September 23, 2008

MEMORANDUM

SUBJECT:	Technical Support Document (TSD) for the Doe Run Herculaneum,
Missouri, Lead State Implementation Plan Revision

FROM:	Gwen Yoshimura

		Richard L. Daye

THRU:	Joshua Tapp

TO:		File MO-259; EPA-R07-OAR-2008-0538

Table of Contents

Regulatory Background

Policy Analyses

Section 110 Requirements

Section 172 Requirements

Technical Background

Nonattainment Boundaries

Location and Topography

Meteorology

Wind

Precipitation and Temperature

Summary of Air Quality Data

Source Characterization

Development of Dispersion Model Inputs

Model Input Development

Model Selection

Study Period

Meteorological Data

Emissions Inventory

Actual Value, Design Value, and Control Strategy Dispersion Modeling

 Actual Value Dispersion Modeling & Emissions Reconciliation

Design Value Dispersion Modeling

Control Strategy Dispersion Modeling

Modeled Background

Control Strategy/Contingency Measures

Source Contributions

Description of Selected Control Measures

Concentrate Delivery 

Railcar Tipper Building

Sinter Mix Room Fume, Sinter Mix Room Concentrate and Sinter

Process Building Fugitives: Sinter Plant, Blast Furnace, and Dross &
Refinery Buildings

Sinter Plant Building Fugitives

Blast Furnace Building Fugitives

Dross & Refinery Building Fugitives

Kettle Heat Stack Cameras and Refinery Kettle Heat Stack

Number 5 Baghouse Fume Handling

Roads: Street Sweeping and Unpaved Haul Road

Street Sweeping

“North End of Blast Furnace to Refinery Dock” Haul Road

Reasonably Available Control Measures (RACM) including Reasonably
Available Control Technology (RACT)

Reasonable Further Progress (RFP)

Contingency Measures

Contingency Measure Schedule

Contingency Measure Impacts

Implementation of Contingency Measures

Enforceability

Enforceable Requirements

Process Weight Limits

Hours of Operation, Moisture Content, and Chemical Stabilization

Emission Limits

Ventilation Requirements

Roads

Reporting

Sanctions and Penalties

Process for Making Substitutions, Alterations, or Additions to the SIP

Continual Improvement

Regulatory Background

The Environmental Protection Agency (EPA) established the National
Ambient Air Quality Standard (NAAQS) for lead on October 5, 1978 (43 FR
46246).  The NAAQS for lead is set at a level of 1.5 micrograms (µg) of
lead per cubic meter (m3) of air, averaged over a calendar quarter.

	During the 1980s and 1990s, Missouri submitted and EPA approved a
number of State Implementation Plan (SIP) revisions for lead to address
ambient lead concentrations in various areas of the state.  One such
area was Herculaneum, Missouri, where a primary lead smelter has been in
operation since 1892.  The primary lead smelter is currently owned and
operated by the Doe Run Resources Company (hereafter referred to as
“Doe Run”).  Doe Run-Herculaneum is the only currently operating
primary lead smelter in the United States.  

The city of Herculaneum was designated nonattainment for lead in 1991
(56 FR 56694, November 6, 1991, codified at 40 CFR 81.326), pursuant to
new authorities provided by the Clean Air Act Amendments of 1990 (CAA or
Act).  The state also became subject to new SIP requirements under part
D, Title I of the Act, added by the 1990 amendments.  A revised SIP
meeting the part D requirements was subsequently submitted in 1994.  The
plan established 

June 30, 1995, as the date by which the Herculaneum area was to attain
compliance with the lead standard.  However, the plan did not result in
attainment of the standard and monitored lead concentrations in the
Herculaneum area continued to show exceedances of the standard. 
Therefore, on August 15, 1997, after taking and responding to public
comments, EPA published a notice in the Federal Register (62 FR 43647)
finding that the Herculaneum nonattainment area had failed to attain the
lead standard by the June 30, 1995, deadline.

On January 10, 2001, Missouri submitted a revised SIP to EPA for the
Herculaneum area.  The SIP contained control measures to reduce lead
emissions to attain the standard, including building enclosure and
ventilation projects, implementation of work practice standards, process
throughput restrictions, and hours of operation limitations.  As
required by section 172(c)(9) of the Act, the plan also included
contingency measures to be implemented in the event that there were
future exceedances of the lead standard in Herculaneum.  The first
contingency measure included enclosures and installation of additional
process controls.  This measure was to be implemented within six months
following the calendar quarter in which the exceedance occurred.  If
there was a second exceedance of the quarterly lead standard, after
implementation of the initial contingency measure, Doe Run-Herculaneum
would curtail production utilizing one of three emission and/or
production curtailing methods.  A 2000 Work Practices Manual, 2001
Consent Judgment, and Missouri rule 10 CSR 10-6.120 “Restriction of
Emissions of Lead from Specific Lead Smelter-Refinery Installations”
were also submitted as part of the SIP submittal.  

The SIP established August 14, 2002, as the attainment date for the
area.  The plan included permitting, monitoring, and reporting
requirements, an emissions inventory, implementation of all reasonably
available control measures as expeditiously as practicable, provided for
attainment of the NAAQS as demonstrated using modeling, provisions for
reasonable further progress and implementation of contingency measures,
and assurances that the state would be able to implement the plan,
thereby satisfying the CAA section 172(c) nonattainment plan provision
requirements.  EPA approved the SIP on April 16, 2002 (67 FR 18497).  

Doe Run and the Missouri Department of Natural Resources (MDNR) operate
co-located monitors at the Broad Street and Main Street/City Hall
monitoring locations (in addition to other lead monitoring locations in
the nonattainment area).  These monitors are used to show whether or not
the area is in attainment of the standard.  Following the August 2002
attainment date, the Herculaneum area monitored attainment of the lead
standard for 10 consecutive calendar quarters.  In 2005, air quality
monitors in the area again reported exceedances of the 1.5 µg/m3 lead
NAAQS in the first two calendar quarters in 2005.  Monitored values are
quality assured by MDNR and properly entered into the Air Quality
System, EPA’s repository for ambient air monitoring data.  The values
for the first two quarters of 2005 exceed the 1.5 µg/m3 lead standard
and therefore constitute exceedances of the standard for each quarter.  

Typically, an exceedance would trigger implementation of a contingency
measure.  After the 2002 SIP revisions were approved, Doe
Run-Herculaneum implemented both sets of contingency measures.  The
first set of contingency measures was implemented by Doe Run prior to
any monitored exceedances of the lead NAAQS.  This was done in an effort
to reduce emissions earlier than required and beyond what was required. 
Specifically, Doe Run completed the following measures: (1) modification
to the cooler baghouse dilution air intake completed on December 31,
2002, (2) sealing of roof monitor in the Sinter Plant Mixing Room
(sealed instead of installing passive filters as originally described)
completed October 31, 2003, (3) enclosure of north end of the railcar
unloader building to prevent wind blow-through fugitive emissions
completed April 30, 2004 (this contingency measure was altered with MDNR
approval from the original requirement to “Enclose railcar fume
loading station at Number 5 Baghouse” due to modification of Doe
Run’s operation), (4) enclosure of the north end of the railcar
unloader completed April 30, 2004, (5) enclosure of the north end number
1 trestle and bin storage area completed July 31, 2002, and (6)
modification of inlet ducting to number 3 baghouse by removing number 12
fan restriction from ducting completed December 31, 2001.  The second
contingency measure was implemented following exceedance of the lead
standard in the first and second calendar quarters of 2005.  Doe
Run-Herculaneum selected the option of limiting production to 50,000
tons per quarter of finished lead under the second contingency measure. 
Despite implementation of all contingency measures, air monitors in
Herculaneum recorded values above the 1.5 µg/m3 lead standard in the
third quarter of 2005.

Because the exceedance recorded in the third quarter of 2005 occurred
despite implementation of all the control measures contained in the SIP,
including all contingency measures developed and implemented to address
exceedances, EPA proposed a SIP call on December 19, 2005.  The SIP call
proposed to find the SIP substantially inadequate to attain and maintain
the NAAQS for lead and proposed to require the state to revise the lead
SIP for Herculaneum.  

EPA finalized the SIP call on April 14, 2006 (71 FR 19432).  The SIP
call notified the state of EPA’s finding that the SIP was
substantially inadequate to provide for attainment and maintenance of
the lead NAAQS in Herculaneum, and required the state to submit a
revised SIP.  Section 110(k)(5) of the CAA provides that after EPA makes
a finding that a plan is substantially inadequate, it may establish a
reasonable deadline for correcting the deficiencies, but the date can be
no later than 18 months after the state is notified of the finding. 
Based on a number of considerations detailed in the final rule for the
SIP call (71 FR 19432, published April 14, 2006), the SIP call required
submission of the revisions within twelve months following date of
signature of the final rulemaking.  

Along with a deadline for SIP submittal by the state to EPA, the final
SIP call established the date by which the state must demonstrate
attainment of the standard in Herculaneum.  Sections 110(k)(5) and
172(d) provide that EPA may adjust any SIP deadlines that are applicable
under the Act, except that the attainment date may not be adjusted
unless it has elapsed.  For Herculaneum, the attainment date had been
August 2002 (five years after the state was notified that the area
failed to attain).  The attainment date had elapsed, and the area was
not attaining the standard.  The attainment date could therefore be
adjusted pursuant to section 110(k)(5) and section 172(d), and the state
was required to provide for attainment as expeditiously as practicable. 
Based on information described in the final SIP call rule (71 FR 19432,
published April 14, 2006), EPA established an attainment date of April
7, 2008, two years from the date of signature of the final rulemaking. 
MDNR formally commented in support of the timelines contained in the SIP
call, including the SIP submittal deadline and attainment date.

	EPA required MDNR to submit several specific plan elements to EPA in
order to correct the inadequacy of the SIP.  These specific elements
were: (1) A revised emissions inventory, (2) a modeling demonstration
showing what reductions would be needed to bring the area back into
attainment of the lead NAAQS, (3) adopted measures to achieve the
reductions determined necessary by the modeled attainment demonstration,
with enforceable schedules for implementing the measures as
expeditiously as practicable, and (4) contingency measures meeting the
requirements of Section 172(c)(9) of the CAA.

MDNR completed its revision to the SIP, and on April 26, 2007, the
Missouri Air Conservation Commission approved the SIP revision after
completing the required public notification, public hearing and comment
period. On May 31, 2007, EPA received Missouri’s revised SIP for the
Herculaneum area.  EPA received supplemental information from MDNR on
March 19, 2008.

Since the SIP call was issued in April 2006, Herculaneum air monitors
have recorded additional exceedances of the quarterly lead NAAQS.  In
total, since the third quarter of 2002, exceedances have occurred in
the: first, second, third quarters of 2005; first, third, fourth
quarters of 2006; second and third quarters of 2007; and the first
quarter (January – March) of 2008.  The SIP submittal established
April 7, 2008, as the attainment date and required implementation of all
measures required for attainment by that date.  All monitors were below
the quarterly NAAQS standard in the second quarter (April – June) of
2008.  The highest reading in the second quarter of 2008 was 1.20 µg/m3
at the Broad Street monitor, and the second highest reading was 1.04
µg/m3 at the Main Street/City Hall monitor.

Policy Analyses

2.1	Section 110 Requirements

Section 110(k)(5) of the CAA provides that after EPA makes a finding
that a plan is substantially inadequate, it may establish a reasonable
deadline for correcting the deficiencies, but the date cannot be later
than 18 months after the state is notified of the finding.  On April 14,
2006, EPA published the final rule finding Missouri’s SIP inadequate
to provide for attainment of the lead NAAQS in Herculaneum, Missouri,
and requiring the state to submit a plan revision (71 FR 19432).  The
final rule required Missouri to submit revisions to the SIP no later
than April 7, 2007, and required that the SIP provide for attainment of
the lead NAAQS in the Herculaneum nonattainment area as expeditiously as
practicable, but no later than April 7, 2008.

On May 31, 2007, EPA received the Missouri Department of Natural
Resources’ revised SIP entitled, “2007 Revision of the State
Implementation Plan for the Herculaneum Lead Nonattainment Area.” 
Section 110(a)(2) of the CAA outlines a number of elements such plans
must include.  Specific elements include but are not limited to:
enforceable emission limitations and other control measures, schedules
and timetables for compliance, and air quality modeling as necessary. 
MDNR’s submittal addressed each of these elements.

The MDNR submittal includes a control strategy developed to address
emissions from building fugitives, baghouses, materials handling, and
transportation-related emissions.  In addition to construction projects
and work practice requirements, the control strategy includes hours of
operation and throughput production limits.

As also required by 40 CFR part 51, subpart N and CAA section 172(c)(6),
compliance schedules for implementing the control strategy are included
as part of the SIP.  The Consent Judgment entered into between the State
of Missouri and The Doe Run Resources Company, signed by Judge Troy
Cardona, Div. 6 on May 21, 2007, and submitted to EPA as part of the
official SIP submittal, includes the control measures and a schedule for
compliance.  It requires controls to be in place on or before April 7,
2008, with the sole exception of enforceable measures produced as a
result of the ventilation study (see section 5.2.4 of this document for
further discussion on the ventilation study and EPA’s proposed action
related to the ventilation study).  Upon approval by MDNR, the minimum
fan amperages and other requirements identified in the ventilation study
will become enforceable conditions of the Consent Judgment and/or Work
Practices Manual.  

There are a number of important requirements within this SIP revision
which are either currently unspecified (e.g., the enforceable measures
produced as a result of the ventilation study) or are subject to change
upon approval from MDNR (e.g., the concentrate delivery system –
condition A.1 of the Consent Judgment, and the substitute contingency
measures - condition C.5).  While approving these provisions, EPA notes
that any subsequent requirements developed pursuant to the provisions
that could impact emission levels or required monitoring - including
enforceable conditions resulting from the ventilation study - must be
submitted to EPA for approval as revisions to the SIP, thereby also
triggering opportunity for public input through the state and federal
processes.

In support of the revision to the Doe Run-Herculaneum lead SIP, a
dispersion and receptor modeling methodology was developed to predict
ambient lead concentrations as required by section 110(a)(2)(K) of the
CAA and 40 CFR 51.112.  MDNR determined that the Industrial Complex
Short-Term air dispersion model was the appropriate model to use for
this SIP.  (See Section 4.1.1, Model Selection.)

Control strategy modeling submitted by the state shows the area
attaining the standard with all control measures in place, with a
maximum predicted concentration of 1.492 µg/m3.  This concentration
includes the background value for Herculaneum, Missouri, which was
estimated at 0.063 µg/m3 (for more on the background value, see section
4.2.4 Modeled Background).  Further discussion of the individual
controls may be found in section 5.0 of this document. 

As provided in CAA section 110(k)(1)(B), the SIP was reviewed for
completeness before a review of adequacy was conducted.  A completeness
review determines whether the SIP contains the required technical and
administrative elements, as described in 40 CFR part 51, Appendix V. 
These include administrative elements such as evidence the state has
adopted the SIP and has legal authority to implement and enforce the
plan, certification that public hearings were held and compilation of
public comments and the State’s responses, and technical elements such
as submittal of modeling information required to support the state’s
demonstration, evidence that the plan contains emission limitations,
work practice standards, and recordkeeping/reporting requirements as
necessary, compliance/enforcement strategies, and submittal of the
state’s demonstration that the NAAQS is protected if the plan is
approved and implemented.  The SIP revision was deemed complete by
operation of law on December 1, 2007.  

