollutant concentrations are shown on a linear scale in the upper

plot, log scale in the lower plot	 4-19

A4.4.1 FGD Purge and Chemical precipitation effluent data of arsenic
for Pleasant Prairie, prior to excluding any data for reasons
described in Section 3.2. Pollutant concentrations are shown on a

linear scale in the upper plot, log scale in the lower plot	 4-20

A4.4.2 Chemical precipitation effluent data of arsenic for Pleasant
Prairie, prior to excluding any data for reasons described in
Section 3.2. Pollutant concentrations are shown on a linear scale

in the upper plot, log scale in the lower plot	 4-21

A4.4.3 FGD Purge and Chemical precipitation effluent data of mercury
for Pleasant Prairie, prior to excluding any data for reasons
described in Section 3.2. Pollutant concentrations are shown on a

linear scale in the upper plot, log scale in the lower plot	 4-22

A4.4.4 Chemical precipitation effluent data of mercury for Pleasant
Prairie, prior to excluding any data for reasons described in
Section 3.2. Pollutant concentrations are shown on a linear scale
in the upper plot, log scale in the lower plot	 4-23

4-3

-------
Table of Contents
(continued)

Figures	Page

A4.5.1 FGD Purge, Bioreactor influent, and Bioreactor effluent data of
arsenic for Allen, prior to excluding any data for reasons
described in Section 3.2. Pollutant concentrations are shown on a
linear scale in the upper plot, log scale in the lower plot. The red

vertical line indicates the end of the commissioning period	 4-24

A4.5.2 Bioreactor influent and Bioreactor effluent data of arsenic for

Allen, prior to excluding any data for reasons described in Section
3.2. Pollutant concentrations are shown on a linear scale in the
upper plot, log scale in the lower plot. The red vertical line

indicates the end of the commissioning period	 4-25

A4.5.3 Bioreactor effluent data of arsenic for Allen, prior to excluding
any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log
scale in the lower plot. The red vertical line indicates the end of

the commissioning period	 4-26

A4.5.4 FGD Purge, Bioreactor influent, and Bioreactor effluent data of
mercury for Allen, prior to excluding any data for reasons
described in Section 3.2. Pollutant concentrations are shown on a
linear scale in the upper plot; log scale in the lower plot. The red

vertical line indicates the end of the commissioning period	 4-27

A4.5.5 Bioreactor influent and Bioreactor effluent data of mercury for

Allen, prior to excluding any data for reasons described in Section
3.2. Pollutant concentrations are shown on a linear scale in the
upper plot, log scale in the lower plot. The red vertical line

indicates the end of the commissioning period	 4-28

A4.5.6 Bioreactor effluent data of mercury for Allen, prior to excluding
any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log
scale in the lower plot. The red vertical line indicates the end of

the commissioning period	 4-29

A4.5.7 FGD Purge, Bioreactor influent, and Bioreactor effluent data of
Nitrate-Nitrite as N for Allen, prior to excluding any data for
reasons described in Section 3.2. Pollutant concentrations are
shown on a linear scale in the upper plot, log scale in the lower

plot	 4-30

A4.5.8 Bioreactor influent and Bioreactor effluent data of Nitrate-Nitrite
as N for Allen, prior to excluding any data for reasons described
in Section 3.2. Pollutant concentrations are shown on a linear
scale in the upper plot, log scale in the lower plot	 4-31

4-4

-------
Table of Contents
(continued)

Figures	Page

A4.5.9 Bioreactor effluent data of Nitrate-Nitrite as N for Allen, prior to
excluding any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log

scale in the lower plot	 4-32

A4.5.10 FGD Purge, Bioreactor influent, and Bioreactor effluent data of
selenium for Allen, prior to excluding any data for reasons
described in Section 3.2. Pollutant concentrations are shown on a
linear scale in the upper plot, log scale in the lower plot. The red

vertical line indicates the end of the commissioning period	 4-33

A4.5.11 Bioreactor influent and Bioreactor effluent data of selenium for

Allen, prior to excluding any data for reasons described in Section
3.2. Pollutant concentrations are shown on a linear scale in the
upper plot, log scale in the lower plot. The red vertical line

indicates the end of the commissioning period	 4-34

A4.5.12 Bioreactor effluent data of selenium for Allen, prior to excluding
any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log
scale in the lower plot. The red vertical line indicates the end of
the commissioning period	 4-35

A4.6.1 FGD Purge, Bioreactor influent, and Bioreactor effluent data of
arsenic for Belews Creek, prior to excluding any data for reasons
described in Section 3.2. Pollutant concentrations are shown on a
linear scale in the upper plot, log scale in the lower plot. The red

vertical line indicates the end of the commissioning period	 4-36

A4.6.2 Bioreactor influent and Bioreactor effluent data of arsenic for
Belews Creek, prior to excluding any data for reasons described
in Section 3.2. Pollutant concentrations are shown on a linear
scale in the upper plot, log scale in the lower plot. The red

vertical line indicates the end of the commissioning period	 4-37

A4.6.3 Bioreactor effluent data of arsenic for Belews Creek, prior to

excluding any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log
scale in the lower plot. The red vertical line indicates the end of

the commissioning period	 4-38

A4.6.4 FGD Purge, Bioreactor influent, and Bioreactor effluent data of
mercury for Belews Creek, prior to excluding any data for reasons
described in Section 3.2. Pollutant concentrations are shown on a
linear scale in the upper plot, log scale in the lower plot. The red
vertical line indicates the end of the commissioning period	 4-39

4-5

-------
Table of Contents
(continued)

Figures	Page

A4.6.5 Bioreactor influent and Bioreactor effluent data of mercury for
Belews Creek, prior to excluding any data for reasons described
in Section 3.2. Pollutant concentrations are shown on a linear
scale in the upper plot, log scale in the lower plot. The red

vertical line indicates the end of the commissioning period	 4-40

A4.6.6 Bioreactor effluent data of mercury for Belews Creek, prior to

excluding any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot log
scale in the lower plot. The red vertical line indicates the end of

the commissioning period	 4-41

A4.6.7 FGD Purge, Bioreactor influent, and Bioreactor effluent data of
Nitrate-Nitrite as N for Belews Creek, prior to excluding any data
for reasons described in Section 3.2. Pollutant concentrations are
shown on a linear scale in the upper plot, log scale in the lower

plot	 4-42

A4.6.8 Bioreactor influent and Bioreactor effluent data of Nitrate-Nitrite
as N for Belews Creek, prior to excluding any data for reasons
described in Section 3.2. Pollutant concentrations are shown on a

linear scale in the upper plot, log scale in the lower plot	 4-43

A4.6.9 Bioreactor effluent data of Nitrate-Nitrite as N for Belews Creek,
prior to excluding any data for reasons described in Section 3.2.

Pollutant concentrations are shown on a linear scale in the upper

plot, log scale in the lower plot	 4-44

A4.6.10 FGD Purge, Bioreactor influent, and Bioreactor effluent data of
selenium for Belews Creek, prior to excluding any data for
reasons described in Section 3.2. Pollutant concentrations are
shown on a linear scale in the upper plot, log scale in the lower
plot. The red vertical line indicates the end of the commissioning

period	 4-45

A4.6.11 Bioreactor influent and Bioreactor effluent data of selenium for
Belews Creek, prior to excluding any data for reasons described
in Section 3.2. Pollutant concentrations are shown on a linear
scale in the upper plot, log scale in the lower plot. The red

vertical line indicates the end of the commissioning period	 4-46

A4.6.12 Bioreactor effluent data of selenium for Belews Creek, prior to

excluding any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log
scale in the lower plot. The red vertical line indicates the end of
the commissioning period	 4-47

4-6

-------
This Appendix contains plots of all available data for each plant used in calculating the limits for
FGD wastewater based on the chemical precipitation or biological treatment technology options,
prior to any exclusion (described in Section 3.2). Specifically, plots of the available data are shown
for the following plants: Allen, Belews Creek, Hatfield's Ferry, Keystone, Miami Fort, and Pleasant
Prairie. Plots of all available data for Brindisi, Polk, and Wabash River are presented in Sections 8
and 9 of this document. For each plant, the data are plotted on two different scales, linear and
logarithmic. The purpose of the logarithmic is to stretch the scale so that the data are easier to
visualize.

¦	For Hatfield's Ferry, the purple vertical line indicates the point when the plant changed
analytical laboratories, and the red vertical line indicates the start of the
decommissioning period for the plant.

¦	For Hatfield's Ferry, arsenic or mercury in some chemical precipitation effluent samples
was measured at zero. They are plotted at half the lowest non-zero value in the log scale
plot to illustrate that they are at low level but not to the point that the remainder of the
plot is distorted.

¦	For Pleasant Prairie, the secondary clarifier data are not included in the plot.

4-7

-------
Figure A4.1.1 FGD Purge and Chemical precipitation effluent data of arsenic for Hatfield's Ferry,
prior to excluding any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log scale in the lower
plot. The purple vertical line indicates the point when the plant changed analytical
laboratories, and the red vertical line indicates the start of decommissioning period

Hatfield's Ferry

~ FGD Purge

O Chemical Precipitation Effluent
* Not Detected

Date

4-8

-------
Figure A4.1.2 Chemical precipitation effluent data of arsenic for Hatfield's Ferry, prior to excluding
any data for reasons described in Section 3.2. Pollutant concentrations are shown
on a linear scale in the upper plot, log scale in the lower plot. The purple vertical line
indicates the point when the plant changed analytical laboratories, and the red
vertical line indicates the start of decommissioning period

Hatfield's Ferry

a>

cn

O Chemical Precipitation Effluent

* Not Detected	Date

4-9

-------
Figure A4.1.3 FGD Purge and Chemical precipitation effluent data of mercury for Hatfield's Ferry,
prior to excluding any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log scale in the lower
plot. The purple vertical line indicates the point when the plant changed analytical
laboratories, and the red vertical line indicates the start of decommissioning period

Hatfield's Ferry

800000 H

600000

400000 -

200000

a>
c

Jun 2009

Aug 2010

Sep 2011

Jun 2009
~ FGD Purge

O Chemical Precipitation Effluent
* Not Detected

Aug 2010

Sep 2011

l

Oct 2012

~T
Dec 2013

Oct 2012

Dec 2013

Date

4-10

-------
Figure A4.1.4 Chemical precipitation effluent data of mercury for Hatfield's Ferry, prior to

excluding any data for reasons described in Section 3.2. Pollutant concentrations
are shown on a linear scale in the upper plot, log scale in the lower plot. The purple
vertical line indicates the point when the plant changed analytical laboratories, and
the red vertical line indicates the start of decommissioning period

Hatfield's Ferry

O Chemical Precipitation Effluent
* Not Detected

Date

4-11

-------
Figure A4.2.1 FGD Purge and Chemical precipitation effluent data of arsenic for Keystone, prior to
excluding any data for reasons described in Section 3.2. Pollutant concentrations
are shown on a linear scale in the upper plot, log scale in the lower plot

5000

2500 -

Sep 2010

Keystone

Aug 2011

Jul 2012

Jun 2013

Apr 2014

1000

100 -

10

1 H

Sep 2010

~ FGD Purge

O Chemical Precipitation Effluent
* Not Detected

Aug 2011

Jul 2012

Date

Jun 2013

Apr 2014

4-12

-------
Figure A4.2.2 Chemical precipitation effluent data of arsenic for Keystone, prior to excluding any
data for reasons described in Section 3.2. Pollutant concentrations are shown on a
linear scale in the upper plot, log scale in the lower plot

Keystone

Sep 2010

Aug 2011

Jul 2012

Jun 2013

Apr 2014

1 H

Sep 2010

Aug 2011

Jul 2012

Jun 2013

Apr 2014

O Chemical Precipitation Effluent
* Not Detected

Date

4-13

-------
Figure A4.2.3 FGD Purge and Chemical precipitation effluent data of mercury for Keystone, prior
to excluding any data for reasons described in Section 3.2. Pollutant concentrations
are shown on a linear scale in the upper plot, log scale in the lower plot

Sep 2010

=3
O

i—

21000000

100000 -

Sep 2010

Keystone

Aug 2011

Aug 2011

Jul 2012

Jul 2012

Jun 2013

Jun 2013

Apr 2014

Apr 2014

~ FGD Purge

O Chemical Precipitation Effluent
* Not Detected

Date

4-14

-------
Figure A4.2.4 Chemical precipitation effluent data of mercury for Keystone, prior to excluding any
data for reasons described in Section 3.2. Pollutant concentrations are shown on a
linear scale in the upper plot, log scale in the lower plot

5000 -

2500 -

5" 0 -

O)

Sep 2010	Aug 2011	Jul2012	Jun2013	Apr2014

CD

1000 -

100 -

Sep 2010	Aug 2011	Jul 2012	Jun2013	Apr 2014

O Chemical Precipitation Effluent
* Not Detected

Date

Keystone

IU







4-15

-------
Figure A4.3.1 FGD Purge and Chemical precipitation effluent data of arsenic for Miami Fort, prior
to excluding any data for reasons described in Section 3.2. Pollutant concentrations
are shown on a linear scale in the upper plot, log scale in the lower plot

Miami Fort

Jul 2009

Aug 2010

Sep 2011

Oct 2012

Dec 2013

1000 -

100 -

10

i

\	A,,,





1

rnt i * I



Ju! 2009

~ FGD Purge

O Chemical Precipitation Effluent
* Not Detected

Aug 2010

Sep 2011

Date

Oct 2012

Dec 2013

4-16

-------
Figure A4.3.2 Chemical precipitation effluent data of arsenic for Miami Fort, prior to excluding any
data for reasons described in Section 3.2. Pollutant concentrations are shown on a
linear scale in the upper plot, log scale in the lower plot

Miami Fort

500 -





400 -





300 -





200 -



|

100 -

\ ^ 1

I

T 0 -

*<««»»»- """ ««•««•»



•~=	Jul 2009	Aug 2010	Sep 2011	Oct 2012	Dec 2013

C
0)

CO

Jul 2009	Aug 2010	Sep 2011	Oct 2012	Dec 2013

O Chemical Precipitation Effluent
* Not Detected

Date

4-17

-------
Figure A4.3.3 FGD Purge and Chemical precipitation effluent data of mercury for Miami Fort, prior
to excluding any data for reasons described in Section 3.2. Pollutant concentrations
are shown on a linear scale in the upper plot, log scale in the lower plot

Miami Fort

Jul 2009	Aug 2010	Sep 2011	Oct 2012	Dec 2013

Jul 2009	Aug 2010	Sep 2011	Oct 2012	Dec 2013

~ FGD Purge

O Chemical Precipitation Effluent
* Not Detected

Date

4-18

-------
Figure A4.3.4 Chemical precipitation effluent data of mercury for Miami Fort, prior to excluding
any data for reasons described in Section 3.2. Pollutant concentrations are shown
on a linear scale in the upper plot, log scale in the lower plot

Miami Fort

800

600 -

Jul 2009

Aug 2010

Sep 2011

Oct 2012

Dec 2013

100 -

10 -

Jul 2009

Aug 2010

Sep 2011

Oct 2012

Dec 2013

O Chemical Precipitation Effluent
* Not Detected

Date

4-19

-------
Figure A4.4.1 FGD Purge and Chemical precipitation effluent data of arsenic for Pleasant Prairie,
prior to excluding any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log scale in the lower
plot

Pleasant Prairie

	T	

Apr 2013

n	

Nov 2007

Mar 2009

Jul 2010

	1	

Dec 2011

	r

Apr 2013

~ FGD Purge

O Chemical Precipitation Effluent	_

* Not Detected	Date

150 -

100 -

100 -

Nov 2007

	1	

Mar 2009

	1	

Jul 2010

	1	

Dec 2011

4-20

-------
Figure A4.4.2 Chemical precipitation effluent data of arsenic for Pleasant Prairie, prior to

excluding any data for reasons described in Section 3.2. Pollutant concentrations
are shown on a linear scale in the upper plot, log scale in the lower plot

Pleasant Prairie

Nov 2007	Mar 2009	Jul 2010	Dec 2011	Apr 2013

C

CD
CO

Nov 2007	Mar 2009	Jul 2010	Dec 2011	Apr 2013

O Chemical Precipitation Effluent
* Not Detected

Date

4-21

-------
Figure A4.4.3 FGD Purge and Chemical precipitation effluent data of mercury for Pleasant Prairie,
prior to excluding any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log scale in the lower
plot

Pleasant Prairie

4000000 -

3000000 -

2000000 -

1000000 -

100000 -

10000 -

1000 -

May 2009

Nov 2010

	1	

Jun 2012

	1	

Dec 2013

"I	

Nov 2007

1000000

	1	

May 2009

I

Nov 2010

I

Jun 2012

	1	

Dec 2013

100 -

Nov 2007

~ FGD Purge

O Chemical Precipitation Effluent
* Not Detected

4-22

-------
Figure A4.4.4 Chemical precipitation effluent data of mercury for Pleasant Prairie, prior to

excluding any data for reasons described in Section 3.2. Pollutant concentrations
are shown on a linear scale in the upper plot, log scale in the lower plot

Pleasant Prairie

3000

2000

1000 -

O)

c
=3

Nov 2007

May 2009

Nov 2010

Jun 2012

Dec 2013

1000

100 -

Nov 2007

O Chemical Precipitation Effluent
* Not Detected

May 2009

Nov 2010

Jun 2012

Dec 2013

Date

4-23

-------
Figure A4.5.1 FGD Purge, Bioreactor influent, and Bioreactor effluent data of arsenic for Allen,
prior to excluding any data for reasons described in Section 3.2, Pollutant
concentrations are shown on a linear scale in the upper plot, log scale in the lower
plot. The red vertical line indicates the end of the commissioning period

CT>
3.

O
'c
CD

Allen

Mar 2009

Apr 2010

Jun 2011

Aug 2012

Oct 2013

500 -

50 -

5 -

Mar 2009

Apr 2010

~	FGD Purge

O	Bioreactor Inf.
A Bioreactor Eff.

*	Not Detected

Jun 2011

Date

Aug 2012

Oct 2013

4-24

-------
Figure A4.5.2 Bioreactor influent and Bioreactor effluent data of arsenic for Allen, prior to

excluding any data for reasons described in Section 3.2. Pollutant concentrations
are shown on a linear scale in the upper plot, log scale in the lower plot. The red
vertical line indicates the end of the commissioning period

Allen

"I	

Mar 2009

Mar 2009

Apr 2010

Juri 2011

Aug 2012

Oct 2013

O Bioreactor Inf.
A Bioreactor Eff.
* Not Detected

150 -

100 -

Apr 2010

	1	

Jun 2011

l

Aug 2012

	1	

Oct 2013

4-25

-------
Figure A4.5.3 Bioreactor effluent data of arsenic for Allen, prior to excluding any data for reasons
described in Section 3.2. Pollutant concentrations are shown on a linear scale in the
upper plot, log scale in the lower plot. The red vertical line indicates the end of the
commissioning period

150 -

100 -

200 H

Allen

Mar 2009

	1	

Apr 2010

Jun 2011

	1	

Aug 2012

	1	

Oct 2013

Mar 2009	Apr 2010	Jun 2011	Aug 2012	Oct 2013

A Bioreactor Eff.

* Not Detected

Date

4-26

-------
Figure A4.5.4 FGD Purge, Bioreactor influent, and Bioreactor effluent data of mercury for Allen,
prior to excluding any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot; log scale in the lower
plot. The red vertical line indicates the end of the commissioning period

Allen

150000 -

50000 -

Mar 2009

Apr 2010

Jun 2011

Aug 2012

Oct 2013

5 100000 -
10000 -
1000 -
100 -
10 -
1 -

Mar 2009	Apr 2010	Jun 2011	Aug 2012	Oct 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

4-27

-------
Figure A4.5.5 Bioreactor influent and Bioreactor effluent data of mercury for Allen, prior to

excluding any data for reasons described in Section 3.2. Pollutant concentrations
are shown on a linear scale in the upper plot, log scale in the lower plot. The red
vertical line indicates the end of the commissioning period

Allen

O Bioreactor Inf.
A Bioreactor Eff.
* Not Detected

Date

4-28

-------
Figure A4.5.6 Bioreactor effluent data of mercury for Allen, prior to excluding any data for reasons
described in Section 3.2. Pollutant concentrations are shown on a linear scale in the
upper plot, log scale in the lower plot. The red vertical line indicates the end of the
commissioning period

Allen

Mar 2009	Apr 2010	Jun2011	Aug 2012	Oct 2013

A Bioreactor Eff.

* Not Detected

Date

4-29

-------
Figure A4.5.7 FGD Purge, Bioreactor influent, and Bioreactor effluent data of Nitrate-Nitrite as N
for Allen, prior to excluding any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log scale in the lower
plot

Allen

125 -
100 -
75 -
50 -
25 -

100
10

1 H
0.1
o.oi H

Aug 2010	May 2011	Mar 2012	Dec 2012	Oct 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

Aug 2010	May 2011	Mar 2012	Dec 2012	Oct 2013

4-30

-------
Figure A4.5.8 Bioreactor influent and Bioreactor effluent data of Nitrate-Nitrite as N for Alien, prior
to excluding any data for reasons described in Section 3.2. Pollutant concentrations
are shown on a linear scale in the upper plot, log scale in the lower plot

Allen

Aug 2010	May 2011	Mar 2012	Jan 2013	Oct 2013

O Bioreactor Inf.

Bioreactor Eff.