2.2	Section 172 Requirements

As required under section 172(a)(2) of the CAA Amendments, the
attainment date for an area designated nonattainment with respect to a
national primary ambient air quality standard shall be the date by which
attainment can be achieved as expeditiously as practicable, but no later
than 5 years from the date the area was designated as nonattainment
under section 107(d).  As previously described, section 110(k)(5) also
applies in this case, requiring plan submittal no later than 18 months
following notification of the finding.  Section 172(d) further specifies
that EPA may adjust SIP deadlines if the plan revision is required in
response to a finding of plan inadequacy.  On April 14, 2006, EPA found
Missouri’s lead SIP inadequate, and required Missouri to submit
revisions to the SIP no later than April 7, 2007, and required that the
SIP provide for attainment of the lead NAAQS in the Herculaneum
nonattainment area as expeditiously as practicable, but no later than
April 7, 2008.

Section 172(d) of the CAA also provides that a plan revision required by
a SIP call under section 110(k)(5) must correct the deficiencies
specified by EPA, and must meet all other applicable plan requirements
under section 110 and Part D of Title I of the CAA.  As required by
172(c) of the CAA, MDNR submitted several specific plan elements to EPA
in order to correct the inadequacy of the SIP.  These specific elements
included:  (1) a revised emissions inventory, (2) a modeling
demonstration showing what reductions will be needed to bring the area
back into attainment of the lead NAAQS, (3) adoption of measures to
achieve reductions determined necessary by the attainment demonstration,
with enforceable schedules for implementing the measures as
expeditiously as practicable, and (4) contingency measures.  

The SIP also addresses Reasonably Available Control
Technology/Reasonably Available Control Measures (RACT/RACM)
determination and Reasonable Further Progress (RFP) determination. 
Section 172(c)(1) of the CAA for nonattainment areas requires the
implementation of all RACM, including emissions reduction through the
adoption of RACT, as expeditiously as practicable.  In the April 14,
2006, SIP call, EPA did not list a new RACT analysis as a required
element of the SIP submittal.  Even though not required by the 2006 SIP
call, a RACT/RACM analysis is still included with the 2007 SIP
submittal.  No additional RACT measures were identified that would
expedite attainment or reasonable further progress, and the plant has
not changed significantly from when the previous RACT/RACM evaluation
was completed.  Some previously implemented RACT/RACM measures, i.e.,
types of controls, were strengthened through incorporation of more
detailed, enforceable work practices in the Work Practices Manual.  See
section 5.3 in this TSD for further RACT/RACM discussion. 

	Section 172(c)(2) of the CAA requires SIPs to provide for Reasonable
Further Progress (RFP) as defined in section 171(1) of the CAA.  MDNR
has demonstrated RFP as required under section 172(c)(2) of the CAA. 
Doe Run is subject to a compliance schedule for implementing: (1)
installation of emission control equipment, (2) enclosure and
ventilation projects to reduce lead emissions, (3) process throughput
restrictions and hours of operation limitations, and (4) work practice
standards.  These are but a few of the SIP controls that are enforceable
through the Consent Judgment and/or the Work Practices Manual.  For
further RFP discussion, see section 5.4 of this TSD.

As required by section 172(c)(9) of the CAA, contingency measures are
also submitted as part of the SIP.  Should the air quality data for the
second calendar quarter of 2008, or any quarter thereafter, exceed the
1.5 µg/m3 NAAQS, the first two contingency measures will be completed
within 6 months of notification of exceedance.  Additional contingency
measures will be required should there be exceedances following
implementation of the first set of contingency measures.  (See 6.0
Contingency Measures.) 

Technical Background

Nonattainment Boundaries

The nonattainment designation was codified in 40 CFR part 81 and became
effective on January 6, 1992 (56 FR 56694).  The nonattainment
designation applies to that part of Jefferson County, Missouri, which is
within the city limits of the town of Herculaneum.  (See 

Attachment 1: Map of Herculaneum City Boundaries.)

Location and Topography

	The Doe Run facility is located in Herculaneum, Missouri, on the west
bank of the Mississippi River approximately 30 miles south of the city
of St. Louis.  Doe Run is bounded on the east by the Mississippi River
and by Joachim Creek to the south.  The Doe Run facility is bounded on
the west and north sides by Herculaneum residential areas.

	Overall, the topography in the vicinity of the smelter can be described
as relatively level in the immediate area of the facility and in the
agricultural land in Illinois east of the facility.  Beyond the
immediate facility area, higher terrain is found in the river bluffs
that extend from the south through the west to the north-northeast. 
Terrain to the west of the facility rises steeply.  The site location
can be described in modeling terms as rural, not urban, because of the
sparsely-inhabited plains east of the facility and lack of industrial
facilities to the west.  

Meteorology

3.3.1	Wind

	Based upon meteorological data collected at the Doe Run meteorological
monitoring station immediately to the east of the process buildings, the
predominant wind direction for the April 1, 1997, to March 31, 1999,
period was from the west-northwest.  However, the wind directions vary
considerably from quarter to quarter and from month to month.  Wind
speeds are typically light, especially during the summer months when
they average about 1.70 meters per second.  Topographical effects, e.g.,
the steeply rising terrain to the west of the facility, the more
gradually rising terrain to the north and north-northeast, and the creek
and river valleys south and east of the lead facility, influence the
wind speed, wind direction, and temperature.

Precipitation and Temperature

Except for the local effects on wind speed, wind directions, and
temperature, climatological conditions at Herculaneum are typical of
eastern Missouri.  Annual precipitation is mostly in the form of rain
and the precipitation averages about 39 inches.  Monthly average
temperatures range from a low of approximately 30 degrees Fahrenheit in
January to a high around 79 degrees Fahrenheit in July.  The lowest
minimum temperatures, approximately -18 degrees Fahrenheit, occur in
January, while the highest maximum temperatures of about 107 degrees
Fahrenheit occur in July and August.  

Summary of Air Quality Data

	EPA’s Air Quality System (AQS) database contains quality assured data
for Herculaneum area monitors starting in 1982.  Over the years, the
monitoring network has altered to reflect the changing ambient air
concentration profile in Herculaneum.  As the smelter has improved
controls and reduced emissions, MDNR and Doe Run adapted the monitoring
network.  Table 1: Herculaneum Lead Monitors shows all AQS historical
and current Herculaneum lead monitors.   

Table 1.  Herculaneum Lead Monitors















Other Names	AQS Monitor Name*	Reporting Agency	Site ID	POC	Start date
End date**	Most Recent Sampling Frequency

City Hall	Main St.	MDNR	29-099-0004	1	2004	running	every day

City Hall	Main St.	DR	29-099-0004	2	2004	running	every day

High School	Dunklin	MDNR	29-099-0005	1	1982	running	1 in 3

High School	Dunklin	MDNR	29-099-0005	2	2006	running	1 in 3

High School	Dunklin	DR	29-099-0005	3	1982	running	1 in 3

High School	Dunklin	DR	29-099-0005	4	2006	3/31/2008	1 in 3

Golf Course	DRH Site # 2, Golf Course	DR	29-099-0008	1	1982	10/2005	1 in
6

Dow Chemical	Pevely	DR	29-099-0009	1	1982	10/2005	1 in 6

Dow Chemical	Pevely	MDNR	29-099-0009	2	2002	6/2003	1 in 3

-	Ursuline	DR	29-099-0010	1	1982	10/2005	1 in 6

-	Ursuline	MDNR	29-099-0010	2	2001	6/2003	1 in 3

Rutz's Home	Bluff	DR	29-099-0011	1	1982	running	1 in 3

-	Bluff	MDNR	29-099-0011	2	2001	running	1 in 3

Division Manager's Home	Sherman	DR	29-099-0013	1	1982	running	1 in 6

Division Manager's Home	Sherman	MDNR	29-099-0013	2	2001	11/2003	1 in 3

-	Broad St.	DR	29-099-0015	1	1992	running	every day

-	Broad St.	MDNR	29-099-0015	2	2001	running	every day

-	Thurwell	DR	29-099-0016	1	2004	8/2005	1 in 3

-	Mott St.	DR	29-099-0020	1	4/2008	running	1 in 3

-	Mott St.	DR	29-099-0020	2	4/2008	running	1 in 3

-	Circle St.	DR	29-099-0021	1	4/2008	running	every day

-	Circle St.	MDNR	29-099-0021	2	2006	running	every day

-	South Cross	DR	29-099-0022	1	4/2008	running	every day

-	North Cross	DR	29-099-0023	1	4/2008	running	every day

-	Church St.	DR	29-099-0024	1	4/2008	running	1 in 3

-	Church St.	DR	29-099-0024	2	4/2008	running	1 in 3









	Bold indicates currently operating monitors, as of 8/2008.  Italics
indicates monitors no longer operating.

*AQS stands for "Air Quality System," the official EPA database for
quality assured monitoring data.

**"running" indicates active monitors for which data was entered into
AQS as of August 2008.

	

In 1982, the network consisted of seven monitors at six locations
dispersed over several miles.  (See Attachment 2: Map of Herculaneum
Monitors, 1982-2005.)  These monitors were sited using dispersion
modeling in accordance with the U.S. EPA Guidelines for Lead Monitoring,
January 1980.  As the graphs below indicate, by mid-2002, with the
implementation of the 2002 SIP, air concentrations substantially
decreased.  Monitors further from the smelter consistently reported
quarterly values well below the 1.5 µg/m3 standard.  As a result,
between 2003 and 2005 monitoring was discontinued at the Golf Course,
Pevely, Ursuline, and Thurwell monitors, as well as the collocated
Sherman monitor.  The last exceedance of the standard at one of these
monitors was in 1986, and the highest quarterly reading since the 2002
SIP controls were implemented was 0.193 µg/m3.  The chart titled
“Ursuline Monitor (29-099-0010, POC 1): Quarterly Averages for
Complete Life of Monitor, 1982 – 2005” shows the quarterly averages
over time for one such discontinued monitor.  The charts, “Ursuline
(29-099-0010, POC 1 & 2), Sherman (POC 2), Thurwell (POC 1): Quarterly
Averages for Last 5 Years of Monitoring, 2000 – 2005” and “Golf
Course (29-099-0008, POC 1), Pevely (29-009-0009, POC 1 & 2): Quarterly
Average for Last 5 Years of Monitor, 2000 – 2005” show the quarterly
averages over the last five years the monitors were operated.  Gaps in
data indicate quarters in which data did not satisfy summary criteria. 
An example of a situation that would result in the data not satisfying
summary criteria is if power at a monitor was lost for a number of days
during the quarter.  This could result in data loss and too few days
with data within the quarter to produce a statistically significant
quarterly average.

Many of the air lead monitoring stations located in Herculaneum have
recorded quarterly averages below the NAAQS for several consecutive
quarters.  The graph below shows quarterly averages for Dunklin and
Broad Street monitors, from 1982 through 2007.  Note the decreased
concentrations following implementation of the 2002 SIP.  There were ten
quarters of attainment, starting the third quarter of 2002, until an
exceedance in the first quarter of 2005.

The state and Doe Run installed new monitors after 2000 to better
characterize areas showing the highest concentrations following
implementation of the 2002 SIP and locations about which citizens voiced
concerns.  Notably, an additional two monitors were installed at the
Dunklin High School location; the state installed monitors at Broad
Street and Bluff, where Doe Run monitors were already located, and Doe
Run and the state collocated a new set of monitors at Main Street/City
Hall.  As shown on the map (Attachment 2: Map of Herculaneum Monitors,
1982 - 2005), Broad Street and Main Street/City Hall were the monitors
closest to the smelter.  Since 2002, only Broad Street and Main
Street/City Hall monitoring locations have recorded exceedances of the
1.5 µg/m3 lead NAAQS.  

	Table 2 provides the historical quarterly averages monitored at the
Dunklin High School monitors.  Tables 3 and 4 provide the historical
quarterly averages monitored at the Broad Street monitors and the Main
Street/City Hall monitors, respectively.

Table 2.  Dunklin High School Monitor (29-099-0005): Ambient Air Lead
Concentrations (µg/m3)

(MDNR monitor poc 1/ Doe Run monitor poc 3)















Year	1st Qtr	2nd Qtr	3rd Qtr	4th qtr

 	(MDNR	/	Doe Run)	(MDNR	/	Doe Run)	(MDNR	/	Doe Run)	(MDNR	/	Doe Run)

1982	-	/	-	-	/	-	0.81	/	1.18	na	/	3.81

1983	-	/	1.12	-	/	1.87	2.42	/	1.62	1.55	/	4.33

1984	1.92	/	0.99	1.56	/	0.10	1.01	/	0.59	3.28	/	1.28

1985	1.02	/	0.48	2.04	/	1.91	1.51	/	4.31	0.81	/	1.76

1986	2.46	/	2.41	1.82	/	1.49	1.36	/	1.68	1.23	/	3.46

1987	1.45	/	1.14	1.39	/	2.68	na	/	1.88	1.56	/	1.85

1988	0.87	/	3.67	1.97	/	1.38	1.61	/	1.45	2.20	/	1.54

1989	1.88	/	1.16	na	/	1.57	1.78	/	1.12	2.34	/	1.32

1990	0.72	/	0.91	0.98	/	1.61	1.29	/	1.19	2.18	/	1.86

1991	na	/	1.45	1.78	/	0.71	0.71	/	1.37	0.83	/	1.65

1992	0.47	/	0.65	0.61	/	1.27	1.75	/	1.95	2.09	/	2.54

1993	0.69	/	0.33	1.00	/	1.71	1.37	/	0.61	1.85	/	0.64

1994	2.27	/	0.43	1.57	/	0.68	0.75	/	1.32	1.70	/	1.44

1995	0.72	/	0.58	1.42	/	0.66	1.30	/	1.20	1.53	/	1.75

1996	1.21	/	1.87	1.47	/	1.24	0.92	/	0.61	1.32	/	1.81

1997	1.20	/	0.75	0.73	/	1.28	0.67	/	1.11	1.16	/	1.34

1998	1.43	/	1.13	1.99	/	1.45	1.35	/	0.77	0.77	/	1.42

1999	4.50	/	1.28	1.87	/	1.43	2.18	/	0.40	1.25	/	0.96

2000	0.37	/	1.19	0.73	/	1.35	0.38	/	1.23	0.14	/	0.90

2001	0.28	/	3.78	0.24	/	1.80	0.19	/	1.89	0.39	/	1.28

2002	0.38	/	0.39	0.46	/	0.61	0.47	/	0.38	0.25	/	0.16

2003	0.19	/	0.37	0.34	/	0.25	0.26	/	0.25	0.37	/	0.54

2004	0.24	/	0.51	0.36	/	0.68	0.32	/	0.47	0.37	/	0.32

2005	0.13	/	0.21	0.26	/	0.30	0.50	/	0.29	0.39	/	0.27

2006	0.23	/	0.29	0.33	/	0.36	0.29	/	0.34	0.29	/	0.36

2007	0.11	/	0.18	0.26	/	0.29	0.52	/	0.53	0.42	/	0.42

2008	0.42	/	0.45	0.33	/	0.36





















“-” indicates monitor was not yet installed and operating.



	"na" indicates the mean did not satisfy summary criteria.



	

Table 3.  Broad Street Monitor (29-099-0015): Ambient Air Lead
Concentrations (µg/m3)

(MDNR monitor poc 2/ Doe Run monitor poc 1)















Year	1st Qtr	2nd Qtr	3rd Qtr	4th qtr

 	(MDNR	/	Doe Run)	(MDNR	/	Doe Run)	(MDNR	/	Doe Run)	(MDNR	/	Doe Run)

1992	-	/	-	-	/	-	-	/	-	-	/	5.530

1993	-	/	3.720	-	/	5.450	-	/	2.060	-	/	2.310

1994	-	/	3.450	-	/	3.650	-	/	3.910	-	/	3.120

1995	-	/	6.540	-	/	2.450	-	/	4.130	-	/	6.310

1996	-	/	2.300	-	/	5.740	-	/	3.990	-	/	1.600

1997	-	/	3.970	-	/	6.810	-	/	1.650	-	/	8.530

1998	-	/	11.590	-	/	4.140	-	/	3.950	-	/	5.400

1999	-	/	6.750	-	/	4.070	-	/	2.940	-	/	4.190

2000	-	/	4.310	-	/	4.930	-	/	6.860	-	/	2.760

2001	-	/	2.950	-	/	6.780	-	/	9.130	1.760	/	1.800

2002	1.130	/	1.109	2.240	/	1.848	1.000	/	0.924	1.010	/	1.019

2003	1.464	/	1.491	1.461	/	1.349	1.352	/	1.363	0.950	/	1.025

2004	1.200	/	1.390	1.380	/	1.480	1.380	/	1.460	1.270	/	1.360

2005	1.880	/	1.930	1.460	/	1.610	1.600	/	1.730	1.080	/	1.220

2006	1.670	/	1.820	1.310	/	1.500	1.490	/	1.620	1.340	/	1.310

2007	1.210	/	1.370	1.500	/	1.650	1.630	/	1.740	1.260	/	1.310

2008	2.73	/	2.89	1.02	/	1.20





















“-” indicates monitor was not yet installed and operating.