* Not Detected

Date

4-31

-------
Figure A4.5.9 Bioreactor effluent data of Nitrate-Nitrite as N for Ailen, prior to excluding any data
for reasons described in Section 3.2. Pollutant concentrations are shown on a linear
scale in the upper plot, log scale in the lower plot

Allen

Aug 2010	May 2011	Mar2012	Jan2013	Oct2013

0.05

Aug 2010	May 2011	Mar 2012	Jan 2013	Oct 2013

A Bioreactor Eff.
* Not Detected

4-32

-------
Figure A4.5.10 FGD Purge, Bioreactor influent, and Bioreactor effluent data of selenium for Allen,
prior to excluding any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log scale in the lower
plot. The red vertical line indicates the end of the commissioning period

Allen

Mar 2009	Apr 2010	Jun2011	Aug 2012	Oct 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

10000 -
7500 -
5000 -
2500 -
0 -

1000 -

100 -

0)
-------
Figure A4.5.11 Bioreactor influent and Bioreactor effluent data of selenium for Alien, prior to

excluding any data for reasons described in Section 3.2. Pollutant concentrations
are shown on a linear scale in the upper plot, log scale in the lower plot. The red
vertical line indicates the end of the commissioning period

Allen

O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

4-34

-------
Figure A4.5.12 Bioreactor effluent data of selenium for Allen, prior to excluding any data for

reasons described in Section 3.2. Pollutant concentrations are shown on a linear
scale in the upper plot, log scale in the lower plot. The red vertical line indicates the
end of the commissioning period

CD

¥

'c

0)
-------
Figure A4.6.1 FGD Purge, Bioreactor influent, and Bioreactor effluent data of arsenic for Belews
Creek, prior to excluding any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log scale in the lower
plot. The red vertical line indicates the end of the commissioning period

Belews Creek

2500 -
2000 -
1500 -
1000 -
500 -

o -

O)

-5	Feb 2008

o
'c
dJ
w

1—

<

500 -

Jul 2009	Jan 2011	Jun2012	Nov 2013

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

~ FGD Purge
O Bioreactor Inf.
A Bioreactor Eff.
* Not Detected

Date

4-36

-------
Figure A4.6.2 Bioreactor influent and Bioreactor effluent data of arsenic for Belews Creek, prior to
excluding any data for reasons described in Section 3.2. Pollutant concentrations
are shown on a linear scale in the upper plot, log scale in the lower plot. The red
vertical line indicates the end of the commissioning period

Belews Creek

CO

o
'c
<1)



*'«¦*	• •Vfi1-'



T

•¦WV.JK Ui1, jmmn,.

Wr\

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

50

5 -

iMmmUmw#' f ~ f 'ft

•iWFWwm-

Feb 2008

Jul 2009

Jan 2011

T

Jun 2012

i

Nov 2013

O Bioreactor Inf.
A Bioreactor Eff.
* Not Detected

Date

4-37

-------
Figure A4.6.3 Bioreactor effluent data of arsenic for Beiews Creek, prior to excluding any data for
reasons described in Section 3.2. Pollutant concentrations are shown on a linear
scale in the upper plot, log scale in the lower plot. The red vertical line indicates the
end of the commissioning period

Beiews Creek

CT>
3.

O
'c
CD

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

50

5

Jul 2009

Jan 2011

I

Jun 2012

I

Nov 2013

Feb 2008

A Bioreactor Eff.

* Not Detected

Date

4-38

-------
Figure A4.6.4 FGD Purge, Bioreactor influent, and Bioreactor effluent data of mercury for Belews
Creek, prior to excluding any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log scale in the lower
plot. The red vertical line indicates the end of the commissioning period

Belews Creek

Feb 2008	Jul 2009	Jan 2011	Jun2012	Nov 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

4-39

-------
Figure A4.6.5 Bioreactor influent and Bioreactor effluent data of mercury for Belews Creek, prior
to excluding any data for reasons described in Section 3.2. Pollutant concentrations
are shown on a linear scale in the upper plot, log scale in the lower plot. The red
vertical line indicates the end of the commissioning period

Belews Creek

60000

40000

20000

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

80000 -|

8000 -

800 -

Feb 2008	Jul 2009	Jan 2011	Jun 2012	Nov 2013

O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

4-40

-------
Figure A4.6.6 Bioreactor effluent data of mercury for Belews Creek, prior to excluding any data for
reasons described in Section 3.2. Pollutant concentrations are shown on a linear
scale in the upper plot log scale in the lower plot. The red vertical line indicates the
end of the commissioning period

Belews Creek

Ol

c

13
O

Feb 2008

Jul 2009

Jan 2011

Jun 2012

r

Nov 2013

800

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

A Bioreactor Eff.
» Not Detected

Date

4-41

-------
Figure A4.6.7 FGD Purge, Bioreactor influent, and Bioreactor effluent data of Nitrate-Nitrite as N
for Belews Creek, prior to excluding any data for reasons described in Section 3.2.
Pollutant concentrations are shown on a linear scale in the upper plot, log scale in
the lower plot

Belews Creek

c»
E

-------
Figure A4.6.8 Bioreactor influent and Bioreactor effluent data of Nitrate-Nitrite as N for Belews
Creek, prior to excluding any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log scale in the lower
plot

Belews Creek

®	Jun 2010	Apr 2011	Mar 2012	Jan 2013	Nov 2013

Jun 2010	Apr 2011	Mar 2012	Jan 2013	Nov 2013

O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

4-43

-------
Figure A4.6.9 Bioreactor effluent data of Nitrate-Nitrite as N for Beiews Creek, prior to excluding
any data for reasons described in Section 3.2. Pollutant concentrations are shown
on a linear scale in the upper plot, log scale in the lower plot

Beiews Creek

0.05 -

Jun 2010	Apr 2011	Mar 2012	Jan 2013	Nov 2013

A Bioreactor Eff.

* Not Detected

0.05 -

0.01 -

I

Jun 2010

	1	

Apr 2011

	1	

Mar 2012

	1	

Jan 2013

	1—

Nov 2013

4-44

-------
Figure A4.6.10 FGD Purge, Bioreactor influent, and Bioreactor effluent data of selenium for Belews
Creek, prior to excluding any data for reasons described in Section 3.2. Pollutant
concentrations are shown on a linear scale in the upper plot, log scale in the lower
plot. The red vertical line indicates the end of the commissioning period

Belews Creek

D
C

CD

Feb 2008	Jul 2009	Jan 2011	Jun2012	Nov 2013

~ FGD Purge
O Bioreactor Inf.

A Bioreactor Eff.

* Not Detected

Date

4-45

-------
Figure A4.6.11 Bioreactor influent and Bioreactor effluent data of selenium for Belews Creek, prior
to excluding any data for reasons described in Section 3.2. Pollutant concentrations
are shown on a linear scale in the upper plot, log scale in the lower plot. The red
vertical line indicates the end of the commissioning period

3000
2500
2000
1500
1000
500
0

Feb 2008

Jul 2009

Jan 2011

Jun 2012

Nov 2013

O Bioreactor Inf.
A Bioreactor Eff.
* Not Detected

Belews Creek

Feb 2008	Jul 2009	Jan 2011	Jun 2012	Nov 2013

400 -
40 -
4 -

4-46

-------
Figure A4.6.12 Bioreactor effluent data of selenium for Belews Creek, prior to excluding any data

for reasons described in Section 3.2. Pollutant concentrations are shown on a linear
scale in the upper plot, log scale in the lower plot. The red vertical line indicates the
end of the commissioning period

Belews Creek

Feb 2008	Jul 2009	Jan 2011	Jun2012	Nov 2013

D
C

A Bioreactor Eff.

* Not Detected

Date

4-47

-------
Appendix 5. Plots to Identify Potential Outlier

Observations Not Consistent with
BAT/NSPS Operation of the Treatment
System

5-1

-------
Table of Contents

Figures	Page

A5.1	Plot of available effluent arsenic and mercury data for Hatfield's

Ferry after excluding all data that warranted exclusion except

outliers	 5-4

A5.2	Plot of available effluent arsenic and mercury data for Keystone

after excluding all data that warranted exclusion except outliers	 5-5

A5.3	Plot of available effluent arsenic and mercury data for Miami Fort

after excluding all data that warranted exclusion except outliers	 5-6

A5.4	Plot of available effluent arsenic and mercury data for Pleasant

Prairie after excluding all data that warranted exclusion except

outliers. The red vertical line indicates the end of TMT 15 use	 5-7

A5.5	Plot of available effluent Nitrate-Nitrite as N and selenium data

for Allen after excluding all data that warranted exclusion except

outliers	 5-8

A5.6 Plot of available effluent Nitrate-Nitrite as N and selenium data
for Belews Creek after excluding all data that warranted exclusion
except outliers	 5-9

5-2

-------
The plots in this appendix are intended as an aid to help identify observations that are outliers to
facilitate evaluating whether the observations are consistent with data for a properly operated
BAT/NSPS model treatment system. The plots presented already exclude data that were already
determined to warrant exclusion, such as those with problem associated with the initial
commissioning period for the treatment system, treatment system upset due to pipe broken, plant
decommissioning period, analytical issues (e.g., analytical interferences, analytical methods not
approved for NPDES purposes, etc.), or other previously identified abnormal operation. These
known data abnormalities were excluded from the plots in this appendix because they distort the
observed data. The resulting plots presented here in Appendix 5 are therefore helpful in identifying
and illustrating the outliers.

¦	Concentrations of arsenic and mercury in the chemical precipitation effluent are plotted
for Hatfield's Ferry, Keystone, Miami Fort, and Pleasant Prairie. Mercury data
associated with the use of TMT-15 in the treatment system at Pleasant Prairie, arsenic
data with problem associated with start up at Pleasant Prairie, and arsenic and mercury
data associated with treatment system update or abnormal operation at Hatfield's Ferry
and Pleasant, are highlighted in pink.

¦	Concentrations of selenium and nitrate-nitrite as N in the bioreactor effluent are plotted
for Allen and Belews Creek.

5-3

-------
Figure A5.1 Plot of available effluent arsenic and mercury data for Hatfield's Ferry after
excluding all data that warranted exclusion except outliers

Hatfield's Ferry

O)

a
'c

0)
V)

30

20

10

0

Jun 2009

Jul 2010

Aug 2011

Sep 2012 Oct 2013

4000

3000

O)

^ 2000
1 1000

0

Jun 2009 Jul 2010 Aug 2011 Sep 2012 Oct 2013

O Chemical Precipitation Effluent	Datp

O Excluded Data	U3le

* Not Detected

5-4

-------
Figure A5.2 Plot of available effluent arsenic and mercury data for Keystone after excluding ali
data that warranted exclusion except outliers

CJ)
Z3

O

'c
CD
to

Keystone

0

Sep 2010 Aug 2011

Jul 2012

Jun 2013 Apr 2014

O)
ZJ

o

I—

CD

Sep 2010

Aug 2011

O Chemical Precipitation Effluent
* Not Detected

Jul 2012
Date

Jun 2013

Apr 2014

5-5

-------
Figure A5.3 Plot of available effluent arsenic and mercury data for Miami Fort after excluding all
data that warranted exclusion except outliers

Miami Fort

O)
23

O
'a
en

CO

Jul 2009

Aug 2010 Sep 2011

Oct 2012

Dec 2013

a>
c

=3

o

1—

d)

800

600

^ 400

200

0

Jul 2009 Aug 2010 Sep 2011

O Chemical Precipitation Effluent
* Not Detected

Oct 2012 Dec 2013

Date

5-6

-------
Figure A5.4 Plot of available effluent arsenic and mercury data for Pleasant Prairie after

excluding all data that warranted exclusion except outliers. The red vertical line
indicates the end of TMT 15 use

Pleasant Prairie

O)
O

'c
CD

C/3

25









20

O







15









10







o

oo

5



o

° o

2 00
8

8oo
o

0









Nov 2007

May 2009

Nov 2010

Jun 2012 Dec 2013

4000
3000

cn

2000
§ 1000

0



o

o



o
o

o



o ©



Kt#/i

° % ° 8 °
j^i§y^8 .ft Q $ jL





Nov 2007 May 2009

O Chemical Precipitation Effluent
O Excluded Data
* Not Detected

Nov 2010
Date

Jun 2012 Dec 2013

5-7

-------
Figure A5.5 Plot of available effluent Nitrate-Nitrite as N and selenium data for Alien after
excluding all data that warranted exclusion except outliers

Allen

05

E

to
TO

(1)

CD

-f-»

CO

15

10

0

Jul 2009

Aug 2010 Sep 2011

Sep 2012 Oct 2013

40

O)