	

Table 4.  Main Street/City Hall Monitor (29-099-0004): Ambient Air Lead
Concentrations (µg/m3)

(MDNR monitor poc 1/ Doe Run monitor poc 2)















Year	1st Qtr	2nd Qtr	3rd Qtr	4th qtr

 	(MDNR	/	Doe Run)	(MDNR	/	Doe Run)	(MDNR	/	Doe Run)	(MDNR	/	Doe Run)

2004	0.974	/	0.874	1.238	/	1.259	1.416	/	1.452	1.475	/	1.383

2005	0.880	/	0.820	1.090	/	0.950	0.890	/	0.670	1.090	/	1.120

2006	1.450	/	1.380	0.990	/	0.990	1.400	/	1.370	1.930	/	1.830

2007	1.110	/	1.090	1.140	/	1.110	1.180	/	1.060	1.460	/	1.360

2008	2.01	/	2.00	1.02	/	1.04









	In 2006, the state installed a monitor at Circle Street in anticipation
of a change to Doe Run’s fence line.  Seven new monitors at five
locations (collocated monitors at Mott, one monitor at South Cross, one
monitor at North Cross, collocated monitors at Church, and a collocated
monitor at Circle Street - see Attachment 13) were also installed to
monitor the ambient air concentrations along the new fence line, and
began reporting data to EPA’s Air Quality System (AQS) database in
April 2008.  Doe Run completed installation of a new fence line to the
west of the facility (see Attachment 13) on August 25, 2008.  With the
MDNR-approved completion of the new fence line, the collocated Broad
Street monitors are now inside of the fence line, and are therefore not
located in ambient air.  The Consent Judgment requires Doe Run to
continue running its Broad Street monitor, sampling once every six days
or more frequently.  Information from the Broad Street monitors will
continue to be entered into AQS and used for informational purposes, but
will not be used to determine attainment with the lead NAAQS as the
Broad Street monitors are no longer located in ambient air.

In summary, there are currently nineteen ambient air lead monitors
located in the vicinity of the Doe Run-Herculaneum facility that are
reporting to the EPA AQS database.  There are ten monitoring locations,
several with collocated Doe Run and MDNR monitors:  Main Street/City
Hall (29-099-0004: 1 MDNR monitor, 1 Doe Run monitor), Dunklin High
School (29-099-0005: 

2 MDNR monitors, 2 Doe Run monitors), Bluff (29-099-0011: 1 MDNR
monitor, 1 Doe Run monitor), Sherman (29-099-0013: 1 Doe Run monitor),
Broad Street (29-099-0015: 1 MDNR monitor, 1 Doe Run monitor), Mott
Street (29-099-0020: 2 Doe Run monitors), Circle Street (29-099-0021: 1
MDNR monitor, 1 Doe Run monitor), South Cross (29-099-0022: 1 Doe Run
monitor), North Cross (29-099-0023: 1 Doe Run monitor), and Church
Street (29-099-0024: 2 Doe Run monitors).  The monitors are operated in
accordance with the requirements and specifications of 40 CFR part 60,
appendix B.  Laboratory analysis of filters is conducted in accordance
with 40 CFR part 50, appendix G.  The monitors are run for a 24-hour
period on a frequency of at least once every six days.  As of August
2008, daily monitoring occurs at the Broad Street, Main Street/City
Hall, Circle Street, South Cross, and North Cross monitors, which are
the locations closest to the facility (see Attachment 13).

	Table 5 contains the quarterly averages monitored at all monitors
operating in the first quarter (January – March) and second quarter
(April – June) of 2008.  As previously noted, several monitors began
reporting data to AQS in the second quarter of 2008.  The first quarter
of 2008 saw abnormally high values.  Doe Run was completing construction
projects associated with SIP-required facility modifications, and
attributes the level of emissions to these activities.  On April 7,
2008, at the beginning of the second quarter, all SIP-required controls
were in place (with the sole exception of the ventilation study
controls, as discussed in section 5.2.4 of this document).  The highest
value in the second quarter of 2008 was 1.20 µg/m3.

Table 5.  All Herculaneum Monitors: Ambient Air Lead Concentrations,
Quarterly Average, 1st and 2nd Quarters of 2008.

AQS Monitor Name	Reporting Agency	Site ID	POC	1st Qtr 2008* (µg/m3)	2nd
Qtr 2008** (µg/m3)

	Main St.	MDNR	29-099-0004	1	2.01	1.02

	Main St.	DR	29-099-0004	2	2.00	1.04

	Dunklin	MDNR	29-099-0005	1	0.42	0.33

	Dunklin	MDNR	29-099-0005	2	0.41	0.32

	Dunklin	DR	29-099-0005	3	0.45	0.36

	Dunklin	DR	29-099-0005	4	0.48	-***

	Bluff	DR	29-099-0011	1	0.75	0.28

	Bluff	MDNR	29-099-0011	2	0.63	0.25

	Sherman	DR	29-099-0013	1	0.41	0.24

	Broad St.	DR	29-099-0015	1	2.89	1.20

	Broad St.	MDNR	29-099-0015	2	2.73	1.02

	Mott St.	DR	29-099-0020	1	-	1.01

	Mott St.	DR	29-099-0020	2	-	0.84

	Circle St.	DR	29-099-0021	1	-	0.85

	Circle St.	MDNR	29-099-0021	2	1.14	0.72

	South Cross	DR	29-099-0022	1	-	0.52

	North Cross	DR	29-099-0023	1	-	0.45

	Church St.	DR	29-099-0024	1	-	0.36

	Church St.	DR	29-099-0024	2	-	0.35

	*Doe Run states that the abnormally high 1st quarter 2008 values are
attributable to emissions from

	construction activity associated with SIP-required facility
modifications.

	**All SIP-required controls, except those associated with the
ventilation study (see 5.2.4 of this document), were 

completed by April 7, 2008.

***With MDNR’s approval, Doe Run moved their Dunklin POC 4 monitor to
Mott Street, where it now operates as one of the collocated Mott
samplers.



Source Characterization

The Doe Run-Herculaneum facility was opened in 1892 and is the only
primary lead smelter currently operating in the United States.  The
annual total production capacity of the facility is approximately
250,000 tons of refined lead.

	

	The primary lead smelting process begins with lead concentrate.  Doe
Run-owned mining and milling operations located in south-eastern
Missouri are the primary source of Doe Run-Herculaneum’s lead ore and
lead concentrate.  Lead ore, typically 6 percent to 8 percent lead, is
mined from underground ore deposits.  The ore is crushed and then
processed into lead concentrate at the mills.  Lead concentrate contains
approximately 75 percent lead by weight.  Lead concentrate was
previously transported from the mines/mills to the Herculaneum smelter
by rail, but since 2002 has been transported exclusively by truck to
Herculaneum.  Once delivered to the Herculaneum primary lead smelter,
the process of smelting the lead concentrate into high purity lead can
be divided into three main steps: sintering, reducing (smelting), and
refining.

	Once delivered to Herculaneum, the concentrate is first processed
through the sinter plant.  The concentrate is mixed and crushed with
other feedstock materials such as silica, iron ore, and limestone
fluxes.  Recycled process material such as returned sinter, blast
furnace slag, and baghouse fume may also be added to this mixture to
produce the sinter feed.  A thin layer of sinter feed enters the sinter
machine and is ignited by a series of natural gas burners.  A main
sinter feed layer is then laid on top of this ignition layer.  This
layered sinter bed enters the updraft portion of the sinter machine,
where air is drawn across the sinter bed from the bottom to the top,
driving the thermal reaction.  The lead sulfide contained in the feed is
oxidized, producing lead oxide and releasing sulfur dioxide.  Off-gasses
from the sintering process are sent to a baghouse which removes
particulate matter.  The off-gasses continue on to the acid plant where
sulfur dioxide is recovered as sulfuric acid.  The sinter machine
produces a continuous feed of sinter cake (also called sinter roast)
which is crushed and sorted by size.  The larger pieces are transported
to the blast furnace or to temporary storage, while the undersized
pieces return to the mix room to await reprocessing through the sinter
machine.

	Smelting takes place in Doe Run-Herculaneum’s two blast furnaces. 
Sinter cake is mixed with coke and other feed materials and transferred
to the top of a furnace.  Air feeds through the bottom of the furnace,
resulting in coke combustion.  The coke combustion heats the sinter cake
to approximately 3,000 degrees Fahrenheit and produces carbon monoxide. 
The carbon monoxide reacts with lead and other metal oxides to produce
molten lead, waste slag, and carbon dioxide.  The lead bullion settles
to the bottom of the furnace, where it is tapped into holding pots and
transferred to the drossing area for further refining.  The slag (a
sand-like byproduct with small amounts of lead, copper, zinc, and other
materials) floats to the top of the furnace, is tapped off and either
recycled into the sinter feed or transported to the slag storage area at
the south end of the facility.  

	Impurities are further separated and removed from the lead in the
dross/refinery departments.  The lead bullion from the blast furnace is
first transferred to one of four large drossing kettles where it is
allowed to cool.  As the bullion cools, copper, nickel, and other
impurities are skimmed from the surface layer, known as the “dross.”
 Next, the decopperized lead is transferred to a series of natural
gas-heated refining kettles where additional impurities are removed. 
Zinc is added to the lead to facilitate the removal of silver.  The
zinc-silver dross that forms at the surface of the kettle is removed and
then further processed in order to recover the silver.  Excess zinc is
removed by vacuum distillation and chemical conversion.  The resulting
lead is more than 99.999 percent pure and is cast into 60-pound and
100-pound pigs, as well as 1-ton ingots.  Precise amounts of other
metals may be added to the molten lead in order to produce lead alloys
for specific industrial uses.

	Process flow diagrams, depicting potential points of lead emissions,
are included as Attachments 4 and 5.

	More detailed information on the primary lead smelting process can be
found in AP-42, Fifth Edition, Section 12.6: Primary Lead Smelting.

	Development of Dispersion Model Inputs

4.1	Model Input Development

4.1.1 	Model Selection

In support of the revision to the Doe Run-Herculaneum lead SIP, a
dispersion modeling methodology was developed to predict ambient lead
concentrations as required by section 110(a)(2)(K) of the CAA and 40 CFR
51.112.  

EPA Region 7 initially considered modeling using the CALPUFF air
dispersion model to estimate lead concentrations in the Herculaneum
area.  Situations where there are complex winds, or the winds are light
or calm, are handled well by a non-steady state puff model such as
CALPUFF.  However, MDNR and EPA determined that recent available
meteorological data were not of sufficient quality to use in the CALPUFF
model to ensure minimum performance criteria required to perform a
performance evaluation.  Significant sections of data were missing due
to instrument error and were thus unusable in CALPUFF.

	EPA’s AERMOD model was also considered.  It has been a guideline
model since December 2005 and contains more recently updated algorithms
than the ISC model.  As with the CALPUFF model, the recent available
meteorological data were not of sufficient quality to ensure minimum
performance criteria required to perform a performance evaluation in
AERMOD due to missing or erroneous data.  

MDNR selected the Industrial Source Complex Short-Term (ISC3P) model. 
The principal difference between ISC3P, version 04269 and the ISCST
model (used in modeling for the 2002 SIP) is the downwash algorithm. 
The ISC3P model includes a prime downwash algorithm that is similar to
the downwash algorithm in the AERMOD and CALPUFF models, while ISCST
includes an older downwash algorithm.  

ISC3P can be used to assess concentrations from several types of sources
associated with industrial source complexes, can account for building
downwash, urban or rural dispersion coefficients, flat or elevated
terrain, and averaging periods from one hour to one year.  ISC3P
requires fewer meteorological parameters than AERMOD and CALPUFF.  EPA
approved the use of ISC3P model for the lead SIP control strategy. 
Refer to February 12, 2007, memorandum from Jeffry D. Bennett to John
Rustige, titled, “Doe Run – Herculaneum State Implementation Plan
(SIP) Dispersion Modeling Review” for additional information on the
modeling methodology.

4.1.2 	Study Period

	Due to the short time period allotted to develop the SIP revision and
the substantial technical support already in existence, MDNR relied on
existing information as the starting point for the SIP development.  As
described in further detail in sections below, model refinement
completed using the actual value modeling was conducted using 2005
inputs.  Once the model performance evaluation was complete, the
attainment demonstration modeling was conducted using meteorological
data from April 1997 - March 1999 and January – March 2005.  These
quarters of meteorological data represented recent, complete
meteorological data, as well as a larger block of time over which a
range of meteorological conditions occurred.

Meteorological Data

	The meteorological data set used in the actual value/model performance
modeling was developed from data collected in 2005.  The 2005 data used
on-site data from the river site located east of the plant and mixing
height data from the Lincoln, Illinois (WBAN 04833) radiosonde site. 
Solar Radiation/Delta-T (SRDT) data were available for 112 days during
2005.  For 124 other days in 2005, either the solar radiation or
temperature difference data were missing for at least part of the day. 
The sigma-A method was utilized to calculate the stability class for
these days.  This method is a turbulence-based method which uses the
standard deviation of the wind direction in combination with the scalar
mean wind speed (reference: Meteorological Monitoring Guidance for
Regulatory Modeling Applications, pages 6-18, EPA-454/R-99-005, February
2000).  Finally, one day (March 22, 2005) was missing data for hours 2
p.m. and 3 p.m.  Examining the hours with available data, wind speeds
decreased from noon through 8 p.m., wind direction varied by only 24.2
degrees between hour 5 and hour 21, and maximum radiation and
temperatures were reached earlier in the day.  Simple interpolation
using the two data points on either side of the gap was used to fill the
two-hour data gap (See Table 6.  Gap Filling Summary for 3/23/2005,
hours 2 p.m. and 3 p.m.).  In total, data from 247 days in 2005 were
used for the actual value modeling.  (See Table 7.  Meteorological Block
Descriptions.)  Additional information on how the meteorological data
were developed may be found in the February 1, 2007, Shell Engineering
document titled, “Design Value Modeling Analysis in Support of the
Revision to the Herculaneum, Missouri Lead SIP.”  

Table 6.  Gap Filling Summary for 3/23/2005, hours 2 p.m. and 3 p.m.



Time	Wind Speed (mph)	Wind Direction (degrees)	Sigma	Temp 2 meters
(degrees F)	Solar Radiation (cal_cm-2_min-1)	Net Radiation (Wm-2)	BP
(Inches Hg)	Rain (Inches)

midnight	6.82	5.03	16.27	41.51	0	-2.7	29.14	0.04

1am	7.07	14.65	14.12	41.66	0	-2.55	29.13	0.02

2am	6.72	353.5	15.46	41.56	0	-2.41	29.11	0.05

3am	5.53	337.3	16.44	41.53	0	-2.6	29.11	0.06

4am	5.63	329.9	18.93	41.47	0	-2.97	29.12	0.06

5am	5.69	323.3	18.83	41.18	0	-3.2	29.14	0.13

6am	6.62	318.4	19.47	40.35	0	-3.55	29.16	0.03

7am	7.24	317.4	19.03	39.83	0.004	-1.51	29.19	0.01

8am	6.77	323.5	19.8	39.49	0.02	12.36	29.22	0

9am	6.74	323.8	19.97	39.81	0.05	28.36	29.24	0

10am	7.68	330	20.79	40.23	0.12	63.38	29.26	0

11am	8.06	313.7	20.41	40.42	0.14	74	29.28	0

Noon	8.76	306.4	20	40.37	0.11	61.72	29.31	0

1pm	7.62	308.1	22.47	40.47	0.11	61.7	29.32	0

2pm	7.27	311.5	22.5	40.42	0.09	50.06	29.33	0

3pm	6.92	315.0	22.53	40.37	0.07	38.42	29.34	0

4pm	6.57	318.4	22.56	40.32	0.05	26.78	29.35	0

5pm	6.23	330.6	18.96	40.3	0.02	9	29.36	0

6pm	5.03	318.7	19.92	40.35	0.007	-0.13	29.39	0

7pm	4.66	313.1	17.19	40.19	0	-4.18	29.4	0

8pm	3.74	312.3	20.87	39.65	0	-4.47	29.42	0

9pm	3.84	288.4	17.11	39.32	0	-3.68	29.44	0.01

10pm	3.02	291.4	22.38	39.14	0	-3.2	29.45	0.01

11pm	3.78	281.9	19.33	39.01	0	-3.37	29.45	0



Table 7.  Meteorological Block Descriptions







	Block	Start	End	Days	Comment

1	1/1/2005	1/9/2005	9	No data missing, use SRDT

2	1/10/2005	3/22/2005	72	SRDT missing, use sigma-A method to calculate
stability

3	3/23/2005	3/23/2005	1	Primary parameters missing for 2 hours, filled
by interpolation.  SRDT missing, use sigma-A method to calculate
stability

4	3/24/2005	5/5/2005	43	SRDT missing, use sigma-A method to calculate
stability

5	5/6/2005	5/30/2005	25	No data missing, use SRDT

6	5/31/2005	6/22/2005	23	Primary parameters missing, days not modeled

7	6/23/2005	8/21/2005	60	No data missing, use SRDT

8	8/22/2005	8/26/2005	5	Primary parameters missing, days not modeled

9	8/27/2005	9/5/2005	10	No data missing, use SRDT

10	9/6/2005	10/5/2005	30	Primary parameters missing, days not modeled

11	10/6/2005	11/1/2005	27	No data missing, use SRDT

12	11/2/2005	12/31/2005	60	Primary parameters missing, days not modeled



The design value and control strategy/attainment demonstration modeling
both used nine quarters of data, from April 1997 - March 1999 and
January – March 2005.  The January – March 2005 data were complete
except for the exceptions mentioned in the preceding paragraph.  The
April 1997 – March 1999 data were approved under the previous SIP
attainment demonstration.  Concentrations modeled over these nine
quarters of meteorological data are therefore representative of an
assortment of meteorological conditions, and use of these quality
assured data provides confidence that this SIP control strategy was
evaluated over a variety of meteorological conditions.

Emission Inventory

A revised emission inventory was developed for this SIP revision.  Shell
Engineering’s February 1, 2007, document titled, “2005 Hourly Lead
Emission Inventory for the Doe Run Company’s Herculaneum, Missouri
Smelter,” discusses the development of the individual emission point
characteristics.  In general, 2005 hourly lead emissions were based upon
facility daily production records.  Many of the processes and sources of
emissions had not been altered since the previous SIP and associated
emission rates were assumed to be unchanged.  To clarify where changes
were made and where source characteristics remained unaltered, Doe
Run’s contractor, Shell Engineering, developed a chart (Attachment 3. 
Source Chart, 2002 SIP versus 2007 SIP) outlining source changes from
the 2002 SIP versus the 2007 revised SIP.  Rates were estimated using
equations developed from source testing at the facility or from
published emission factors.  In some instances, the emission equations
include meteorological parameters to account for wind-driven emissions.

Actual Value, Design Value, and Control Strategy Dispersion Modeling

Actual Value Dispersion Modeling & Emissions Reconciliation

Actual value dispersion modeling was conducted to determine whether the
model was performing adequately to pursue attainment demonstration
modeling.  This was determined through three comparisons: (1)
determining the model’s ability to replicate actual monitored daily
(24 hour, midnight to midnight) lead concentrations, (2) comparing the
actual value modeling results with filter analysis results from specific
days in 2005, and (3) determining the model’s ability to replicate
averaged lead concentration values when averaged over a longer period of
time.  As described in the previous section 4.1.3 Meteorological Data,
the meteorological data set used in the actual value/model performance
modeling was developed from data collected in 2005.  Actual stack
heights, flagpole heights at monitor/receptor locations, and the 2005
hour-by-hour inventory were used.

	The first comparison, evaluating the model output versus the monitored
values on a day-to-day basis, was completed for Broad St., Main
Street/City Hall, Bluff, and Dunklin High School monitor sites. 
Overall, the model predictions matched general increases and decreases
in daily values.  The precise predicted daily concentrations varied from
the measured concentrations.  This was attributed to uncertainties in
the meteorological measurements, model algorithms, and the emission
inventory.  Charts developed by MDNR in support of this analysis are
included as Attachment 6.	

The second comparison looked at the filter analysis versus the model. 
The actual value modeling was completed for 247 days in 2005, as
previously discussed.  Of these days, filter analysis was completed on
12 days worth of filters from the Broad St. monitor as well as 11 days
worth of filters from the Main Street/City Hall monitor.  These filter
results may be found in Attachments 7a and 7b, and filter analysis in
Attachments 8a and 8b.  By combining fingerprint data from the previous
SIP with updated source profiles, the filters were analyzed for the
percent contribution from several facility source categories.  This
percent source category contribution profile was compared against the
source category percentage contribution profile indicated by the
modeling.  

A number of corrections were made by the state modelers as a result of
these comparisons.  Corrections included:

Filter analysis pointed to the model under-predicting sinter plant
fugitives.  Review of sinter plant fugitive emission development
indicated possible under estimations.  As a result, sinter plant
fugitives were scaled up within the model to better match filter
results.

Filter analysis pointed to the model over predicting emissions from
Number 7 & 9 Baghouse, and Number 8 Baghouse (dross and refinery
building baghouses).  It was determined that the model had not yet been
revised to include the new baghouse stack parameters.  Number 7 & 9
Baghouse, and Number 8 Baghouse stack parameters were corrected to
actual build height.  Diameter, exit velocity, and locations were
adjusted to represent the combined 7 & 9 stack.  As a result of these
corrections, impacts from the dross and refinery areas within the model
diminished and more closely reflected the filter analysis results.

The model results were compiled after these corrections were made and
compared against the filter analysis.  The filter analysis and model
results showed relatively good agreement.  Table 8 summarizes
calculations contained in Attachments 8a and 8b, comparing the filter
and ISC-Prime source category contribution estimates.  

Table 8.  Filter Culpability vs. ISC-Prime Culpability.









	Filter	Sinter / Trestle	Refinery	Blast Furnace	Roads & Yards	Dross

Main St/City Hall, with 4/6 a	20.3%	17.6%	25.3%	34.7%	2.1%

Main St/City Hall, with 4/6 b	16.0%	17.8%	28.2%	36.0%	2.1%

Broad St., new filters	28.3%	17.3%	13.1%	41.3%	0.0%







	Dispersion Model	Sinter / Trestle	Refinery	Blast Furnace	Roads & Yards
Dross

Main St/City Hall	26.5%	19.2%	13.1%	31.6%	9.2%

Broad St., new filters	22.5%	18.0%	18.4%	32.0%	8.7%















Difference, Model vs Filter Analysis	Sinter / Trestle	Refinery	Blast
Furnace	Roads & Yards	Dross

Main St/City Hall, with 4/6 a	6.2%	1.6%	-12.2%	-3.0%	7.1%

Main St/City Hall, with 4/6 b	10.5%	1.4%	-15.0%	-4.3%	7.1%

Broad St., new filters	-5.8%	0.7%	5.3%	-9.3%	8.7%







*Two results were given for Main Street/City Hall filter analysis for
4/6/2005.  Result a is represented in one result row, Result b is
represented in the next.

**As shown in Attachment 8a, MDNR included an additional 9 days of
filter data from Broad Street.  These days were not analyzed using the
most recent source profiles and are, therefore, not reflected in the
above tables or Attachment 8b.  



The third comparison, looking at the model’s ability to replicate
actual quarterly monitored lead concentrations, also gave favorable
results.  The Sherman monitoring location was added to the four
monitoring locations used in the first comparison.  The model over
predicted concentrations at three of the four monitor sites.  The Broad
Street monitor is the monitor located closest to the smelter and is the
monitor that has registered the majority of the exceedances since 2002. 
At the Broad Street monitor site, the model over predicted at one of the
collocated monitors and under predicted at the other collocated monitor.
 As shown by an examination of actual monitored values, it is not
uncommon for collocated monitors to show different values.  In the 12
quarters from the first quarter of 2005 through the fourth quarter of
2007, the Doe Run monitor at the Broad Street location measured higher
concentrations 11 out of the 12 quarters.  At the Main Street/City Hall
monitor location, the MDNR monitor measured the higher concentration in
11 of the 12 quarters.  The Doe Run monitor at the Dunklin High School
site had higher concentrations in 9 of the quarters and one quarter the
collocated MDNR and Doe Run monitors were almost equal.  Table 9 below
shows monitored versus modeled values.  Breakdowns of quarterly averages
are provided in Attachment 9.

Table 9.  Model Comparison for the Entire Study Period (247 days in
2005).











Monitor	Monitored* (µg/m3)	ISC-Prime Predicted (µg/m3)





Broad Street**	1.361 / 1.215	1.23





Main St/City Hall	0.811 / .917	.859 / .982





Dunklin High School	.265 / .247	0.467





Bluff	.362 / .241	0.484





Sherman	0.178	0.295













* Sites with two values represent collocated samplers: Doe Run's monitor
/ MDNR's monitor.

	** The Broad Street average does not include the 3/21 and 3/22 event
days because the emissions

were not adjusted in the model on these days to reflect these
high-emission events.

	BOLD denotes monitors with an every day sampling schedule.



	

	The three comparisons matched single-day monitored concentrations,
measured source category contributions, and averaged monitored
concentrations against predicted modeled values.  The state concluded
that the model performs extremely well when compared against the
monitoring sites with the most data (highest number of sampling days)
and the highest concentrations.  Generally, the model accurately
predicted lower concentrations at sites further from the facility, and
predicted lower concentrations on days the monitors also showed lower
concentrations.  However, the model tended to have a higher numerical
value associated with these comparatively low concentration days and
sites.  MDNR attributes this difference to the 0.063 µg/m3 background
concentration.  The background concentration was included in all actual
value modeling calculations and was applied uniformly to all days and
all areas modeled.  (For more information on the background
concentration, please see section 4.2.4 of this document.)  There were
therefore zero modeled days with concentrations lower than 0.063 µg/m3.
 However, there are many instances of monitored values lower than 0.063
µg/m3.  For example, in the first quarter of 2005, the Sherman monitor
(which samples once every 6 days) collected 15 samples, of which 11
showed concentrations between 0.01 – 0.04 µg/m3 (the detection limit
is 0.002 µg/m3).  Overall, the state concluded that the model has a
slight overprediction bias, which provides additional confidence that
the model is adequate to develop control strategies and an attainment
demonstration.

Design Value Dispersion Modeling

The receptor network for the design value modeling contained 50m spacing
at the existing fenced boundary and 100m spacing covering most areas
with predicted 1 µg/m3 concentrations.  The 100m spacing covered all
areas where the 1.5 µg/m3 lead NAAQS might be exceeded based on initial
runs completed using a coarse receptor network (250 meter and 500 meter
spacing).  The state did not place receptors on all external roads, but
included receptors on the road closest to the facility (Main Street). 
EPA expects this road would experience the highest total concentration
out of all external roads due to transportation activity and proximity
to the facility and associated emission sources.  The removal of
receptors on some of the more distant roads that were originally
included in the design value modeling does not affect the conclusion
that the NAAQS will be attained.  The background concentration value of
0.063 µg/m3 was added to the predicted air dispersion calculations (see
section 4.2.4 of this document for more on the background
concentration), and the modeling used the nine quarters of data from
April 1997 – March 1999 and January – March 2005.  Good Engineering
Practice (GEP) stack height was used for the main stack, and actual
heights were used for all other stacks (as they were at or below the GEP
stack height).  Also, receptors at the monitor sites were not simulated
at flagpole heights.

Design value modeling was conducted to identify which sources may be
significant contributors in a hypothetical scenario where all processes
operate for as many hours as possible, and throughput is as high as
possible.  The design value modeling was completed for a worst-case
scenario (highest production and maximum operational hours) without
consideration of the 2007 proposed controls and without assuming the
controls resulting from the previous 2002 SIP.  Results from this
worst-case scenario modeling indicated sources or groups of sources that
may significantly contribute to lead concentrations.  Identified source
areas included: south-end storage, all process building fugitives,
Baghouse 7/9 stack, Baghouse 8 stack, unloader area, and in-plant roads.
 The state then took a closer look at the effectiveness of existing
controls and the technological and economic feasibility for additional
controls at these sources. 

Control Strategy Dispersion Modeling

The receptor network for the control strategy modeling matched that of
the design value modeling network.  It contained 50m spacing at the
existing property boundary and 100m spacing covering most areas where
the design modeling predicted 1 µg/m3 concentrations.  The 100m spacing
covered all areas where the 1.5 µg/m3 lead NAAQS might be exceeded
based on initial runs completed using a coarse receptor network (250
meter and 500 meter spacing).  The state did not place receptors on all
external roads, but included receptors on the road closest to the
facility (Main Street).  EPA expects this road would experience the
highest total concentration out of all external roads due to
transportation activity and proximity to the facility and associated
emission sources.  The removal of receptors on some of the more distant
roads that were originally included in the design value modeling does
not affect the conclusion that the NAAQS will be attained.  The
background concentration value of 0.063 µg/m3 was added to the
predicted air dispersion calculations (see section 4.2.4 of this
document for more on the background calculation), and the modeling used
the nine quarters of data from April 1997 – March 1999 and January –
March 2005.  Good Engineering Practice (GEP) stack height was used for
the main stack, and actual heights were used for all other stacks (as
they were at or below the GEP stack height).  Receptors were not
simulated at flagpole heights.

The control strategy modeled was developed by applying capture and/or
control efficiencies to a number of emission points, changing source
characteristics to reflect modified stacks, and considering limitations
on either process throughputs or hours of operation.  The control
strategy modeling incorporates all changes made as a result of the
actual and design value modeling refinement runs.  

One specific set of control efficiencies included in the control
strategy modeling were those attributed to process buildings as a result
of operating conditions required by a ventilation study.  The Consent
Judgment requires Doe Run to conduct a building ventilation study for
the Sinter Building, Blast Furnace Building, and Refinery Building. 
Building openings, ventilation sources with either continuous or varying
rates of operation, and a procedure for measuring inflow into the
buildings must be identified within the study.  The study must also
include enforceable conditions developed to ensure that particles
emitted within the process buildings are being appropriately captured by
the ventilation systems.  Upon MDNR’s approval of the study and its
findings, the enforceable conditions identified in the study will become
part of the Consent Judgment and/or Work Practices Manual.  The control
strategy modeling includes control efficiencies based on the expected
results from the ventilation study, as described further in section
5.2.4 of this document.  For discussion on how EPA proposes to address
the ventilation study, see the Notice of Proposed Rulemaking, proposed
conditional approval of Missouri’s attainment demonstration SIP for
the lead NAAQS nonattainment area of Herculaneum, Missouri. 