3
E

'c
0)
CD
CO

30
20
10
0

<9

o
o

o o Jr* y ™ * * *** *

*0 *H|P #

Oi
**

oj*°»

• °tSm

o
o

o



Jul 2009

O Bioreactor Eff.
O Excluded Data
* Not Detected

Aug 2010

Sep 2011
Date

Sep 2012

Oct 2013

5-8

-------
Figure A5.6 Plot of available effluent Nitrate-Nitrite as N and selenium data for Belews Creek
after excluding all data that warranted exclusion except outliers

Belews Creek

Jun 2008

Oct 2009

Mar 2011

Jul 2012

Nov 2013

40

O)

E

5

'c
_CD
CD
CD

30
20
10
0

8

o°° °

A°

Jun 2008

O Bioreactor Eff.
* Not Detected

o
o

8

O o ©

° °* aVj? %

ill iMfc H ikik — **•

o o

CO



Oct 2009

Mar 2011
Date

Jul 2012

Nov 2013

5-9

-------
Appendix 6. Data Editing Criteria Res '^s

6-1

-------
Table of Contents

Table	Page

A6.1	Summary of the results for data editing procedure performed to

select a dataset for pollutant at each plant in calculating the limits	 6-5

6-2

-------
This appendix contains a summary of the results of the data editing criteria on a pollutant-by-
pollutant basis to select a dataset at each plant to be used for calculating the effluent limits for each
technology option (described in Section 4 of the document). These criteria are referred to as the
long-term average test (or LTA test). EPA established the LTA test to ensure that the pollutants for
which limits are being set were present in the influent at sufficient concentrations to evaluate
treatment effectiveness at the plant. The data editing procedure is specified as follows: first, the
influent had to pass a basic requirement that 50% of the influent measurements for the pollutant
have to be detected at any concentration. If the dataset for a pollutant at a plant passed the basic
requirement, it then had to pass one of the following two criteria to pass the LTA test:

Criterion 1. At least 50% of the influent measurements in a dataset at a plant are detected at

levels equal to or greater than 10 times the baseline value (shown in Section 3.3).

Criterion 2. At least 50% of the influent measurements in a dataset at a plant are detected at

any concentration and the influent arithmetic average equal to or greater than 10 times the

baseline value (shown in Section 3.3).

If the dataset at a plant failed the basic requirement, then EPA automatically set both Criteria 1 and
2 to "fail." If the dataset for a plant failed the basic requirement, or passed the basic requirement but
failed both criteria, EPA would exclude the plant's effluent data for that pollutant when calculating
the limits. Through the application of the LTA test, EPA ensures that the limits result from
treatment of the wastewater and not simply the absence or substantial dilution of that pollutant in
the waste stream.

Table A6.1 shows how each pollutant at each plant fared with respect to the basic requirement and
the two criteria of the LTA test. The bullets below provide a brief description for the important
columns in the table:

¦	Column "N" presents the total number of influent observations (detected and non-
detected combined)

¦	Column "Percent Detected for Influent (%)" shows the percent of influent
observations that are detected.

¦	Column "Basic Requirement" contains an indicator of whether the dataset passed (or
failed) the basic requirement in the data editing criteria.

¦	Column "Baseline Value" shows the baseline value for each pollutant.

¦	Column "% Influent values >10*baseline value" contains the percent of influent
observations that are at least 10 times the baseline values.

6-3

-------
¦	Column "Criterion 1" contains an indicator of whether the dataset passed (or failed)
criterion 1 of the LTA test.

¦	Column "Mean Influent" shows the mean of the influent concentrations.

¦	Column "Criterion 2" contains an indicator of whether the dataset passed (or failed)
criterion 2 of the LTA test.

The influent sampling location at each plant is given in the bullets below:

¦	FGD Purge: Allen, Belews Creek, Hatfield's Ferry, Keystone, Miami Fort, and Brindisi.

¦	Vapor Compression Evaporator Influent: Polk and Wabash River.

As shown in Table A6.1, the data for each pollutant at each of the plants all passed the LTA test,
except for arsenic and mercury at Wabash River. Since the data for these pollutants failed the LTA
test, EPA excluded the Wabash River data for these pollutants in developing the limits for
gasification wastewater.

6-4

-------
Table A6.1 Summary of the results for data editing procedure performed to select a dataset for pollutant at each plant in calculating
the limits

Plant Name

Pollutant
(unit)

N1

Percent Detected
for Influent (%)

Basic
Requirement2

Baseline
Value

% Influent values
> 10* baseline

(%)

Criterion1

Mean Influent

Criterion2

Allen

Arsenic (Mg/L)

173

100

Passed

2

100

Passed

217.3

Passed

Allen

Mercury (ng/L)

172

100

Passed

0.5

100

Passed

52,073.0

Passed

Allen

Nitrate Nitrite as N (mg/L)

32

100

Passed

0.05

100

Passed

50.5

Passed

Allen

Selenium (Mg/L)

173

100

Passed

5

100

Passed

2,091.5

Passed

Belews Creek

Arsenic (Mg/L)

201

100

Passed

2

100

Passed

288.1

Passed

Belews Creek

Mercury (ng/L)

199

100

Passed

0.5

100

Passed

203,333.2

Passed

Belews Creek

Nitrate Nitrite as N (mg/L)

38

100

Passed

0.05

100

Passed

11.5

Passed

Belews Creek

Selenium (Mg/L)

207

100

Passed

5

100

Passed

5,155.0

Passed

Hatfield's Ferry

Arsenic (Mg/L)

8

100

Passed

2

100

Passed

1,796.3

Passed

Hatfield's Ferry

Mercury (ng/L)

7

100

Passed

0.5

100

Passed

496,285.7

Passed

Keystone

Arsenic (Mg/L)

130

94.6

Passed

2

100

Passed

1,569.9

Passed

Keystone

Mercury (ng/L)

130

93.1

Passed

0.5

100

Passed

438,659.2

Passed

Miami Fort

Arsenic (Mg/L)

8

100

Passed

2

100

Passed

744.8

Passed

Miami Fort

Mercury (ng/L)

62

100

Passed

0.5

100

Passed

286,137.9

Passed

Pleasant Prairie

Arsenic (Mg/L)

16

93.8

Passed

2

87.5

Passed

117.0

Passed

Pleasant Prairie

Mercury (ng/L)

166

100

Passed

0.5

100

Passed

1,382,747.0

Passed

Brindisi

Arsenic (Mg/L)

2

100

Passed

2

100

Passed

54.0

Passed

Brindisi

Mercury (ng/L)

2

100

Passed

0.5

100

Passed

24,000.0

Passed

Brindisi

Selenium (Mg/L)

2

100

Passed

5

100

Passed

255.0

Passed

Brindisi

TDS (mg/L)

2

100

Passed

10

100

Passed

14,000.0

Passed

Polk

Arsenic (Mg/L)

4

100

Passed

2

100

Passed

280.0

Passed

Polk

Mercury (ng/L)

4

100

Passed

0.5

100

Passed

70.4

Passed

Polk

Selenium (Mg/L)

4

100

Passed

5

100

Passed

1,277.5

Passed

Polk

TDS (mg/L)

4

100

Passed

10

100

Passed

4,575.0

Passed

Wabash River

Arsenic (Mg/L)

4

50

Passed

2

0

Failed

4.5

Failed

Wabash River

Mercury (ng/L)

4

0

Failed

0.5

75

Failed

8.7

Failed

Wabash River

Selenium (Mg/L)

4

100

Passed

5

100

Passed

920.0

Passed

Wabash River

TDS (mg/L)

4

100

Passed

10

100

Passed

4,225.0

Passed

1Total number of observations (detected and non-detected combined).

2if a dataset fails the basic requirement, EPA automatically set both Criteria 1 and 2 to Failed.

6-5

-------
Appendix 7. Summary of the Results for the

Engineering Review of the Limits

7-1

-------
Table of Contents

Tables	Page

A7.1	List of all daily values that are above the daily limits for arsenic

and mercury based on chemical precipitation treatment for FGD

wastewater	 7-20

A7.2	List of monthly averages that are above the monthly limitation

for arsenic and mercury and the daily values that went into the

monthly average	 7-25

A7.3	List of all daily values that are above the daily limits for selenium

for FGD wastewater	 7-31

A7.4	List of monthly averages that are above the monthly limitation

for selenium and the daily values that went into the monthly

average	 7-34

A7.5	Listing of all daily values that are above the daily limits for

vapor-compression evaporation treatment for gasification

wastewater	 7-46

A7.6	Summary statistics for the influent (FGD Purge) concentrations

for plants from which the data were used as the basis for
calculating the limits for chemical precipitation treatment

technology option for FGD wastewater	 7-47

A7.7	Summary statistics for the influent (FGD Purge) concentrations

for plants from which the data were used as the basis for
calculating the limits for biological treatment technology option

for FGD wastewater	 7-48

A7.8	Summary statistics for the influent (FGD Purge) concentrations

for plants from which the data were used as the basis for
calculating the limits for vapor-compression evaporation

treatment technology option for FGD wastewater	 7-48

A7.9	Summary statistics for the influent (vapor compression

evaporator influent) concentrations from plants for which the
data were used as the basis for calculating the limits for vapor-
compression evaporation treatment technology option for

gasification wastewater	 7-49

7-2

-------
Table of Contents
(continued)

Figures	Page

A7.1	Arsenic daily limitation and daily concentrations (ng/L) at

Hatfield's Ferry, Keystone, Miami Fort, and Pleasant Prairie used
to calculate the limitations based on chemical precipitation

treatment of FGD wastewater	 7-10

A7.2	Arsenic daily limitation and daily concentrations (ng/L) at

Hatfield's Ferry used to calculate the limitations based on

chemical precipitation treatment of FGD wastewater	 7-11

A7.3	Arsenic daily limitation and daily concentrations (ng/L) at

Keystone used to calculate the limitations based on chemical

precipitation treatment of FGD wastewater	 7-12

A7.4	Arsenic daily limitation and daily concentrations (ng/L) at Miami

Fort used to calculate the limitations based on chemical

precipitation treatment of FGD wastewater	 7-13

A7.5	Arsenic daily limitation and daily concentrations (ng/L) at

Pleasant Prairie used to calculate the limitations based on

chemical precipitation treatment of FGD wastewater	 7-14

A7.6	Mercury daily limitation and daily concentrations (ng/L) at

Hatfield's Ferry, Keystone, Miami Fort, and Pleasant Prairie used
to calculate the limitations based on chemical precipitation

treatment of FGD wastewater	 7-15

A7.7	Mercury daily limitation and daily concentrations (ng/L) at

Hatfield's Ferry used to calculate the limitations based on

chemical precipitation treatment of FGD wastewater	 7-16

A7.8	Mercury daily limitation and daily concentrations (ng/L) at

Keystone used to calculate the limitations based on chemical

precipitation treatment of FGD wastewater	 7-17

A7.9	Mercury daily limitation and daily concentrations at Miami Fort

(ng/L) used to calculate the limitations based on chemical

precipitation treatment of FGD wastewater	 7-18

A7.10 Mercury daily limitation and daily concentrations at Pleasant

Prairie (ng/L) used to calculate the limitations based on chemical

precipitation treatment of FGD wastewater	 7-19

A7.ll Arsenic monthly limitation and monthly average concentrations
(Mg/L) at Hatfield's Ferry and Keystone for chemical

precipitation treatment of FGD wastewater	 7-22

A7.12 Arsenic monthly limitation and monthly average concentrations
(Mg/L) at Hatfield's Ferry for chemical precipitation treatment of
FGD wastewater	 7-22