The hourly source emission option was used in the model to represent
emissions from sources that were not throughput or process-dependent. 
These sources included fugitives from the drossing area and wind erosion
areas (see sources 70002 – 70010 in Attachment 3).  Unless specific
hourly or daily operating limitations were applied to a process or
activity, sources at the plant were modeled based on a quarterly
average.  Many emission sources at Doe Run do not run continuously,
twenty-four hours a day, seven days a week.  However, in order to
account for the variety of meteorological conditions simulated in the
analyses, the model was run using an average emission rate assuming
24-hour operation for most sources.  

One source where emissions are not expected to be uniform across all
days is roadways.  The control strategy modeling attributes a 95 percent
control efficiency to paved in-plant roads and paved truck haul routes
external to the plant.  This control percentage was modeled uniformly
across all days modeled, but hourly emissions varied according to the
amount of traffic.  Given typical operating conditions, the Herculaneum
smelter generally experiences somewhat less activity on weekends than on
weekdays.  The Consent Judgment and Work Practices Manual require
operation of the wet sweeper a minimum of Monday through Friday, and the
regenerative air sweeper must operate Monday through Friday as well as
any days concentrate is scheduled for delivery.  While requirements for
a continuously-operating sprinkler system, truck tarping and truck
washing add an additional layer of emission controls, the state
attributed a control efficiency of 95 percent to the sweepers alone. 
Further discussion on road controls may be found in section 5.2.7.1 of
this document. 

The resulting maximum predicted quarterly lead concentration from the
state’s control strategy modeling was 1.492 µg/m3, including the
0.063 µg/m3 background concentration.  A comparison of the
“worst-case” emissions modeled for the 2002 SIP with the
“worst-case” emissions modeled for the 2007 SIP shows a decrease of
about 15 pounds per day or 1,350 pounds per quarter.  Most of the
decreasing emissions occur in the low-level sources that have a
significant impact on the concentrations close to the facility.  The
submitted control strategy modeling shows the area attaining the
standard with all control measures in place.

4.2.4	Modeled Background

A background lead concentration value was calculated for Herculaneum and
included in all modeled concentrations.  The background concentration
represents any naturally occurring lead in the atmosphere, air
concentrations of lead from distant sources, and sources of lead not
captured by the Herculaneum lead emissions inventory.  In MDNR’s
Response to Comments Received on Proposed SIP Revision, the state
commented that the background concentration is likely resulting from
secondary (e.g., re-entrained historical lead deposition from the plant)
or primary impacts caused by Doe Run sources not modeled or not
accurately characterized.  

The background value was determined by analyzing the concentrations at
the Ursaline, Bluff and High School air quality monitors during the
period April 2000 – May 2005.  These monitors were used in the 2002
SIP background concentration calculations, and have different geographic
orientations to the smelter: Ursaline - distant south, Bluff –
proximate north, High School – middle scale northwest (see Attachment
2: Map of Herculaneum Monitors, 1982 - 2005).  By considering
geographically dispersed monitors, MDNR sought to better represent the
entire area, as well as decrease the possibility of one specific,
unidentified source unduly influencing the background concentration. 
The evaluation eliminated concentrations due to sources at or around the
smelter by only using monitor results from days the wind trajectory was
not directed from the Doe Run facility toward the individual monitor. 
MDNR deemed, and EPA agreed, the wind directions appropriate for
estimating background values at the Ursaline monitor were from 40
degrees to 280 degrees.  Winds from 256 degrees to 142 degrees were
included as valid background wind directions for the Bluff monitor.  The
Dunklin High School had valid background wind directions of 241 degrees
to 88 degrees.  The background concentrations were averaged over these
days, producing a background concentration of 0.063 µg/m3. 

With the 2002 SIP, using the same three monitors and the same
representative wind directions, a background concentration of 0.130
µg/m3 was calculated.  The difference in background values found during
the 2002 SIP development and the 2007 SIP development is believed to be
caused by a decrease in primary and secondary (e.g., re-entrained
historical lead deposition from the plant) plant emissions due to
effectiveness of 2002 SIP controls.  

	Control Strategy/Contingency Measures

5.1	Source Contributions

	Table 10 provides the control strategy throughput, control efficiency,
and time restriction inputs for sources in and around the plant.  

Table 10.  Summary of Attainment Demonstration Modeling and Control
Efficiencies.







	Source ID	Source Description	Throughput (TPD or VMT/day)	Capture /
Control Efficiency*	Time Restriction**

10001A1	Concentrate Truck Unloading	1800	90%	6 am - 10 pm

10001A2	Transfer Concentrate to Storage/Railcar	1800	80%	6 am - 10 pm

10001B1	Loading Concentrate for Storage into Railcar	1187	50%	6 am - 10
pm

10001B2	Unloading at Tipper	1800	90%	6 am - 10 pm

20001A	Load Sinter to Railcar	500****	N/A	N/A

20001B	Railcar Sinter Dump to Unloader	500****	N/A	N/A

20002	Load Sinter to Truck	500****	N/A	N/A

20003	Unload Sinter Truck at Storage	500****	N/A	N/A

20004	Baghouse #5 Fume Loading to Railcar	13***	95%	N/A

20004B	Fume Unloading at South Storage	13***	90%	12 pm - 6 pm

20004C	Loading of Fume to Railcar at South Storage	13***	90%	6 am - 10
pm

20005	Sinter Mix Room Concentrate and Sinter	1800	90%	N/A

20005	Sinter Mix Room Fume	13***	90%	N/A

20006	Sinter Plant Building Fugitives	2160	94%	N/A

20007	Baghouse No. 3 Fugitives	2160	N/A	N/A

30001	Main Stack	4944	+	N/A

30002	Blast Furnace Fugitives	2160	97%	N/A

30011-13	Baghouse No. 5 Vents	2160	N/A	N/A

40004/5	Dross Kettle Heat Stacks	1260	50%	N/A

40006	Dross Plant Fugitives	1260	90%	N/A

50006	Refinery Plant Fugitives	888	90%	N/A

50007	Baghouse No. 8	888	++	N/A

50008	Baghouse No. 7 & 9	1260/888	+++	N/A

50011-18	Kettle Setting Heat Stacks	888	50%	N/A

60001/2	Strip Mill Heat Stacks	100.8	N/A	N/A

60003	Strip Mill Baghouse	100.8	N/A	N/A

60004	Low Alpha Baghouse	0.96	N/A	N/A

60005-8	Strip Mill Roof Vents	100.8	N/A	N/A

70100-122	Road AB	155.30	95%	Traffic Scalars

70150-213	Road BC	425.13	95%	Traffic Scalars

70250-252	Road CD	15.43	95%	Traffic Scalars

70300	Road DE	1.37	95%	Traffic Scalars

70350-358	Road EF	9.63	95%	Traffic Scalars

70400-406	Road DF	13.09	95%	Traffic Scalars

70450-454	Road FG	7.13	95%	Traffic Scalars

70500-513	Road GH	20.88	95%	Traffic Scalars

70550-553	Road GK	0.60	95%	Traffic Scalars

70600-612	Road HL	2.68	95%	Traffic Scalars

70650-669	Road IJ	10.73	95%	Traffic Scalars

70700-703	Road KM	1.01	N/A	Traffic Scalars

*Capture/control efficiencies are applied to a zero control baseline.

**Time restrictions, as required by the Consent Judgment.  Also included
in the Work Practice Manual.

***Fume loading and unloading was based on the 1170 tons per quarter SIP
throughput limit scaled back to daily throughputs shown above.

****Sinter to storage transfers are limited to 45,000 tons per quarter
SIP throughput limit scaled back to the daily throughput shown above.

+ Main stack emissions are limited to 794 lbs/day, as required in
Missouri Rule 10 CSR 10-6.120.

++ Baghouse 8 emissions are limited to 8.2 lbs/day, as required in
Missouri Rule 10 CSR 10-6.120.

+++ Emissions from Baghouses 7 & 9 combined stack are limited to 56.6
lbs/day, as required in Missouri Rule 10 CSR 10-6.120.



	Emissions from the roads were modeled every hour and varied according
to the predicted amount of traffic.  The largest number of vehicles on
the roads occurs during the daylight hours.  

The hourly source emission option was used in the model to represent
emissions from sources that were not throughput or process-dependent. 
These sources included fugitives from the drossing area and wind erosion
areas (see sources 70002 – 70010 in Attachment 3).  

Description of Selected Control Measures

Control measures employed by Doe Run focus on reducing fugitive
emissions through engineering and work practices associated with
materials handling, processes, buildings, and transportation. 
Descriptions of the specific individual strategies follow below.  For
estimates of emission reductions associated with implementation of these
controls, see Attachment 12, Clarification Letter from MDNR to EPA,
March 19, 2008.

Concentrate Delivery (Sources 10001A1, 10001A2, 10001B1)

Lead concentrate is delivered to Herculaneum via truck.  Upon arrival,
truck operators elevate one end of their truck bed and transfer the
concentrate into a hopper at the unloading station.  The concentrate is
conveyed up out of the hopper on a conveyor belt and deposited either
into an awaiting railcar, or if no railcar is available, into a storage
pile in the yard.  When a railcar becomes free, the concentrate is
picked up with a front end loader and deposited into the railcar.  The
railcar transports the concentrate to the railcar tipper building where
it is unloaded and transferred to the sinter plant mixing bins.  Sources
associated with the concentrate delivery process are: Concentrate Truck
Unloading (10001A1), Transfer Concentrate to Storage/Railcar (10001A2),
and Loading Concentrate from Storage into Railcar (10001B1).

The 2007 Consent Judgment and Work Practices Manual limit concentrate
handling hours for all three sources to between 6 a.m. – 10 p.m. 
Delivery (10001A1, 10001A) is limited to 1,800 tons of concentrate per
day, and loading from storage to railcar (10001B1) is limited to 1,187
tons per day.  Further controls minimizing concentrate handling
emissions are described below.

Concentrate Truck Unloading (10001A1).  

When trucks arrive at the Herculaneum facility, they will back into a
new “unloading station” consisting of a three-sided and roofed
enclosure, with the truck backing into the fourth-side opening and
dumping the concentrate into a hopper.  Currently there is no wind
barrier or enclosure.  The enclosure will significantly decrease wind
exposure.  According to AP-42 13.2.4, Aggregate Handling and Storage
Piles, reducing exposure to the wind, e.g., reducing wind speeds from
medium (7.5 mph) to low (1.3 mph), can reduce emissions up to 90%.  The
new unloading station enclosure is therefore conservatively attributed
an 80% control efficiency.  

The 2007 Consent Judgment also requires the concentrate to have a six
percent minimum moisture content.  Increasing the moisture content
causes particles to aggregate and smaller, easily airborne particles to
cement to larger particles.  As concentrate is transferred out of the
truck, Doe Run will take random grab samples and send them to the
laboratory for analysis.  Data on moisture content per concentrate truck
are collected by the company to not only ensure a six percent minimum
moisture content, but also in order to accurately pay for the amount of
product minus water delivered, creating a financial incentive for Doe
Run to have an accurate evaluation.  AP-42 13.2.4 states that PM
emissions can be reduced up to 60% when going from moderate to high
moisture conditions.  MDNR applied a 50 percent control efficiency to
the three concentrate delivery emission sources due to the moisture
control.  

Taken together, the reductions expected from increasing the moisture
content of the lead-bearing material coupled with additional control of
the partial enclosure result in an aggregate reduction of 90% during
truck unloading.  

In addition, trucks are required to be tarped during transit, whether
full or not.  After dropping off concentrate, all trucks must be
processed through the water wash system.  These requirements, which are
also required under EPA’s Resource Conservation and Recovery Act
(RCRA) 7003 Order, will reduce the amount of dust and concentrate
emitted from trucks as they travel to and from the plant.  These
controls add further confidence to emission reduction estimates but were
not assigned individual control efficiencies.

Transfer Concentrate to Storage/Railcar  (10001A2).  

	Once the concentrate is unloaded into the hopper, it moves up a
conveyor belt and is released either to a railcar or to a storage pile. 
This transfer process is represented by Source 10001A2 Transfer
Concentrate to Storage/Railcar.  Reducing the amount of handling and
material movement reduces opportunities for dust to release into the
air.  The most direct transfer of concentrate deposits it directly from
the hopper to the railcar, thereby bypassing the storage pile step.  The
Consent Judgment requires Doe Run to directly load a minimum of eleven
railcars each day concentrate is delivered, and Doe Run will “endeavor
to convey concentrate directly to railcars at all times feasible”
(Work Practices Manual, pg. 14).  Doe Run is unable to coordinate
concentrate delivery with railcar availability and sinter processing at
this time, making the storage pile alternative a necessary fallback.  

	Emissions from source Transfer Concentrate to Storage/Railcar (10001A2)
will be reduced through application of a drop sleeve to the end of the
conveyor.  As material falls through the air, it displaces the air it
moves through.  In uncontrolled situations, this displaced air is able
to interact with the falling material and produce fugitive emissions.  A
drop sleeve reduces emissions by minimizing the amount of interaction
the displaced air has with the material being deposited.  Although AP-42
does not provide documentation on the performance of drop sleeves, these
devices are widely used throughout the grain industry, where they can
substantially minimize visible emissions.  

	The six percent minimum moisture content described under Source 10001A1
also applies to this source.  Applying the 50 percent moisture control
efficiency, as well as a drop sleeve efficiency of 60 percent, results
in an 80 percent overall control efficiency for Source 1001A2.

Loading Concentrate from Storage into Railcar (10001B1).  

	As described above under Source 10001A2, in the event that there are no
available railcars, concentrate is deposited into storage piles.  When a
railcar becomes available, a front end loader lifts the concentrate from
the storage pile into the railcar.  The 50 percent moisture control
efficiency described under Source 10001A1 also applies to this source. 
Adding additional confidence is the Work Practices Manual requirement to
chemically stabilize inactive concentrate storage piles once every
month.

Railcar Tipper Building (10001B2).  

Delivered concentrate as well as a number of other materials may be
loaded into railcars and fed into the sinter machine.  These materials
include a combination of concentrate, sinter that was not fully
processed the first time through the sinter machine, slag from the blast
furnace, baghouse fume, plant cleanup/spills, sump cleanings, and flux
agents.  Once the railcars are loaded, they move to the railcar tipping
building, where the railcar is tipped and emptied into the railcar
hopper.  Contents are then conveyed to and mixed in the sinter plant mix
room, crushed, and fed into the sinter machine.  

	The Consent Judgment and Work Practices Manual require the railcar
tipper building’s north and south doors to be closed except when
moving cars in and out.  Doors must remain closed until all material is
conveyed out of the hopper.  This will allow time for the material to
settle before the doors are reopened.  Emissions will be significantly
reduced due to the complete enclosure of the railcar tipper.  As the
drop operation equation in AP-42 indicates, mean wind speed
significantly influences the quantity of emissions.  Enclosing the
operation is thus attributed a 90% emissions reduction.  

Sinter Mix Room Fume, Sinter Mix Room Concentrate and Sinter (20005).  

Concentrate, sinter being recycled through the sinter machine, slag from
the blast furnace, baghouse fume, plant cleanup/spills, sump cleanings,
and flux agents are combined in the Sinter Mix Room (20005).  The sinter
mix room is currently totally enclosed, with zero hot processes housed
in the room.  The control efficiency builds on a baseline of no controls
(in this case, a mix room without any enclosure), and a 90 percent
control efficiency is therefore conservatively attributed to this
enclosure.

Process Building Fugitives: Sinter Plant (20006), Blast Furnace (30002),
and Dross and Refinery Buildings (40006 & 50006)

Design value modeling identified all process building fugitive emissions
as source areas of concern.  In order to minimize the amount of dust
released within the buildings, and the amount of air (potentially
containing dust) that escapes from the buildings without first being
processed through a control device, a number of controls are required
for each process building (the Sinter Plant, Blast Furnace, and Refinery
and Dross Buildings).  Three different yet closely related building
enclosure and ventilation controls are required for all three of these
buildings.  First, Doe Run must modify all personnel doors to
automatically close, and shift supervisors will be responsible for
minimizing the opening of large equipment doors.  Second, siding
inspections will occur every two weeks and repairs promptly addressed. 
These first two measures will minimize the number of openings in the
buildings that air could flow in and out of, potentially creating draft
and disturbing the expected air flow conditions inside the buildings. 
Third, Doe Run will complete a ventilation study and identify
enforceable conditions, such as fan amperages, necessary to ensure
effective particle capture and appropriate ventilation.  The details and
findings of the ventilation study will be submitted by the state to EPA.
 Following submittal of study findings and approval through State and
Federal processes, the measures will become  enforceable under the Work
Practices Manual and/or the Consent Judgment.  Collectively, these three
controls will improve process emissions capture (and thereafter send
them to a control device), and better ensure any other lead emissions
released within the building fall within the confines of the building. 
Taken together, a 90 percent control was considered appropriate for
these three controls.

The overall objective, or control measure, these three controls work to
achieve is effective building enclosure.  By minimizing building
openings (siding and doors) and ensuring adequate ventilation, the
building enclosures will be operated in such a fashion as to minimize
the escape of fugitive emissions.  The SIP requires this overall control
measure, and also requires adequate ventilation under the ventilation
study element.  The control efficiency described above is attributed to
the overall building enclosure control measure and is so included in the
attainment demonstration modeling.  However, although the adequate
ventilation and overall building enclosure control measures are required
under the SIP, the submitted SIP does not include all necessary
enforceable conditions (such as fan amperages or flow rates) associated
with the ventilation study to ensure that these control measures are
met.  EPA is therefore proposing conditional approval of the SIP as it
provides substantial progress toward improving air quality, and the
State asserts that it will adopt and submit the missing enforceable
ventilation conditions to EPA by date certain, which will be no later
than one year following any EPA approval of the plan.  See the Notice of
Proposed Rulemaking, proposed conditional approval of Missouri’s
attainment demonstration SIP for the lead NAAQS nonattainment area of
Herculaneum, Missouri, for a description of what action EPA proposes to
take with respect to the ventilation study.  See Attachment 14,
“Ventilation Study Commitment Letter from MDNR to EPA, August 18,
2008,” for the state’s ventilation study submittal commitment.

Sinter Plant Building Fugitives (20006)

Sinter feed is conveyed from the sinter mixing room to the sinter
machine.  The sinter machine heats the feed in order to remove a
significant amount of SO2.  Inside of the sinter machine, material
passes through on top of a moving grate.  Air pushes up through the
grate, oxidizing the lead sulfide into lead oxide and releasing SO2. 
The gases from the front of the sinter machine are vented to an
electrostatic precipitator and then to an acid plant, which produces
sulfuric acid as a smelting byproduct.  All gases from the sinter
machine are ultimately vented to the Number 3 Baghouse.  The sinter is
sorted for size, and sinter that is too small is cooled and sent back to
be remixed and reprocessed through the sinter machine.  The remaining
sinter roast is crushed in the smooth roll crusher and sent on to the
blast furnace.  Air from the cooling process is vented to the Cooler
Baghouse, and air from the crushing process is vented to the Crusher
Baghouse.  (See Attachments 4 and 5, Process Flow Diagrams.)

The Carrier Cooler Baghouse and sinter plant building are attributed a
90 percent building enclosure control (as described under 5.2.4).  The
sinter plant building enclosure includes enclosing the sinter loading
area on the northeast corner of the sinter area.  Accounting for the 90
percent enclosure control leaves 10 percent of the original sinter plant
fugitive emissions.  This remaining 10 percent is controlled an
additional 40 percent by the sinter machine wheel tunnel ventilation
project, through wheel tunnel enclosure and ventilation.  The grate
holding the sinter feed is attached to two wheel tunnels which run
lengthwise along the edges of the grate, and rotate to propel the grate
forward through the sinter machine.  As the sinter is heated and excess
air pushes up through the grate, hot gases and dust can escape out the
sides of the sinter machine, producing building fugitives.  Doe Run will
enclose the hot gas section of the sinter machine’s wheel tunnels, to
capture the potential fugitives.  The wheel tunnel ventilation will be
designed with a rate of fifteen thousand (15,000) actual cubic feet per
minute, and the captured gases routed to Number 3 Baghouse.  The wheel
tunnel ventilation will be continuously operated, and the ventilation
rate or an equivalent fan amperage measurement will be monitored. 
Within the sinter building, approximately 40 percent of airborne lead
particles originate from the sides of the sinter machine.  Enclosing the
wheel tunnel will trap the gases and dust, while the wheel tunnel
ventilation will pull the captured air toward the Number 3 Baghouse.  In
all, the sinter plant building fugitive source is attributed a 94
percent control efficiency.

Number 3 Baghouse, which services the sinter plant, will also have an
upgraded bag and bag cleaning system.  No increased control efficiency
was attributed to this requirement.

Blast Furnace Building Fugitives (30002)

The sinter roast is conveyed on a charge belt and fed into the blast
furnace together with coke, slag, other byproducts, and other charge
agents.  This mix is fed into the top of the furnace and descends into
the smelting zone of the furnace, where the blast furnace feed becomes
molten.  The combination of feed rate and the amount of air flow,
regulated by the tuyeres at the bottom of the blast furnace, determines
the blast furnace process rate.  Air allowed through the tuyeres passes
through the column of material and is vented to the Number 5 and Number
6 Baghouses.  The molten material is tapped into settlers, where the
lead separates to the bottom and slag is tapped off the top to be
recycled back as sinter feed or stored at the plant’s south-end slag
storage area. 

	

	Multiple projects were undertaken in the blast furnace building.  Four
additional controls are included in the Consent Judgment, which in
aggregate earn a 70 percent control efficiency relative to zero
controls.  First, the Number 1 blast furnace will be relocated closer to
the ventilation source, reducing ductwork as well as charge belt length
and an improved ventilation hood system will be installed on the front
of the blast furnace.  Second, ductwork will be installed allowing
continuous ventilation of the blast furnace doghouse area.  Third,
Number 5 Baghouse ventilation will be maintained at 300,000 actual cubic
feet per minute and Number 6 Baghouse ventilation will be maintained at
50,000 actual cubic feet per minute.  These ventilation rates will be
verified as identified in the ventilation study (see 5.2.4).  Fourth,
emissions associated with the tuyeres will be reduced through operation
of automatic flow controllers.  Occasionally voids form in the molten
liquid contained in the furnace.  These voids create passages for high
amounts of air flow to pass through, carrying excess amounts of dust out
the furnace top.  The flow controllers will sense where voids form and
reduce the amount of air being released beneath.  Together, these
multiple controls account for a 70 percent control efficiency, which,
combined with the 90 percent building enclosure control, results in a
total 

97 percent control efficiency for the blast furnace building.

	An alarm system will also be installed at the Number 5 Baghouse.  The
alarm system will alert operators when there is a malfunction in the
Number 5 Baghouse fans.  No additional control efficiency was associated
with this requirement.

Dross and Refinery Building Fugitives (40006 & 50006)

The lead bullion from the blast furnace is next transferred to the Dross
and Refinery Buildings, where it is processed through a number of
kettles and material is further separated.  First, the lead bullion is
cooled, allowing a copper dross to rise to the top and be separated from
the lead dross.  The lead dross is heated and sulfur dross removed. 
Lead dross produced from these activities is progressively poured into
large heated kettles, allowing separation of copper, silver, and zinc,
and resulting in a product of 99.999 percent pure lead.  Number 7
Baghouse services the dross section of the building, while Numbers 8 and
9 Baghouses service the refinery.  Dross kettle heat stacks and refinery
kettle heat stacks are additional source points from this portion of the
process, and are further described in Section 5.2.5 below.

As described in further detail in section 5.2.4 above, three different
controls are required to improve fugitive capture and ventilation within
the refinery and dross building.  Together, the door closure, siding
inspection and repair, and ventilation study are attributed a 90 percent
control efficiency.

Doe Run must also modify the stack height of the Number 7 & 9 Baghouse
stack to a minimum of one hundred fifty feet, and not exceed a 24 hour
stack emissions limit of 56.6 pounds.  Pleated filters will also be
installed in the Number 7 & 9 Baghouse.  No additional control
efficiency is associated with these changes.

Kettle Heat Stack Cameras (Dross Heat Stacks (40004-5) & Refinery Kettle
Heat Stack (50011-18))

In order to further purify the lead bullion and separate material with
high lead content from that with only trace amounts of lead, the last
steps transfer molten bullion into large kettles.  These kettles are
heated and remaining trace metal impurities are separated and removed in
the dross and refinery departments.  Occasionally the kettles crack or
break, resulting in the molten content leaking into the hot kettle pits
and ultimately vaporizing.  These vapor emissions are vented to the
dross and refinery kettle heat stacks, creating a visible plume of
smoke.  

	Doe Run did not closely monitor kettle cracks or breaks in the past and
had no formally documented response procedure or responsible official. 
The 2007 Consent Judgment requires cameras to monitor the opacity of the
kettle heat stacks, and operators to watch and take immediate action to
mitigate emissions should they observe a problem.  The Work Practices
Manual specifies that operators in the sinter plant control room will be
primarily responsible for watching and reacting to the monitors, and the
security office will act as a backup.  The camera will additionally feed
into the environmental department, blast furnace control room, and the
general foreman’s office.  Upon observation of smoke, the refinery
department will be contacted and will immediately shut off the
appropriate kettle burners, transfer the material out of the problem
kettle, and inspect both the kettle and kettle setting prior to resuming
use.  The monitoring requirement is expected to improve response time,
while specified work practices will improve clean ups.  Work practices
(as required in the refinery section of the Doe Run-Herculaneum Work
Practices Manual) include kettle and kettle setting inspections prior to
use.  Recognizing there will still be vaporization between the time the
leak occurs and when operators are able to clean the very hot kettle and
kettle settings, this control is attributed a 50 percent control
efficiency.

Number 5 Baghouse Fume Handling (Sources 20004, 20004B, 20004C)

Number 5 Baghouse Fume Loading to Railcar (20004), Fume Unloading at
South Storage (20004B), Loading of Fume to Railcar at South Storage
(20004C).  

	Air from the blast furnace is ventilated through Number 5 Baghouse. 
When the sinter plant is operational, the fine blast furnace baghouse
fume is collected and fed directly into the sinter plant, allowing for
reprocessing.  When the sinter plant is not operational, the fume is
screw conveyed to a railcar, employing a “pugger” process.  The
railcar unloads the fume onto a fume storage pile at the south end of
the plant.  When this fume is needed, a front-end loader lifts it from
the storage pile and deposits it into a railcar.  The railcar feeds back
into the sinter plant feed circuit, ultimately tipping material into the
sinter plant mixing bins.

	The Consent Judgment and Work Practices Manual require the
screw-conveyed fume to contain an eight percent daily average moisture
content.  Doe Run may unload no more than 

13 cars of fume to storage per quarter, for a maximum total of 1,170
tons of fume unloaded to storage per calendar quarter.  The Consent
Judgment and Work Practices Manual also limit unloading hours to between
noon and 6 p.m., and fume loading from storage pile to railcar to
between the hours of 6 a.m. and 10 p.m.  

All three fume-related sources (Sources 20004, 20004B, and 20004C)
received a 

90 percent control efficiency due to the eight percent moisture control.
 A moisture sample will be taken as the fume drops from the screw
conveyor to the railcar.  The Work Practices Manual additionally
requires the fume storage pile to be sampled daily when it is being
worked on, and maintained at an average daily eight percent moisture. 
MDNR describes the material as “…a fine baghouse fume with
characteristics similar to powdered gypsum.  The material is
screw-conveyed from a saturated container so the material is exceedingly
wet” (as communicated in MDNR’s March 19, 2008, clarification letter
to EPA, included as Attachment 12).  The inactive portion of the fume
storage pile will be chemically stabilized once every month.  AP-42
13.2.4 lists watering and the use of chemical wetting agents as the
principal means for control of storage pile emissions.  As discussed in
that document, continuous chemical treatment or more extensive watering
of material loaded onto piles can reduce total particulate emissions
from storage operations by up to 90 percent.  

Doe Run is also required to install and maintain a drop sleeve that
reaches from the bottom of the screw conveyor to the top of the railcar.
 As material falls through the air, it displaces the air it moves
through.  In uncontrolled situations, this displaced air is able to
interact with the falling material and produce fugitive emissions.  A
drop sleeve reduces emissions by minimizing the amount of interaction
the displaced air has with the material being deposited.  Although AP-42
does not provide documentation on the performance of drop sleeves, these
devices are widely used throughout the grain industry, where they can
substantially minimize visible emissions.  An additional 50 percent drop
sleeve control efficiency was applied to 

Source 20004, Number 5 Baghouse Fume Loading to Railcar, resulting in a
combined efficiency of 95 percent.  

	

Roads: Street Sweeping and Unpaved Haul Road

Street Sweeping (70100-122, 70150-213, 70250 – 252, 70300, 70350-358,
70400-406, 70450-454, 70500-513, 70550-553, 70650-669)

	Doe Run is required to operate a regenerative air sweeper on paved
truck haul routes external to the plant and a wet sweeper on paved areas
in the plant.  The Consent Judgment requires operation of both sweepers
for a minimum of 6 hours per day, Monday through Friday.  The Work
Practice Manual additionally requires the regenerative air sweeper to be
operated on any day that concentrate is scheduled for delivery to
Herculaneum.

	

	The regenerative air sweeper employs newer, more effective technology
than the previously operated Doe Run sweeper.  Quarterly road samples
will be taken at four sites, to verify sweeper effectiveness.  In
addition to the wet sweeper activity on the paved areas in the plant,
in-plant sprinklers will be used continuously, 7 days a week, on
in-plant roads.  The Consent Judgment also requires concentrate delivery
trucks coming to or leaving the smelter to be tarped except during the
unloading and washing process.  Following unloading, trucks must proceed
through a water washing system to remove visible material from the truck
tires and chassis.  The concentrate receiving procedure is further
outlined in the Work Practices Manual.   

	Sweeping of the paved in-plant roads and paved truck haul routes
external to the plant was attributed a 95 percent control efficiency. 
The state modeled this control percentage uniformly across all days
modeled, i.e., seven days a week.  However, as discussed above, the
Consent Judgment and Work Practices Manual only require operation of the
wet sweeper a minimum of Monday through Friday, and the regenerative air
sweeper must be operated Monday through Friday as well as any days
concentrate is scheduled for delivery.  The Herculaneum smelter
generally experiences somewhat less activity on weekends than on
weekdays, and the continuously-operating sprinkler system, truck tarping
and truck wash requirements add an additional layer of emission
controls.  The model did not attribute any separate control efficiency
to the sprinkler system, truck tarping or truck wash requirements, nor
did it decrease activities on weekends.  There is therefore a degree of
uncertainty regarding the controls required to maintain modeled levels
of emissions control on weekends.  In the event that Herculaneum
monitors an exceedance of the 1.5 µg/m3 standard, EPA will pursue
negotiation of additional weekend road emissions control strategies with
Doe Run and MDNR. 

“North End of Blast Furnace to Refinery Dock” Haul Road (70600 –
612)

This unpaved road is used approximately two weeks out of each calendar
quarter to transport slag from the north end of the blast furnace to
slag storage.  During any day or part of the day the road is being used
to haul slag, water will be applied at a rate of 350 gallons once every
four hours.  MDNR has clarified the relationship between rainfall and
wetting of the road:

A calculation was made to determine the amount of water present on the
slag haul road during a one-quarter inch rain.  The road is 1,073 feet. 
If a conservative width of 12 feet is assumed, the road surface is
(1073)(12) = 12,876 square feet.  A one-quarter inch rain is equivalent
to 0.020833 feet of rain.  A conservative estimate of the volume of rain
on the road is therefore:

(12,876)(.020833)(gal/0.13368 cubic feet) = 2007 gallons.