7-3

-------
Table of Contents
(continued)

Figures	Page

A7.13 Arsenic monthly limitation and monthly average concentrations
(ng/L) at Keystone for chemical precipitation treatment of FGD

wastewater	 7-23

A7.14 Mercury monthly limitation and monthly average concentrations
(ng/L) at Hatfield's Ferry and Pleasant Prairie for chemical

precipitation treatment of FGD wastewater	 7-23

A7.15 Mercury monthly limitation and monthly average concentrations
(ng/L) at Hatfield's Ferry for chemical precipitation treatment of

FGD wastewater	 7-24

A7.16 Mercury monthly limitation and monthly average concentrations
(ng/L) at Pleasant Prairie for chemical precipitation treatment of

FGD wastewater	 7-24

A7.17 Nitrite Nitrate as N daily limitation and daily concentrations
(mg/L) at Allen and Belews Creek used to calculate the

limitations for the biological treatment of FGD wastewater	 7-28

A7.18 Nitrite Nitrate as N daily limitation and daily concentrations

(mg/L) at Allen used to calculate the limitations for the biological

treatment of FGD wastewater	 7-29

A7.19 Nitrite Nitrate as N daily limitation and daily concentrations

(mg/L) at Belews Creek used to calculate the limitations for the

biological treatment of FGD wastewater	 7-29

A7.20 Selenium daily limitation and daily concentrations (ng/L) at Allen
and Belews Creek used to calculate the limitations for biological

treatment of FGD wastewater	 7-30

A7.21 Selenium daily limitation and daily concentrations (ng/L) at Allen
used to calculate the limitations for biological treatment of FGD

wastewater	 7-30

A7.22 Selenium daily limitation and daily concentrations (ng/L) at
Belews Creek used to calculate the limitations for biological

treatment of FGD wastewater	 7-31

A7.23 Nitrite Nitrate as N monthly limitation and monthly average

concentrations (mg/L) at Belews Creek for biological treatment

of FGD wastewater	 7-32

A7.24 Selenium monthly limitation and monthly average concentrations
(ng/L) at Allen and Belews Creek for biological treatment of

FGD wastewater	 7-33

A7.25 Selenium monthly limitation and monthly average concentrations

(ng/L) at Allen for biological treatment of FGD wastewater	 7-33

7-4

-------
Table of Contents
(continued)

Figures	Page

AH .26 Selenium monthly limitation and monthly average concentrations
(ng/L) at Belews Creek for biological treatment of FGD

wastewater	 7-34

A7.27 Arsenic daily limitation and daily concentrations (ng/L) at

Brindisi used to calculate the limitations for Vapor-Compression

Evaporation treatment for FGD wastewater	 7-35

A7.28 Mercury daily limitation and daily concentrations (ng/L) at

Brindisi used to calculate the limitations for Vapor-Compression

Evaporation treatment for FGD wastewater	 7-36

A7.29 Selenium daily limitation and daily concentrations (ng/L) at

Brindisi used to calculate the limitations for Vapor-Compression

Evaporation treatment for FGD wastewater	 7-37

A7.30 Total dissolved solid daily limitation and daily concentrations
(mg/L) at Brindisi used to calculate the limitations for Vapor-

Compression Evaporation treatment for FGD wastewater	 7-38

A7.31 Arsenic daily limitation and daily concentrations (ng/L) at Polk
used to calculate the limitations for the vapor-compression
evaporation treatment technology option for gasification

wastewater	 7-39

A7.32 Mercury daily limitation and daily concentrations (ng/L) at Polk
used to calculate the limitations for the vapor-compression
evaporation treatment technology option for gasification

wastewater	 7-40

A7.33 Selenium daily limitation and daily concentrations (ng/L) at Polk
and Wabash River used to calculate the limitations for the vapor-
compression evaporation treatment technology option for

gasification wastewater	 7-41

A7.34 Selenium daily limitation and daily concentrations (ng/L) at Polk
used to calculate the limitations for the vapor-compression
evaporation treatment technology option for gasification

wastewater	 7-42

A7.35 Selenium daily limitation and daily concentrations (ng/L) at
Wabash River used to calculate the limitations for the vapor-
compression evaporation treatment technology option for

gasification wastewater	 7-43

A7.36 Total dissolved solid daily limitation and daily concentrations
(mg/L) at Polk and Wabash River used to calculate the
limitations for the vapor-compression evaporation treatment
technology option for gasification wastewater	 7-44

7-5

-------
Table of Contents
(continued)

Figures	Page

A7.37 Total dissolved solid daily limitation and daily concentrations
(mg/L) at Polk used to calculate the limitations for the vapor-
compression evaporation treatment technology option for

gasification wastewater	 7-45

A7.38 Total dissolved solid daily limitation and daily concentrations

(mg/L) at Wabash River used to calculate the limitations for the

vapor-compression evaporation treatment technology option for

gasification wastewater	 7-46

7-6

-------
As described in Section 11, to evaluate whether the limits are reasonable, EPA performed an
engineering review to verify that the limits are reasonable based upon the design and expected
operation of the control technologies. EPA performed two types of comparisons for this evaluation.
First, EPA compared the limits to the effluent data used to develop the limits. Second, EPA
compared the limits to the influent data.

Section 1 of this appendix presents the results of the comparisons between the limits and all effluent
data that were used to calculate the limits for each technology option. A series of plots is presented
for each technology option, showing how each of the daily observations compare to the daily limit.
All daily observations that were above the daily limit are summarized in a table following the plots of
daily data. Another series of plots is presented for each technology option, showing how the effluent
values compare to the monthly average limits, for those periods where there were sufficient data to
represent weekly monitoring. Values for those months where the average exceeds the monthly limit,
along with each daily observation for those months, are presented in a table following the plots of
monthly averages.

Section 2 of this appendix contains the results of the comparisons between the limits and influent
data for each technology option.

7-7

-------
i.

Comparison of the Limits to the Effluent Data Used as the
Basis for the Limits

A series of plots is presented for each regulated parameter, showing how each of the daily
observations compare to the daily limit. For limits based on data from more than one plant, the first
plot in a series presents the data for all of the plants (for example, see Figure A7.1). Each grey panel
on that plot has data from one plant, indicated by the label on the horizontal axis. Within each grey
panel, the thin dashed line shows the plant-specific LTA. The wider dashed line that crosses the data
from all plants is the LTA for the technology option. The solid line shows the limitation established
for the parameter and technology option. Within each grey panel the observed concentrations are
ordered by date and plotted by order; thus observations that are relatively far apart in time may plot
next to each other. Following the first plot in a series are additional plots that show the data for each
plant individually (for example, see Figures A7.2 through A7.5). These plots use the sample
collection date for the horizontal axis. As is the case for the first plot in a series, the plots include
horizontal lines depicting the plant-specific LTA, option-level LTA, and the limitation.

Following all plots for a technology option comparing the daily observations to the daily limit, all
daily observations that were above the daily limit are summarized in a table (for example, see Table
A7.1).

Another series of plots is presented for each technology option, showing how the effluent values
compare to the monthly average limits, for those periods where there were sufficient data to
represent weekly monitoring. Values for those months where the average of all daily values is above
the monthly limit, along with each daily observation for those months, are presented in a table
following the plots of monthly averages. Again, a series of plots is presented for each regulated
parameter, showing how the monthly averages calculated for each calendar month compare to the
monthly limit, for those periods where there are sufficient data to represent weekly monitoring. For
limits based on data from more than one plant, the first plot in a series presents the data for all of
the plants (for example, see Figure A7.ll). Each grey panel on that plot has data from one plant,
indicated by the label on the horizontal axis. Following the first plot in a series are additional plots
that show the data for each plant individually, with the sample collection date on the horizontal axis
(for example, see Figures A7.12 and A7.13). As is the case for the first plot in a series, the plots
include horizontal lines depicting the plant-specific LTA, option-level LTA, and the limitation.

7-8

-------
For each plant and analyte, monthly averages for calendar months were calculated using all daily
effluent values within the calendar month. As described in Section 11, this comparison to the
monthly limit requires that the observations are representative of weekly monitoring. EPA
accounted for this by calculating monthly averages for all time periods where there were at least four
daily values with the sample dates spread across a period of at least several weeks (i.e., that spanned
a range of at least 21 days and for which the maximum number of days between sequential samples
was 10 days or less). Data not representative of weekly (or more frequent) monitoring generally were
not valid for this evaluation. As a result, samples collected in a short time frame, or samples that
were not spread across the month were not used when comparing the monthly averages to the
monthly limits.

Following all plots for a technology option comparing the monthly average to the monthly limit, all
results above the monthly limit, along with all daily observations for the month, are summarized in a
table (for example, see Table A7.2).

Arsenic and Mercury Limits for FGD Wastewater, Based on Chemical Precipitation Treatment Technology

The limitations based on chemical precipitation treatment for FGD wastewater were calculated
using data without adjusting for baseline substitution. As a result, the effluent limits were compared
to the reported concentrations without baseline adjustment.

Figures A7.1 through A7.5 show the arsenic daily limitation and daily effluent concentrations used
to calculate the arsenic limitations for FGD wastewater, based on chemical precipitation treatment.
All observations for two plants were equal to or below the daily maximum limit (24 observations at
Keystone; 9 observations at Miami Fort). At Pleasant Prairie, all but one of the 20 observations were
equal to or below the daily limit. At Hatfield's Ferry, 102 observations were equal to or below the
limit; 28 of the 130 observations were above the daily limit.

Figures A7.6 through A7.10 show the mercury daily limitation and daily effluent concentrations used
to calculate the mercury limitations for FGD wastewater, based on chemical precipitation treatment.
All observations for two plants were equal to or below the daily maximum limit (8 observations at
Keystone; 68 observations at Miami Fort). At Pleasant Prairie, 370 of the 375 observations were
equal to or below the daily limit; only 5 observations were above the limit. At Hatfield's Ferry, 217
of the 219 observations were equal to or below the limit; only 2 observations were above the daily
limit. All values above the daily limit are listed in Table A7.1.

7-9

-------
The observations that are equal to or below the arsenic and mercury daily limits can be determined
by comparing the observations in Table A7.1 to the data listed in DCN SE06277.