This exceeds the required 350 gallons required every four hours of the
day 

(350 gallons * 24 hrs/4 = 2100 gallons/day).  We have concluded that a
quarter-inch rain serves as an equivalent control.

MDNR additionally provided the following calculations in support of the
350 gallons once every four hours:

Annual Evaporation Rate	47	in/yr





	Potential Average Hourly Daytime Evaporation Rate	0.3055	mm/hr















Quarterly Slag Tonnage	15000	ton/qtr





	Slag Payload	24	tons





	Quarterly Slag Haulage	625	Truck/qtr















	2	wk/qtr







7	dy/wk







12	hr/dy







168	hr/qtr















Average Hourly Traffic Rate	7.44	Truck/hr















Application Intensity	0.73	L/m2	0.0179	gal/ft2	0.028743	in-Precip
Equivalent

Application Interval	4	hr















Control Efficiency (%)	90.04	%















Segment Length	1229	ft





	Road Width	15	ft





	Water Area	18435	ft2















	330	gal/application







Use 350 gallons/application











	Reference:  Inspection Manual for PM10 Emissions from Paved/Unpaved
Roads and 

Storage Piles, USEPA, October 1989, Equation 4-2.







Within 72 hours of completing batch hauling, Doe Run will apply a road
surfactant to the haul road, decreasing emissions from wind or other
road traffic.  Considering the water and chemical surfactant together, a
90 percent control is attributed to this stretch of road.

Reasonably Available Control Measures (RACM) including Reasonably
Available Control Technology (RACT) 

Section 172(c)(1) of the CAA requires nonattainment areas to implement
all Reasonably Available Control Measures (RACM), including emissions
reduction through the adoption of Reasonably Available Control
Technology (RACT), as expeditiously as practicable.  EPA interprets this
as requiring all nonattainment areas to consider all available controls
and to implement all measures that are determined to be reasonably
available, except that measures which will not assist the area to more
expeditiously attain the standard are not required to be implemented. 
See 58 FR 67751, December 22, 1993, for a discussion of this
interpretation as it relates to lead.  

In the April 14, 2006, SIP call, EPA did not list a new RACT analysis as
a required element of the SIP submittal.  Even though not required by
the 2006 SIP call, a RACT/RACM analysis is still included with the 2007
SIP submittal.  No additional RACT measures were identified that would
expedite attainment or reasonable further progress, and the plant has
not changed significantly from when the previous RACT/RACM evaluation
was completed.  Some previously implemented RACT/RACM measures, i.e.,
types of controls, were strengthened through incorporation of more
detailed, enforceable work practices in the Work Practices Manual. 
Table 11 summarizes EPA’s RACT/RACM analysis, reflecting improvements
as a result of the 2007 Work Practices Manual.

Table 11.  Doe Run Herculaneum RACM/RACT Analysis.

Description of Measure	Explanation	Used in Control Strategy

Pave, vegetate or chemically stabilize access points where unpaved
traffic surfaces adjoin paved roads	All primary traffic routes inside
the plant have been paved.  Unpaved areas are used only for material
storage.  The unpaved slag haul route, typically used for a two-week
period once a calendar quarter, will be watered during slag hauling and
chemically stabilized upon completion of hauling as required by the May
2007 Consent Judgment (CJ) and January 2007 Work Practices Manual (WPM).
Yes.

Require dust control plans for construction or land clearing projects
Such sources have not been identified in the emission inventory or
modeling study and are therefore not addressed in the control strategy. 
Nearly all of the land near the active areas of the plan has been
cleared and much of it is paved.  The WPM contains construction
guidelines for capital construction projects.  Doe Run will address
these types of issues on a case-by-case basis after the plan controls
are implemented.	No.

Require haul trucks to be covered	This measure is required as part of
the CJ.  Concentrate trucks must be properly tarped except when
unloading or during washing.  Tarps must be maintained with no tears or
openings.	Yes.

Provide for traffic rerouting around or rapid cleanup of temporary
sources of dust on paved roads	Doe Run is required to operate sweeper
machines on paved roads inside and beyond the plant.  This facilitates
quick cleanups of spills of any lead-bearing dust on paved areas.  The
2007 WPM contains clean up specifications for temporary sources of dust
spilled on paved surfaces outside the plant.	Yes.

Develop traffic reduction plans for unpaved roads	All primary traffic
routes inside the plant have been paved.  Unpaved areas are used only
for material storage.  Use of the unpaved slag haul route is to occur
during daylight hours for approximately 2 weeks per calendar quarter, as
specified in the 2007 WPM.	Yes.

Limit use of recreational vehicles on open land	Recreational vehicles
are not permitted in the Doe Run, Herculaneum plant.	Yes.



Require improved material specifications for and reduction of usage of
skid control sand or salt.	These materials do not contain appreciable
amounts of lead and, therefore, control of this material is not
applicable for the lead control strategy.  In addition, use of these
materials is very limited within the Doe Run, Herculaneum plant.	No.

Require curbing and pave or stabilize shoulders of paved roads	All
primary traffic routes inside the plant are paved.  Road shoulders
inside the plant have not been identified as sources of lead-bearing
dust.	No.

Pave or chemically stabilize unpaved roads	All primary traffic routes
inside the plant have been paved.  Unpaved areas are used only for
material storage.  The unpaved slag haul route, typically used for a
two-week period once a calendar quarter, will be watered during slag
hauling and chemically stabilized upon completion of hauling as required
by the 2007 CJ and 2007 WPM.	Yes.

Pave or chemically stabilize unpaved parking areas	Parking areas are
paved.	Yes.

Require dust control measures for material storage piles	The 2007 CJ and
WPM require concentrate and fume being deposited to storage piles to be
maintained at a minimum average daily percent moisture.  Storage piles
will also be wet and chemically stabilized as outlined in the CJ and
WPM.  	Yes.

Provide stormwater drainage to prevent water erosion onto paved roads
Much of the paved areas feed a storm water collection system to minimize
erosion and treat the runoff.  Lead emissions due to water erosion on
paved roads was not identified as a significant contributor to lead
concentrations in Herculaneum.  Doe Run is required to operate sweeper
machines on paved roads inside and outside of the plant.  	Yes.

Require revegetation, chemical stabilization, or other abatement of wind
erodible soil.	The 2007 CJ and WPM require concentrate and fume to be
maintained at a minimum average daily percent moisture.  Storage piles
will be wet and chemically stabilized as outlined in the CJ and WPM. 
The emission inventory and dispersion modeling do not show wind erosion
events as significant contributors of lead emissions at the Doe Run,
Herculaneum facility.	Yes.

Rely upon the soil conservation requirements to reduce emissions from
agricultural operations	No agricultural operations involving soil
disturbance occur at the Doe Run, Herculaneum plant.	No.



Although not directly relevant to RACT/RACM, we note that the
Herculaneum primary lead smelter is also subject to 40 CFR Part 63
subpart TTT, the Federal Maximum Achievable Control Technology (MACT)
standard for Primary Lead Smelters.  Subpart TTT requires the
development and use of standard operating procedures manuals for all
baghouses controlling process, process fugitive, or fugitive lead dust
emissions.

Dispersion modeling analysis was conducted to determine if the controls
required by 

the 2007 Consent Judgment control strategy would be sufficient to bring
the area into 

attainment of the standard.  The dispersion modeling submitted by the
state showed attainment of the 1.5 µg/m3 standard, demonstrating that
the control strategy is adequate to bring the area into attainment of
the standard.  In terms of expeditious attainment we again note that the
time between the SIP submission deadline and the attainment date is only
one year, so that additional measures which could be implemented within
that year and achieve reductions before the end of that year would be
even less likely.  For the reasons stated above, EPA proposes to find
that no additional measures will expedite attainment and that the
RACT/RACM requirement is met.

Reasonable Further Progress (RFP)

Section 172(c)(2) of the CAA requires SIPs to provide for Reasonable
Further Progress (RFP) as defined in section 171(1) of the CAA.  Section
171(1) defines RFP as annual incremental reductions in emissions of the
relevant air pollutants as required by Part D, or emission reductions
that may reasonably be required by EPA to ensure attainment of the
applicable NAAQS by the applicable date.  Part D does not include
specific RFP requirements for lead.  

MDNR has demonstrated RFP as required under section 172(c)(2) of the
CAA.  Doe Run is subject to a compliance schedule for implementing: (1)
installation of emission control equipment, (2) enclosure and
ventilation projects to reduce lead emissions, (3) process throughput
restrictions and hours of operation limitations, and (4) work practice
standards.  These are but a few of the SIP controls that are enforceable
through the Consent Judgment and/or the Work Practices Manual.  Given
that all controls contained in the control strategy were required to be
implemented by April 7, 2008, to provide for attainment by April 7,
2008, EPA does not believe additional incremental reductions are
necessary to meet the RFP requirement.  EPA also notes that, since all
of the new controls in the SIP were required to be implemented within
one year of development of the control strategy (April 2007 to April
2008), and that these controls have been demonstrated to be adequate for
attainment, EPA believes that these controls represent the annual
reductions necessary for RFP and attainment.

	Contingency Measures

As required by Section 172(c)(9) of the CAA, contingency measures have
been prepared that may be implemented if the EPA Regional Administrator
determines that: (1) the nonattainment area has failed to make RFP, (2)
there is a failure to implement a control strategy to attain the NAAQS
by the statutory deadline, or (3) monitoring shows that the
nonattainment area has failed to attain the NAAQS by the statutory
deadline.

Contingency Measure Schedule

The attainment date for the Herculaneum area is April 7, 2008, as set in
the April 14, 2006, final SIP call rulemaking (71 FR 19432).  If the
area has an exceedance of the NAAQS during any quarter following the
April 7, 2008, attainment date, the contingency measures will be
implemented according to the schedule outlined in the May 2007 Consent
Judgment, upon written notification of violation from MDNR.  MDNR may
also require implementation of contingency measures if Doe Run fails to
make reasonable further progress (in this instance, timely
implementation of the control measures).

The implementation schedule contained in the May 2007 Consent Judgment
is as follows: Projects 1.a and 1.b will be implemented within 6 months
of receipt of the notice.  Following implementation of 1.a and 1.b, if
any quarter exceeds the standard or Doe Run fails to make Reasonable
Further Progress, MDNR shall notify Doe Run and project 1.c will be
implemented within 18 months of receipt of the notice.  Project(s) to be
identified in paragraph 1.d shall be completed within a time frame
determined by Doe Run and MDNR.  Following implementation of projects
1.c and 1.d, if any quarter exceeds the standard or Doe Run fails to
make Reasonable Further Progress, MDNR shall notify Doe Run and project
1.e will be implemented within 24 months of receipt of the notice.

	In addition, the production limit contingency outlined in “2.
Production Limit” will be triggered as described below.

	Section 172(c)(9) of the Act provides that contingency measures must be
capable of implementation without any further action by the state or
EPA.  Because project 1.d does not contain specific requirements and
associated deadlines, EPA does not consider it a valid contingency
measure.  Implementation of project 1.e is contingent upon completion of
project 1.d and, therefore, is also not a valid contingency measure
under EPA definition.  EPA encourages implementation of any measures
which would reduce lead emissions and supports implementation of the
activities described in contingency measures 1.d and 1.e.  However,
because these two projects do not contain specific requirements and/or
associated deadlines for implementation, EPA does not consider them
contingency measures under section 172(c)(9) and does not approve them
as part of the Federally-enforceable SIP.  In reviewing the adequacy of
contingency measures, we have considered projects 1.a, 1.b, and 1.c as
well as the production decrease measure, as these contingencies fully
meet the requirements of section 172(c)(9) of the CAA. 

Contingency Measures:

Projects

Enclose sinter plant “pugger” (associated with Fume Unloading at
South Storage, Source 20004B)

Pave “North End of Blast Furnace to Refinery Dock” Haul Road (Source
70600 – 612)

Route Dross Kettle Heat Stack (Source 40004, 40005) and Refinery Kettle
Heat Stack (Source 50011-18) to Main Stack (Source 30001)

Implement contingency measures identified as a result of the
Technological Study for Fugitive Dust Control (as described in the 2007
Consent Judgment)

Install dedicated ventilation to the Sinter Plant.  If Doe Run
demonstrates to MDNR that Doe Run will implement Flubor© technology at
the Herculaneum facility within thirty-six months of the date of
notification of the exceedance that would otherwise require Doe Run to
install dedicated ventilation to the sinter plant, Doe Run shall not be
required to install dedicated ventilation.  If Doe Run does not
implement Flubor© technology at the Herculaneum facility within
thirty-six months as demonstrated, Doe Run must install dedicated
ventilation to the sinter plant within 18 months of the earlier of: (1)
the deadline for implementing Flubor© technology, or (2) when Doe Run
determines it will not implement Flubor© technology.  MDNR may extend
the period in which the Flubor© technology is to be implemented if Doe
Run is making reasonable progress toward implementation.

Production Limit

In addition to the above contingency measure projects, should an
exceedance of the quarterly lead NAAQS occur, the quarterly production
limit for refined lead shall be reduced to 95% of the actual production
during the exceedance quarter.  It shall be reduced by an additional 5%
below actual production for each subsequent quarter in which there is an
exceedance, to a minimum production of 35,000 tons of refined lead per
calendar quarter.  In the event that all monitors show attainment in a
quarter following a production decrease, Doe Run may increase the
production level for refined lead by 5% of the attainment quarter’s
actual production provided that Doe Run implements additional control
measures prior to increasing the production level.  These control
measures must be demonstrated to MDNR to reduce impacts on air quality
to an equal or greater extent than the increased production limit will
increase impacts on air quality.  Production may increase to a maximum
of 62,500 tons per calendar quarter (the level assumed in the attainment
demonstration modeling) if the area continues to monitor attainment of
the lead NAAQS.  

6.2	Contingency Measure Impacts

Enclosure of the pugger, accountable for 0.0006 pounds per year lead
emissions in the attainment demonstration modeling, is projected to
entirely eliminate the pugger as an emission source.  Modeling shows
that this will reduce ambient concentrations by 0.018 µg/m3 at the
monitor closest to the facility, which is also the monitor that has
recorded the most exceedances of the standard (Broad Street monitor). 
Paving the haul road running from the north end of the blast furnace to
the refinery dock is expected to result in a 1.226 pound per hour lead
emission reduction.  The paving and the elimination of emissions from
the pugger together result in a reduction of 0.032 µg/m3 at the Broad
Street monitor.  By rerouting the kettle heat stacks from relatively
short stacks to the tall main stack, emissions will not be reduced but
they will be more dispersed, thus lessening their impact by 0.084 µg/m3
at the Broad Street monitor.  The total reduction at the Broad Street
monitor where the highest concentration is most likely to be measured is
0.116 µg/m3 when all three contingency measures are implemented.

Emission reductions associated with the five percent production cut are
more difficult to calculate, in part because of the daily variations in
the operations of the facility.  In order to present a range of emission
reductions, the 62,500 tons per quarter maximum refined lead production
rate and the 35,000 tons per quarter minimum refined lead production
rate were considered.  To calculate the emission reductions associated
with a five percent production cut, a number of assumptions were made. 
First, the throughput values for the various sections of the process
were calculated relative to the total production.  A ratio of the
example total production (e.g., 62,500 per quarter) to that of the total
production in 2005 (2005 actual production rate was 155,088 tons of
refined lead: 42,289 tons in the first quarter, 29,757 tons in the
second quarter, 40,619 tons in the third quarter, 42,423 in the 4th
quarter, for an average 38,772 tons of refined lead per quarter) was
calculated.  This ratio was then applied to the 2005 throughput values
for the various sections of the process, to produce process section
throughput values associated with the level of example total production.
 This scaling process was used for all of the example total production
levels considered here.  Although the attainment demonstration modeled
conditions at the maximum 62,500 tons per quarter total production, it
also used the maximum allowed throughput values for the various sections
of the process.  The scaled throughput values used in this calculation
were used to better reflect predicted actual throughput under a 62,500
tons per quarter production scenario.  