Figure A7.1 Arsenic daily limitation and daily concentrations (Mg/L) at Hatfield's Ferry, Keystone,
Miami Fort, and Pleasant Prairie used to calculate the limitations based on
chemical precipitation treatment of FGD wastewater

5)BAT Chemical Precipitation Effluent

Arsenic

20

15

o o o

o
o

o

o o

00


O 

00

00

10

5

aoo	ooooo ooo

	O	OOO	-cr - TJET - 	-OCT " ffir Op	OO" '

O O  O OOO O Q ©0 OO
oo ® o o ®o o aoo oo o
—av=DL + Concentration
-------
Figure A7.2 Arsenic daily limitation and daily concentrations (Mg/L) at Hatfield's Ferry used to
calculate the limitations based on chemical precipitation treatment of FGD
wastewater

Hatfields Ferry

Arsenic





o







o







o







o







o o







O OO







o

e

o
o





o

OO OOOOODO OO

OO





ooo ooooo o

o
o



	"O"	

	O-O-O	O		O®	O O (3D- "O	

--XFO--



O m O OQD

O OOO Oq CD O 

o o



 O O OBD

O O <3D O O OO O

OO o











OO O O

° 00 o

o



o o

o +







o



t	1	1	1	1	1	1	1	1	r

Apr Jul Oct Jan Apr Jul Oct Jan Apr Jul Oct Jan
2009	2010	2011	2012

Sample collection date

o

Concentration>=DL + Concentration
-------
Figure A7.3 Arsenic daily limitation and daily concentrations (Mg/L) at Keystone used to
calculate the limitations based on chemical precipitation treatment of FGD
wastewater

10

c S

o
ra

o
O

4

2

Sep Jan May Sep Jan May Sep Jan May Sep Jan May
2010 2011	2012	2013	2014

Sample collection date

o

Concentration>=DL + Concentration
-------
Figure A7.4 Arsenic daily limitation and daily concentrations (Mg/L) at Miami Fort used to
calculate the limitations based on chemical precipitation treatment of FGD
wastewater

Miami Fort
Arsenic

12 -

10

Jul Aug Sep
2010

o Concentration>=DL

Nov Dec Jan Feb Mar Apr May
2011

Sample collection date
-- Facility LTA ...... Option LTA	Option Daily Limit

7-13

-------
Figure A7.5 Arsenic daily limitation and daily concentrations (Mg/L) at Pleasant Prairie used to
calculate the limitations based on chemical precipitation treatment of FGD
wastewater

12 -

10

4 -

Pleasant Prairie

Arsenic

Sep Jan
2009 2010

o Concentration>=DL

Jan
2013

Sample collection date
Facility LTA •••••• Option LTA	Option Daily Limit

7-14

-------
Figure A7.6 Mercury daily limitation and daily concentrations (ng/L) at Hatfield's Ferry,

Keystone, Miami Fort, and Pleasant Prairie used to calculate the limitations based
on chemical precipitation treatment of FGD wastewater

5)BAT Chemical Precipitation Effluent
Mercury

Hatflelds Ferry	Miami Fort

Keystone	Pleasant Prairie

Samples Ordered by date

o

Con:entration>=DL

+

Concentration
-------
Figure A7.7 Mercury daily limitation and daily concentrations (ng/L) at Hatfield's Ferry used to
calculate the limitations based on chemical precipitation treatment of FGD
wastewater

1000

800

600

400

200

Jan
2009

Hatfields Ferry
Mercury

Jan
2010

Jan
2011

Jan
2012

Sample collection date

Concentration>=DL + Concentration
-------
Figure A7.8 Mercury daily limitation and daily concentrations (ng/L) at Keystone used to
calculate the limitations based on chemical precipitation treatment of FGD
wastewater



Keystone



Mercury

800





600 -





c





o





ra





<5 400 -





E





o





200 -













o _











o o O 0



0 -





i i i i i i i i ii
01 Sep 16Sep 01 Oct 160ct 01 Nov 16Nov 01 Dec 16Dec 01 Jan 16Jan

2010 2011



Sample collection date



o Concentration>=DL	Facility LTA	Option LTA 	Option Daily Limit







7-17

-------
Figure A7.9 Mercury daily limitation and daily concentrations at Miami Fort (ng/L) used to
calculate the limitations based on chemical precipitation treatment of FGD
wastewater

Miami Fort
Mercury

800

600

O

5 400 H

=DL

—i	r~

Jan Jul
2011

—i—

Jan
2012

T

Jul

	1	

Jan Jul Jan
2013	2014

Sample collection date
Facility'LTA	Option LTA

Option Daily Limit

7-18

-------
Figure A7.10 Mercury daily limitation and daily concentrations at Pleasant Prairie (ng/L) used to
calculate the limitations based on chemical precipitation treatment of FGD
wastewater

1000

Pleasant Prairie
Mercury

800

o
o

TP"

Jan
2009

Jul

Jan
2010

Jul

Jan
2011

Jul

Jan
2012

Jul

Jan
2013

Jul

Jan
2014

Sample collection date

Con:entration>=DL + Concentration
-------
Table A7.1 List of all daily values that are above the daily limits for arsenic and mercury based
on chemical precipitation treatment for FGD wastewater

Plant

Pollutant (Unit)

Daily Limit

Date

Concentration

Indicator

Hatfield's Ferry

Arsenic (pg/L)

11.00

12/29/2009

12

D

03/02/2010

12

D

04/20/2010

12

D

05/04/2010

19

D

05/11/2010

15

D

06/22/2010

13

D

08/10/2010

13

D

08/24/2010

12

D

08/31/2010

15

D

09/07/2010

14

D

10/12/2010

12

D

11/30/2010

12

D

12/14/2010

13

D

12/21/2010

13

D

12/28/2010

12

D

01/04/2011

12

D

01/11/2011

12

D

01/25/2011

14

D

02/01/2011

16

D

02/08/2011

14

D

03/01/2011

12

D

04/05/2011

13

D

07/12/2011

18

D

07/19/2011

12

D

07/26/2011

22

D

08/02/2011

12

D

11/08/2011

12

D

11/22/2011

12

D

Mercury (ng/L)

788.00

06/05/2012

830

D

12/26/2012

909

D

Pleasant Prairie

Arsenic (pg/L)

11.00

01/30/2013

12

D

Mercury (ng/L)

788.00

06/30/2009

885

D

12/29/2009

905

D

01/13/2010

830

D

02/03/2012

815

D

01/03/2013

855

D

1Daily limit set for each pollutant.
2D = detected and ND = non-detected.

7-20

-------
Figures A7.ll through A7.13 show the monthly average arsenic concentrations for Hatfield's Ferry
and Keystone and the monthly limitation for FGD wastewater, based on chemical precipitation
treatment. Monthly averages could not be calculated for Miami Fort and Pleasant Prairie because
samples were not collected with sufficient frequency at these plants to represent weekly monitoring.
These plots use a format similar to the format used in Figures A7.1 through A7.10, but with one
additional feature that shows each monthly average indicated by a plus (+) sign surrounded by a
circle. The size of the circle is proportional to the number of daily observations that were used to
calculate the monthly average. For Hatfield's Ferry and Keystone, EPA calculated monthly averages
for the time periods where there were sufficient data. For Keystone, all such monthly averages were
below the monthly limit. For Hatfield's Ferry, there were 12 months when the average value was
equal to or below the monthly limit; there were 15 months when the average value was above the
monthly limit (for 8 of those 15 months, the average was only 1-2 ug/L above the limit).

Figures A7.14 through A7.16 show the monthly average mercury concentrations for Hatfield's Ferry
and Pleasant Prairie and the monthly limitation for FGD wastewater, based on chemical
precipitation treatment. Monthly averages for could not be calculated for Keystone and Miami Fort
because samples were not collected with sufficient frequency at these plants to represent weekly
monitoring For Hatfield's Ferry and Pleasant Prairie, EPA calculated monthly averages for the time
periods where there were sufficient data. For Hatfield's Ferry, all but one of the 40 monthly averages
were below the monthly limit For Pleasant Prairie, there were 27 months when the average value
was equal to or below the monthly limit; there were 3 months when the average value was above the
monthly limit (for one of these months, the average is only 4 ng/L above the 356 ng/L limit).
Although Keystone did not collect samples with sufficient frequency to calculate monthly averages,
all of the daily observations for the plant were equal to or below the monthly limit and, therefore,
EPA did not identify any periods of time when the effluent concentrations were higher than the
limit. Miami Fort also did not collect samples with sufficient frequency to calculate monthly
averages; however, it is worth noting that 61 of the 68 daily observations for the plant were below
the monthly limit.

Table A7.2 lists all results above the monthly limits for arsenic and mercury, along with all daily
observations for the month.

7-21

-------
Figure A7.ll Arsenic monthly limitation and monthly average concentrations (Mg/L) at Hatfield's
Ferry and Keystone for chemical precipitation treatment of FGD wastewater

5)BAT Chemical Precipitation Effluent
Monthly Averages, Arsenic

15.0

©

12.5 H

10.0

©

©

©

©© ©



_©	®_®

©

	^





Q ©

©

7.5 H

©

© ©

@

©

©

©

5.0

2.5

Hatfields Ferry
Samples Ordered by date
+ Monthly Mean	Facility LTA ------ Option LTA

©	

©

@

Keystone

Option Monthly Limit

Circle area is proportional to N, N ranges from 4 to 7 sample clays per month

Figure A7.12 Arsenic monthly limitation and monthly average concentrations (Mg/L) at Hatfield's
Ferry for chemical precipitation treatment of FGD wastewater

Hatfields Ferry

Arsenic

Jul
2009

Oct

Jan

2010

Apr Jul Oct

Jan
2011

Apr Jul Oct

Jan
2012

Sample collection date

+ Monthly Average Concentration	Facility LTA ------ Option LTA

	Option Monthly Limit

Circle area is proportional to N, W ranges from 4 to 7 sample days per month

7-22

-------
Figure A7.13 Arsenic monthly limitation and monthly average concentrations (Mg/L) at Keystone
for chemical precipitation treatment of FGD wastewater

Keystone

Arsenic







@







©

©

09Feb 16Feb 23Feb 02Mar 09Mar 16Mar 23Mar 30Mar 06Apr 13Apr 20Apr
2014

Sample collection date

+ Monthly Average Concentration —

— Facility LTA



	Option Monthly Limit





Circle area is proportional to N, N ranges from 4 to 5 sample clays per month

Figure A7.14 Mercury monthly limitation and monthly average concentrations (ng/L) at Hatfield's
Ferry and Pleasant Prairie for chemical precipitation treatment of FGD wastewater

5)BAT Chemical Precipitation Effluent
Monthly Averages, Mercury

600

500

400

@

5 300

© ®" © © ©
" Tl	m 	©...<$©.

200	© "C+i®	„

		©•©	&		 |

100



0

Hatfields Ferry	Pleasant Prairie

Samples Ordered by date

+ Monthly Mean	Facility' LTA	Option LTA 	Option Monthly Limit

Circle area is proportional to N, N ranges from A to 12 sample days per month

7-23

-------
Figure A7.15 Mercury monthly limitation and monthly average concentrations (ng/L) at Hatfield's
Ferry for chemical precipitation treatment of FGD wastewater

Hatfields Ferry
Mercury

400

300

200

100

Jul Jan
2009 2010

Jan
2012

Jan
2013

Jan
2014

Sample collection date

+ Monthly Average Concentration 	- Facility LTA ------ Option LTA

	Option Monthly Limit

Circle area is proportional to N, W ranges from 4 to 11 sample days per month

Figure A7.16 Mercury monthly limitation and monthly average concentrations (ng/L) at Pleasant
Prairie for chemical precipitation treatment of FGD wastewater

600

500

£ 400

u 300

200

100

Pleasant Prairie
Mercury

Jan Jul Jan Jul
2009	201 0

Jan Jul Jan Jul
201 1	2012

Jan Jul
2013

Jan
2014

Sample collection date

+ Monthly Average Concentration	Facility LTA		 Option LTA

	Option Monthly Limit

Circle area is proportional to N, W ranges from 7 to 12 sample days per month

7-24

-------
Table A7.2 List of monthly averages that are above the monthly limitation for arsenic and
mercury and the daily values that went into the monthly average





Daily

Monthly

Monthly





Plant

Pollutant (Unit)

Limit

Limit

Average

Date

Concentration











03/02/2010

12.0











03/09/2010

11.0









9.800

03/16/2010

9.0











03/23/2010

7.0











03/30/2010

10.0











04/06/2010

9.0









10.000

04/13/2010

10.0









04/20/2010

12.0











04/27/2010

9.0











05/04/2010

19.0









12.000

05/11/2010

15.0









05/18/2010

8.0











05/25/2010

6.0











06/01/2010

8.0











06/08/2010

11.0









9.600

06/15/2010

10.0

Hatfield's Ferry

Arsenic (pg/L)

11.00

8.00



06/22/2010

13.0



06/29/2010

6.0











08/03/2010

11.0











08/10/2010

13.0









12.400

08/17/2010

11.0











08/24/2010

12.0











08/31/2010

15.0











09/07/2010

14.0









10.750

09/14/2010

11.0









09/21/2010

9.0











09/28/2010

9.0











12/07/2010

9.2











12/08/2010

4.6











12/09/2010

6.4









9.400

12/10/2010

7.6











12/14/2010

13.0











12/21/2010

13.0











12/28/2010

12.0

7-25

-------
Table A7.2 List of monthly averages that are above the monthly limitation for arsenic and
mercury and the daily values that went into the monthly average (continued)





Daily

Monthly

Monthly





Plant

Pollutant (Unit)

Limit

Limit

Average

Date

Concentration











01/04/2011

12.0











01/11/2011

12.0









11.800

01/12/2011

11.0











01/18/2011

10.0











01/25/2011

14.0











02/01/2011

16.0











02/08/2011

14.0









12.000

02/09/2011

10.0











02/15/2011

9.0











02/22/2011

11.0











03/01/2011

12.0











03/08/2011

9.0









9.833

03/10/2011

10.0









03/15/2011

11.0











03/22/2011

7.0











03/29/2011

10.0











04/05/2011

13.0









9.500

04/12/2011

8.0



Arsenic (pg/L)

11.00

8.00

04/19/2011

8.0



(continued)



04/26/2011

9.0

Hatfield's Ferry









07/05/2011

7.0

(continued)







14.750

07/12/2011

18.0









07/19/2011

12.0











07/26/2011

22.0











08/02/2011

12.0











08/09/2011

10.0









11.000

08/16/2011

11.0











08/23/2011

11.0











08/30/2011

11.0











10/04/2011

9.0









9.000

10/11/2011

10.0









10/18/2011

8.0











10/25/2011

9.0











11/01/2011

5.0











11/08/2011

12.0









8.800

11/15/2011

7.0











11/22/2011

12.0











11/29/2011

8.0











12/04/2012

102.0



Mercury (ng/L)

788.00

356.00

431.750

12/11/2012

272.0



12/18/2012

444.0











12/26/2012

909.0

7-26

-------
Table A7.2 List of monthly averages that are above the monthly limitation for arsenic and
mercury and the daily values that went into the monthly average (continued)





Daily

Monthly

Monthly





Plant

Pollutant (Unit)

Limit

Limit

Average

Date

Concentration











07/01/2009

605.0











07/08/2009

445.0











07/09/2009

265.0











07/14/2009

345.0









378.333

07/15/2009

330.0











07/21/2009

355.0











07/22/2009

355.0











07/28/2009

325.0











07/29/2009

380.0











01/06/2010

425.0











01/08/2010

395.0











01/12/2010

770.0

Pleasant Prairie

Mercury (ng/L)

788.00

356.00

588.750

01/13/2010

830.0









01/18/2010

415.0











01/20/2010

400.0











01/26/2010

780.0











01/27/2010

695.0











03/08/2011

195.0











03/09/2011

335.0











03/18/2011

665.0









360.000

03/19/2011

415.0









03/25/2011

295.0











03/26/2011

210.0











03/29/2011

410.0











03/30/2011

355.0

1Daily limit set for each pollutant.
2Monthly limit set for each pollutant.

Nitrate-nitrite as N and Selenium Limits forFGD Wastewater, Based on Chemical Precipitation Followed by
Biological Treatment

The limitations for nitrate-nitrite as N and selenium based on chemical precipitation followed by
biological treatment were calculated using data without adjusting for baseline substitution. As a
result, the effluent limits were compared to the reported concentrations without baseline adjustment.

Figures A7.17 through A7.19 show the nitrate-nitrite as N daily limitation and daily effluent
concentrations used to calculate the limitations for FGD wastewater. All observations for both Allen

7-27

-------
and Belews Creek were equal to or below the daily maximum limit (30 observations at Allen; 40
observations at Belews Creek).

Figures A7.20 through A7.23 show the selenium daily limitation and daily effluent concentrations
used to calculate the limitations for FGD wastewater. At Allen, 178 of the 182 observations were
equal to or below the daily limit; only 4 observations were above the limit. At Belews Creek, 214 of
the 216 observations were equal to or below the limit; only 2 observations were above the daily limit.
All values above the daily limit are listed in Table A7.3.

The observations that are equal to or below the nitrate-nitrite as N and selenium daily limits can be
determined by comparing the observations in Table A7.3 to the data listed in DCN SE06277.

Figure A7.17 Nitrite Nitrate as N daily limitation and daily concentrations (mg/L) at Allen and
Belews Creek used to calculate the limitations for the biological treatment of FGD
wastewater

6)BAT Biological Treatment Effluent
Nitrate Nitrite as N

15

'¦a 10

CO

0

+o I I I I I +

000+++++++++++

W--H-J300++ I I I I I I I I I I I I I I I I I I I I I I I I 10++++

Allen

Belews Creek

Samples Ordered by date

o Concentration>=DL + Concentration
-------
Figure A7.18 Nitrite Nitrate as N daily limitation and daily concentrations (mg/L) at Allen used to
calculate the limitations for the biological treatment of FGD wastewater

15 -

"ro 10-

"c
0)
o
=DL + Concentration=DL + Concentiation
-------
Figure A7.20 Selenium daily limitation and daily concentrations (Mg/L) at Allen and Belews Creek
used to calculate the limitations for biological treatment of FGD wastewater

6)BAT Biological Treatment Effluent

Selenium

30

o
o

20

o
o

10

o
o

o v

o

o

I ¦ I I ¦	¦¦¦¦¦¦¦!

I ¦ I II ¦¦¦¦¦¦¦¦¦¦I

o
o

-H-

¦	• • • — # - q-	• * OCT. *•••••

0

^p+°

+

¦ ¦¦¦IB II	I ¦ ¦ I	I M ¦ I I 11	I I I ¦¦ II I

$-H

Allen

Belews Creek

Samples Ordered by date

5

HBO

Concentration>=DL + Concentration¦ ii 11 i i 11 ill i r

i	1	1	1	1	1	1	1	1	r

Jul Jan Jul Jan Jul	Jan Jul	Jan Jul	Jan

2009 2010	2011	2012	2013	2014

Sample collection date

o

Concentration>=DL + Concentration
-------
Figure A7.22 Selenium daily limitation and daily concentrations (|Jg/L) at Belews Creek used to
calculate the limitations for biological treatment of FGD wastewater

Belews Creek

Selenium

30 -

o

o



o











o









20 -

Q











o



o









o









Cb °
o

+



O
O

Q

o



o

o

o
©

©

o

10 -

¦ i 11 i	|_i j

11 IT ---1 llllilllllilll i ri

+ +

+-HH- ++

-Hf





	

-rQr---Q





b. pfi.	.JS .

0 -



o

° ~ 1a %- +-W-
o +o +
°°o+ +

Q

+fH+ O Till II 1 1

b

-H- +

oJ% o

¦WW Mill IIII fffi 4P

Jan Jul Jan Jul Jan Jul Jan Jul Jan Jul Jan Jul Jan
2008	2009	2010	2011	2012	201 3	2014

Sample collection date

o

Concentration>=DL + Concentration
-------
average. At Belews Creek, there were 10 months when the monthly average was equal to or below
the monthly limit; there were 2 months when the average was above the limit. Both of these months
occurred shortly after the end of the initial commissioning period for the treatment system.

Table A7.4 lists all results above the monthly limit for selenium, along with all daily observations for
the month.

Figure A7.23 Nitrite Nitrate as N monthly limitation and monthly average concentrations (mg/L)
at Belews Creek for biological treatment of FGD wastewater

4

3 -

L=

O

ra

cr

a) _

u Z -

sz

o

1

0

Q8Sep 15Sep 22Sep 29Sep O60ct 130ct 200ct 270ct 03Nov 10Nov 17Nov
2013

Sample collection date

+ Monthly Average Concentration —

	FacilityLTA









Circle area is proportional to N, N ranges from 5 to 9 sample days per month

Belews Creek
Nitrate Nitrite as N

q	0	.0

7-32

-------
Figure A7.24 Selenium monthly limitation and monthly average concentrations (Mg/L) at Allen
and Belews Creek for biological treatment of FGD wastewater

15

£ 10

5

6)BAT Biological Treatment Effluent
Monthly Averages, Selenium

©





© @ © ©

0

©

©

©

©

Allen Belews Creek
Samples Ordered by date
+ MonthlyMean	Facility LTA 	 Option LTA 	

Option Monthly Limit

Circle area is proportional to N, W ranges from 4 to 13 sample days per month

Figure A7.25 Selenium monthly limitation and monthly average concentrations (Mg/L) at Allen for
biological treatment of FGD wastewater

Allen

Selenium

May Sep Jan May Sep Jan May Sep Jan May Sep
2009	2010	2011	2012

Sample collection date

+ Monthly Average Concentration	Facility LTA	Option LTA

	Option Monthly Limit

Circle area is proportional to N, W ranges from A to 13 sample days per month

7-33

-------
Figure A7.26 Selenium monthly limitation and monthly average concentrations (|Jg/L) at Belews
Creek for biological treatment of FGD wastewater

Belews Creek
Selenium

17.5 -

©





15.0







12.5 -









©





10.0



©



7.5 -
5.0





©

+

©©

T

Jul Jan Jul Jan Jul Jan Jul Jan Jul Jan Jul Jan
2008 2009	2010	201 1	2012	2013	2014

Sample collection date

+ Monthly Average Concentration —

— Facility-LTA ---









Circle area is proportional to N, N ranges from 4 to 12 sample days per month

Table A7.4 List of monthly averages that are above the monthly limitation for selenium and the
daily values that went into the monthly average

Plant

Pollutant (Unit)

Daily Limit

Monthly Limit

Monthly Average

Date

Concentration











08/04/2008

10.0









15.875

08/11/2008

15.0









08/18/2008

18.9











08/25/2008

19.6

Belews Creek

Selenium (pg/L)

23.00

12.00



09/02/2008

12.5











09/08/2008

21.3









16.840

09/15/2008

22.4











09/22/2008

10.1











09/29/2008

17.9

^-Daily limit set for each pollutant.
iMonthly limit set for each pollutant.

7-34

-------
Limits for FGD Wastewater Based on Chemical Precipitation Followed By I Tapor-Compression Evaporation
(Arsenic, Mercury, Selenium, and TPS)

The limitations for vapor-compression evaporation of FGD wastewater were calculated using
baseline adjusted data. As a result, the effluent limits were compared to the reported concentrations
after baseline adjustment.

Figures A7.27 through A7.30 show the daily limitation and daily concentrations from Brindisi for
vapor-compression evaporation of FGD wastewater for arsenic, mercury, selenium, and total
dissolved solids, respectively. These data are for the crystallizer condensate sampling location used
for calculating the limitations. All observations (for both the brine concentrator distillate and
crystallizer condensate) are below the daily limits. Note that although monthly average values were
not calculated, all observations were below the monthly limits.

Figure A7.27 Arsenic daily limitation and daily concentrations (Mg/L) at Brindisi used to calculate
the limitations for Vapor-Compression Evaporation treatment for FGD wastewater

Brindisi

Arsenic

4.2

a
o

c 41

-------
Figure A7.28 Mercury daily limitation and daily concentrations (ng/L) at Brindisi used to calculate
the limitations for Vapor-Compression Evaporation treatment for FGD wastewater

Brindisi
Mercury

40

30

20



10

05 Apr
2011

o Concentration>=DL

06 Apr

Sample collection date
Facility LTA	Option LTA

07 Apr

Option Daily Limit

7-36

-------
Figure A7.29 Selenium daily limitation and daily concentrations (Mg/L) at Brindisi used to

calculate the limitations for Vapor-Compression Evaporation treatment for FGD
wastewater



Brindisi



Selenium

5.2





a





o





-t—¦

n





c=





o





5 5.1





en





U.U





1 1 1

05Apr 06Apr Q7Apr

2

011



Sample collection date



+ Concentration
-------
Figure A7.30 Total dissolved solid daily limitation and daily concentrations (mg/L) at Brindisi

used to calculate the limitations for Vapor-Compression Evaporation treatment for
FGD wastewater

50

40

O
ro

§ 30

a

20

10

05Apr	06Apr	07Apr

2011

Sample collection date

o

Concentration>=DL + Concentration
-------
Figures A7.33 through A7.35 show the selenium daily limitation and daily effluent concentrations
used to calculate the limitations for gasification wastewater. All observations from Wabash River
were below the daily limit. For Polk, three of the four observations were below the daily limit.

Figures A7.36 through A7.38 show the total dissolved solid daily limitation and daily effluent
concentrations used to calculate the limitations for gasification wastewater. All observations for both
plants were below the daily limits.

Table A7.5 lists all daily data that are above the maximum daily limits for each pollutant.

There were too few observations to calculate monthly averages for comparison to the monthly limits
for gasification wastewater.

Figure A7.31 Arsenic daily limitation and daily concentrations (Mg/L) at Polk used to calculate the
limitations for the vapor-compression evaporation treatment technology option for
gasification wastewater

4.2

o

OJ

c 4 1

ai
o
c
o
O

4.0

Polk

Arsenic

1 BOct 190ct 20Oct 21 Oct 220ct 230ct 240ct 250ct 260ct 270ct
2011

Sample collection date

+ Concentration
-------
Figure A7.32 Mercury daily limitation and daily concentrations (ng/L) at Polk used to calculate
the limitations for the vapor-compression evaporation treatment technology option
for gasification wastewater

1.75

1.50

£Z
Q

ra
tz

CD

£ 1.25

o
O

1.00

0.75

1 SOct 190ct 20Oct 21 Oct 220ct 230ct 240ct 250ct 260ct 270ct
2011

Sample collection date

o

Concentration>=DL —

	FacilityLTA

	Option LTA



- Option Daily Limit





7-40

-------
Figure A7.33 Selenium daily limitation and daily concentrations (Mg/L) at Polk and Wabash River
used to calculate the limitations for the vapor-compression evaporation treatment
technology option for gasification wastewater

8)BAT Mechanical Vapor Compression for Gassification

Selenium

500

o

400

o

o 300

"11
(J

o 200 -

Q

	t)'

100 -



1 1

Polk Wabash River





Samples Ordered by date



o

Concentration>=DL + Concentration
-------
Figure A7.34 Selenium daily limitation and daily concentrations (Mg/L) at Polk used to calculate
the limitations for the vapor-compression evaporation treatment technology option
for gasification wastewater

Polk

Selenium

500

400

!=
CD
O

c
o
O

300

200

1 SOct 190ct 20Oct 21 Oct 220ct 230ct 240ct 250ct 260ct 270ct
2011

o Concentration>=DL

Sample collection date
Facility LTA •••••• Option LTA	Option Daily Limit

7-42

-------
Figure A7.35 Selenium daily limitation and daily concentrations (Mg/L) at Wabash River used to
calculate the limitations for the vapor-compression evaporation treatment
technology option for gasification wastewater

Wabash River
Selenium

CLl
O

O

O

500

400

300

200

100

0

21 Feb
2011

22Feb

Sample collection date

o Concentration>=DL + Concentration
-------
Figure A7.36 Total dissolved solid daily limitation and daily concentrations (mg/L) at Polk and
Wabash River used to calculate the limitations for the vapor-compression
evaporation treatment technology option for gasification wastewater

8)BAT Mechanical Vapor Compression for Gassification
Total Dissolved Solid

Polk

Wabash River

Samples Ordered by date

Concentration>=DL + Concentration
-------
Figure A7.37 Total dissolved solid daily limitation and daily concentrations (mg/L) at Polk used to
calculate the limitations for the vapor-compression evaporation treatment
technology option for gasification wastewater

Polk

Total Dissolved Solid

40

35

30 -

25

20

15

10

	1	1	1	1	1	1	1	1	r

180ct 190ct 20Oct 21 Oct 220ct 230ct 240ct 250ct 260ct 270ct
2011

o Concentration>=DL

Sample collection date
Facility LTA ------ Option LTA 	 Option Daily Limit

7-45

-------
Figure A7.38 Total dissolved solid daily limitation and daily concentrations (mg/L) at Wabash
River used to calculate the limitations for the vapor-compression evaporation
treatment technology option for gasification wastewater

40

35

30 -

a
o

"S

1= 25

O)

o
c
o
O

20 -

15 -

10

21 Feb	22Feb	23Feb	24Feb

2011

Sample collection date

o

Concentration>=DL + Concentration
-------
2.

Comparison of the Effluent Limits to Influent Data

In addition to comparing the limits to the effluent data used to develop the limits, EPA also
compared the limits to the influent data for the plants. Sections below provide the influent summary
statistics for all plants from which the data were used as the basis for calculating the limits.

Arsenic and Mercury Limits forFGD Wastewater. Based on Chemical Precipitation Treatment Technology

Table A7.6 presents summary statistics for the treatment system influent (FGD Purge) for arsenic
and mercury at Hatfield's Ferry, Keystone, Miami Fort, and Pleasant Prairie. Also provided in the
table are the daily maximum limits for each pollutant.

Table A7.6 Summary statistics for the influent (FGD Purge) concentrations for plants from
which the data were used as the basis for calculating the limits for chemical
precipitation treatment technology option for FGD wastewater

Pollutant
(daily limits)

Plant Name

Summary Statistics for influent by Pollutant and Plant Name

N

Minimum

Median

Mean

Maximum

Arsenic

(Daily Limit = 11

Ug/L)

Hatfield's Ferry

8

300.0

1,575.0

1,796.3

4,610.0

Keystone

130

200.0

1,580.0

1,569.9

5,250.0

Miami Fort

8

320.0

729.8

744.8

1,330.0

Pleasant Prairie

16

10.7

145.0

117.0

187.0

Mercury

(Daily Limit = 788

ng/L)

Hatfield's Ferry

8

184,000.0

467,000.0

465,125.0

789,000.0

Keystone

130

200.0

434,000.0

438,659.2

950,000.0

Miami Fort

62

22,000.0

277,500.0

286,137.9

1,065,000

Pleasant Prairie

166

120,000.0

1400,000.0

1,382,747.0

4200,000.0

Nitrate-nitrite as N and Selenium Limits forFGD Wastewater, Based on Chemical Precipitation Followed By
Biological Treatment

Table A7.7 presents the influent (FGD Purge) summary statistics for all plants from which the data
were used as the basis for calculating the limits. Also provided in the table are the daily limits for
each pollutant.

7-47

-------
Table A7.7 Summary statistics for the influent (FGD Purge) concentrations for plants from
which the data were used as the basis for calculating the limits for biological
treatment technology option for FGD wastewater

Pollutant
(daily limits)

Plant Name

Summary Statistics for influent by Pollutant
and Plant Name

N

Minimum

Median

Mean

Maximum

Nitrate-nitrite as N
(Daily Limit = 17.0 mg/L)

Allen

32

2.7

42.5

50.5

130.0

Belews Creek

38

0.5

13

11.5

23

Selenium

(Daily Limit = 23 pg/L)

Allen

184

572.0

1,655.0

2,089.2

10,600.0

Belews Creek

217

274.0

4,820.0

5,297.2

26,200.0

Limits for FGD Wastewater Based on Chemical Precipitation Followed By I Tapor-Compression Evaporation
(Arsenic. Mercury. Selenium, and TPS)

Table A7.8 presents the influent (FGD Purge) summary statistics for all plants from which the data
were used as the basis for calculating the limits. Also provided in the table are the daily limits for
each pollutant.

Table A7.8 Summary statistics for the influent (FGD Purge) concentrations for plants from
which the data were used as the basis for calculating the limits for vapor-
compression evaporation treatment technology option for FGD wastewater

Pollutant
(daily limits)

Plant Name

Summary Statistics for influent by Pollutant and Plant Name

N

Minimum

Median

Mean

Maximum

Arsenic

(Daily Limit = 4 pg/L)

Brindisi

2

53.0

54.0

54.0

55.0

Mercury

(Daily Limit = 39 ng/L)

Brindisi

2

21,100.0

24,000.0

24,000.0

26,900.0

Selenium

(Daily Limit = 5 pg/L)

Brindisi

2

220.0

255.0

255.0

290.0

Total Dissolved Solids
(Daily Limit = 50 mg/L)

Brindisi

2

13,000.0

14,000.0

14,000.0

15,000.0

Limits for Gasification Wastewater Based on I rapor-Compression Evaporation (Arsenic. Mercury. Selenium, and
TPS)

Table A7.9 presents the influent summary statistics for all pollutants at Polk and Wabash River. Also
provided in the table is the daily limit for each pollutant. As mentioned above, the data for arsenic
and mercury at Wabash River failed the data editing criteria, thus the data for these two pollutants at
Wabash River were excluded from developing the limits for this technology option. However, for
completeness, the summary statistics for arsenic and mercury are provided in the table below.

7-48

-------
Table A7.9 Summary statistics for the influent (vapor compression evaporator influent)

concentrations from plants for which the data were used as the basis for calculating
the limits for vapor-compression evaporation treatment technology option for
gasification wastewater

Pollutant
(daily limits)

Plant Name

Summary Statistics for influent by Pollutant and Plant Name

N

Minimum

Median

Mean

Maximum

Arsenic

(Daily Limit = 4 pg/L)

Polk

4

220.0

280.0

280.0

340.0

Wabash River1

4

4.0

4.5

4.5

5.0

Mercury

(Daily Limit = 1.8 ng/L)

Polk

4

17.0

85.9

70.4

92.7

Wabash River1

4

5.0

9.9

8.7

9.9

Selenium

(Daily Limit = 453 pg/L)

Polk

4

720.0

1,295.0

1,277.5

1,800.0

Wabash River

4

800.0

890.0

920.0

1,100.0

Total Dissolved Solids
(Daily Limit = 38 mg/L)

Polk

4

4,500.0

4,600.0

4,575.0

4,600.0

Wabash River

4

3,600.0

4,400.0

4,225.0

4,500.0

1 Arsenic and Mercury data from Wabash River were not used to calculate the limits.

7-49

-------