Additionally, stack emission rates were based on recent stack tests,
with adjustments made for post control improvements and production
ratios.  The last assumption was that emissions from storage piles would
not change because of the variable meteorological conditions that drive
these calculations.  The validity of this assumption depends on the
actual conditions experienced.  Wind speeds and precipitation are
meteorological conditions that determine the extent of erosion from
storage piles and other ground surfaces when there is no activity. 
Emissions will be highest when the wind speed is high and the
precipitation is low.  These conditions are very variable and will
determine the extent of erosion.   

With those assumptions, a five percent production cut from the maximum
production rate (62,500 tons per quarter) would result in a quarterly
emission reduction of 1,154.8 pounds of lead or about 12.83 pounds per
day.  The same calculation based on a production rate of 36,842 tons per
quarter (from which a five percent reduction cut would result in a
minimum 35,000 tons per quarter) yields a quarterly emission reduction
of 680.7 pounds of lead or about 7.56 pounds per day. 

Implementation of Contingency Measures

	Doe Run must notify MDNR within 10 days of completion of any
contingency measure.  Sixty days after completion, Doe Run shall propose
an additional quantified contingency measure to be added to the Consent
Judgment, which shall become part of the Consent Judgment and fully
enforceable upon approval by MDNR.  These additional contingency
measures will also be subject to EPA approval as part of the SIP.  Doe
Run may also substitute new control(s) for the above contingency
measure(s) should Doe Run identify and demonstrate to MDNR and EPA’s
satisfaction the alternate control measure(s) would achieve equal or
greater air quality improvements than the contingency measures currently
outlined in the Consent Judgment.  

Changes to contingency measures would require a public hearing at the
state level and EPA approval as a formal SIP revision.  Until such time
as EPA approves any substitute measure, the measure included in the
approved SIP will be the enforceable measure.  EPA does not intend to
approve any substitutions which cannot be implemented in the same
timeframe as the original.  These measures will help ensure compliance
of the lead NAAQS as well as meet the requirements of section 172(c)(9)
of the CAA.  

Enforceability

Enforceable Requirements

As specified in section 172(c)(6) and section 110(a)(2)(A) of the CAA,
and 57 FR 13556, all measures and other elements in the SIP must be
enforceable by the state and EPA.  Missouri’s SIP submittal includes
the May 2007 Consent Judgment and January 2007 Work Practices Manual. 
The Consent Judgment contains all control and contingency measures with
enforceable dates for implementation.  It also includes monitoring,
recordkeeping, and reporting requirements to ensure that the control and
contingency measures are met.  The Work Practices Manual includes these
as well as specific operating procedures and additional reporting
requirements.  The state adopted both documents into the Missouri state
regulations on April 26, 2007, making them state-enforceable.  Upon EPA
approval of the state submission, both documents will become state and
federally enforceable, and enforceable by citizens under section 304 of
the Act.

Process Weight Limits

The majority of the smelter’s emissions are directly linked to the
amount of activity and the level of production occurring within the
plant.  Source emission profiles are created by first determining how
much throughput a process will experience, what the level of production
will be, or what amount of material will be handled. 

Doe Run is required to meet the following daily and quarterly production
limits based on a maximum production limit of 62,500 tons per quarter
and individual process limitations.  Limits were created for each step
of the facility’s process: concentrate and fume handling, sintering,
smelting, and refining.  

Process Weight Limits (Daily)

	Activity	Process Throughput Limit (tons per 24 hours)

Concentrate Delivery	1,800

Concentrate Loaded into Railcars from Ground	1,187

Sinter Produced	2,160

Blast Furnace Charge (of Sinter)	2,160

Rough Lead Produced by Dross Plant	1,260

Refined Lead Produced	888

Process Weight Limits (Quarterly)

	Activity	Process Throughput Limit (tons per calendar quarter)

Sinter to South End Storage	45,000

Fume Handling to Storage on South End	1,170

Sinter Plant Production	169,190

Blast Furnace Sinter Throughput	169,190

Refinery Production	50,000*

*After April 7, 2008, refinery production may be increased by 5,000 tons
per quarter, up 

to a maximum of 62,500 tons.  If the quarterly standard is exceeded,
refinery production 

will be reduced by 5% of the actual production during the exceeding
quarter.  

Required reductions will stop if production reaches 35,000 tons.  See
description contained under 6.1 Contingency Measures Schedule, Project
2.



The previous SIP did not include any daily process weight limits.  It
did contain quarterly throughput limits.  Sinter plant production was
previously limited to 283,920 tons per quarter, blast furnace charge was
limited to 114,0005 tons per quarter, and refinery production was
limited to 80,808 tons per quarter.  Although the blast furnace
throughput limit has been increased, the substantial decrease in sinter
plant production as well as a maximum limit of 62,500 tons per quarter
refined lead produced (down from 80,808 tons per quarter), together with
the other SIP controls and modeling refinements, resulted in the state
modeling attainment.

Each concentrate delivery truck is weighed before and after delivery,
and a moisture content sample taken.  The amount of delivered
concentrate minus moisture is calculated for each truck and recorded.

Front end loaders, used to load concentrate into railcars from the
ground, deposit fume to storage on the south end of the plant, and move
sinter to south end storage, have built-in scales.  Each load is
recorded on a daily log sheet, and transferred to Doe Run’s electronic
database.

Sinter production is calculated using the rate of processing and the
amount of input material.  Blast furnace charge/blast furnace sinter
throughput is recorded from the blast furnace scale belt.  The scale
belt weighs the charge before it is processed in the blast furnace.  The
amount of rough lead produced is calculated from the number of kettles
moved to the dross department and the average kettle content.  Finally,
Doe Run records the amount of finished product from the production
department, equaling the amount of refined lead produced/refinery
production.   

Process weight logs are maintained by Doe Run’s environmental crew. 
Daily/24 hour total throughput records are compiled as: records from the
day shift 5 a.m. – 5 p.m., plus records from the night shift 5 p.m.
– 5 a.m. 

	In addition to these throughput limits, materials handling is further
limited by the following requirements:

Other Materials Handling Quantity Requirements

	Activity	Requirement

Quantity of sinter in stock storage pile	No more than 20,000 tons

Amount of concentrate received directly into railcars on days
concentrate is delivered.	Minimum of 11 concentrate truck loads

Amount of fume unloading to the south end stock area	No more than 13
cars per quarter



Hours of Operation, Moisture Content, and Chemical Stabilization

Handling of material has been limited to the hours specified below.

Hours of Operation

	Activity	Hours Allowed

Concentrate Unloading	6am - 10pm

Concentrate Loading to Railcar	6am - 10pm

Concentrate Rail Tipping	6am - 10pm

Fume Unloading	noon - 6pm

South End Fume Loading into Concentrate Railcars	6am - 10pm



Doe Run’s environmental crew logs will record hours of operation.  

Additional moisture content and chemical stabilization requirements for
storage piles and materials are also included in the Consent Judgment
and Work Practices Manual.  Maintaining moisture and chemically
stabilizing inactive piles further minimizes potential fugitive
emissions from wind erosion and disruption during handling.

Moisture Content Requirements

	Activity	Daily Average Moisture Content Requirement

Concentrate Delivery	6%

Fume to Railcars at South End Stock Area	8%

Fume to Storage Pile at South End Stock Area	8%

Wetting of unpaved Haul Road, North End of Blast Furnace Building to
Refinery Dock	350 gallons applied once every 4 hours during any day or
portion of the day the road is used for hauling slag to storage, minimum
of 1000 gallons per day.

Chemical Stabilization

	Area	Frequency

Unpaved Haul Road, North End of Blast Furnace Building to Refinery Dock
Within 72 hours of completing slag hauling activity

Inactive Concentrate Storage Piles	Once per month

Inactive portions of Fume Storage Pile	Once per month



	Samples must be taken from each concentrate delivery truck, augured
fume, and any fume storage pile being worked on.  These samples are sent
to Doe Run’s laboratory for moisture content determination.  The lab
maintains average daily moisture content records.  A monthly log of
storage pile treatments and records of haul road treatments are
maintained by the Doe Run environmental crew.

Emission Limits

As modeled in the attainment demonstration, Doe Run is required to meet
the following stack emission limits.  

Stack Emissions Limits

	Stack Name	Emission Limitation (pounds per 24 hours)

Main Stack	794

Number 7 & 9 Baghouse Stack	56.6

Number 8 Baghouse Stack	8.2



	Consent Judgment section B3 describes dates for stack testing and
subsequent required actions based on the findings of the specified stack
tests.  The Consent Judgment requires that the main stack be tested by
June 30, 2008, and again by June 30, 2009.  If both tests demonstrate
compliance with the above limit, the main stack’s testing frequency is
set at once every 24 months.  Should any tests exceed the limit, a
continuous or near-continuous monitor must be installed on the stack. 
If EPA has not approved such a monitor, testing frequency increases to
once every calendar quarter until four consecutive tests demonstrate
compliance, at which time tests shall be conducted annually.

	The main stack was tested in June 2008 and tested below the limit.

	The Consent Judgment requires that Number 7 & 9 baghouse stack and
Number 8 baghouse stack be tested every calendar quarter, starting the
second quarter of 2008 through the first quarter of 2009.  For each
stack, if the average of its four tests is greater than 80 percent of
the limits listed above, but less than the limit, the testing frequency
alters to testing every other calendar quarter.  If the average is less
than 80 percent of the limit, testing may be reduced to one test
annually.  If any subsequent annual test is greater than 80 percent, but
less than the limit, the testing frequency must increase to every other
quarter.  Should any test exceed the applicable limit, a continuous or
near-continuous monitor must be installed on the stack.  If EPA has not
approved such a monitor, Doe Run must test every calendar quarter until
four consecutive tests demonstrate compliance.  If the average of the
four consecutive tests is less than 80 percent of the limit, stack tests
may be conducted once annually.  If the average of the four tests is
greater than 80 percent but less than the limit, tests must be conducted
no less than once every other calendar quarter.

	Number 7 & 9 baghouse stack and Number 8 baghouse stack were tested in
June 2008.  Both stacks had emission rates less than 80% of their
respective limits. 

	Methods for stack testing are outlined in Missouri rule 10 CSR
10-6.120, Restriction of Emissions of Lead from Specific Lead
Smelter-Refinery Installations, included in this TSD as Attachment 10.

Ventilation Requirements

In order to improve capture of emissions inside of buildings, the
current SIP, as well as previous SIPs, have included specific
ventilation measures.  Unique to this SIP is a control requiring Doe Run
to complete a ventilation study to determine appropriate ventilation
rates for the Sinter Building, Blast Furnace Building, and
Dross/Refinery Building.  Upon submission and approval from MDNR, the
conditions identified by the study (e.g., minimum fan amperages, minimum
flow rates) will become enforceable under the Work Practices Manual
and/or Consent Judgment.  These enforceable conditions must also be
submitted to EPA for inclusion in the federally-approved SIP.  

On August 14, 2008, MDNR sent a letter to Doe Run further outlining
ventilation study requirements.  The letter is included as Attachment 14
of this technical support document.  In the letter, MDNR requires Doe
Run to establish minimum fan amperages or flow rates for systems without
such specifications (see the “Consent Judgment Ventilation
Requirements” table below).  Doe Run is also required to conduct flow
testing campaigns to demonstrate that the ventilation flow rates in the
Sinter Building, Blast Furnace Building, and Refinery Building are
adequate to create an inflow of 200 feet per minute (fpm) or greater at
all building openings.  The face velocity of 200 fpm has been identified
as a critical velocity in the capture of particulates by the American
Society of Heating, Refrigeration, and Air-Conditioning Engineers.  EPA
Method 204, “Criteria For and Verification of a Permanent or Temporary
Total Enclosure,” also requires that “the average facial velocity
(FV) of air through all natural draft openings (NDOs) shall be at least
3,600 m/hr (200 fpm).  The direction of air flow through all NDO’s
shall be into the enclosure.”  The 200 fpm face velocity will be
applied at Herculaneum to verify that ventilation is adequate to ensure
particulate is not escaping out of building openings.  Once the
ventilation flow rates are set, any person-sized door showing inflow
below 200 fpm must be sealed or replaced with a double door chamber
system.  Any equipment door showing inflow below 200 fpm must have heavy
strip curtains installed, and Doe Run must conduct a fluid modeling
study for all three process buildings to determine if additional
ventilation is appropriate.

See section 5.2.4 and Attachment 14 of this technical support document
for further discussion of the ventilation study.

Ventilation requirements are partially described in the following
tables.  Although the sinter wheel tunnel, Number 3, 5, and 6 baghouses
already have minimum ventilation rates, the buildings they service are a
part of the ventilation study, and the flow testing campaigns and 200
fpm requirements apply to these systems as well as to the Number 7, 8
and 9 baghouses.

 

Consent Judgment Ventilation Requirements

Area	Requirement

Sinter Machine wheel tunnel	Maintain a ventilation rate of 15,000 actual
cubic feet per minute

Number 3 Baghouse	Maintain a ventilation rate of 225,000 actual cubic
feet per minute

Number 5 Baghouse	Maintain a ventilation rate of 300,000 actual cubic
feed per minute

Number 6 Baghouse	Maintain a ventilation rate of 50,000 actual cubic
feet per minute

Number 7 Baghouse	Minimum flow rate or fan amperage requirement to be
determined by ventilation study

Number 8 Baghouse	Minimum flow rate or fan amperage requirement to be
determined by ventilation study

Number 9 Baghouse	Minimum flow rate or fan amperage requirement to be
determined by ventilation study



Work Practices Manual Ventilation Requirements

Area	Design Rate (acfm)

New smooth rolls baghouse	11,000

CV-10 Grizzly	9,000

CV-10 / CV-11 / CV-12 vent	8,000

CV-13 and CV-14 vent	6,000

Scale belt vent	11,000

Crow's nest vent	14,000

"D" kettle fluxing vent	12,000

Blast furnace front end vent	25,000

Sinter Machine wheel tunnel vent.	15,000



	Doe Run is required to maintain a file recording fan amperages,
outages, and ventilation flow rate data as necessary to demonstrate
appropriate process and building ventilation rates.  The Work Practices
Manual ventilation areas will be subject to a volume survey each
calendar quarter.  Should the flow rate drop to 75% of the listed flow
rate, the system will be inspected for maintenance requirement.  

7.1.5	Roads

	Roads, both within and outside of the facility grounds, were noted by
the modeling to contribute to the air concentrations.  Doe Run is
required to run sweepers along paved in-plant and paved out-of-plant
roads, and operate and maintain an in-plant sprinkler system to the
specifications below.  Maps of coverage areas are included in the Work
Practices Manual.

In-Plant and Out-of-Plant Road Requirements

Activity	Requirement

Wet sweeper used on paved routes inside the plant	6 hrs per day, Mon-Fri

Regenerative Air Sweeper used on paved streets	6 hrs per day, Mon-Fri
and any other day concentrate is scheduled for delivery 

Sprinkler System covering Haul Routes within the Plant	Continuously
Operated



	The Regenerative Air Sweeper is subject to quarterly tests, to
demonstrate that it is working effectively.  Sweeper use logs are
required by the Work Practices Manual.

Reporting

In addition to daily logs and weekly, monthly, and quarterly reports
required by the Work Practices Manual, the following summary reports and
information collection are required by the Consent Judgment:

Frequency	Information	Associated Activity / Process

Quarterly	Throughput	sinter machine

Quarterly	Throughput	blast furnace

Quarterly	Produced	refined lead

Quarterly	Report	EMS requirements

Quarterly	Report	TEOM filter analysis as required in Consent Judgment
B10.

Record	date, time, findings, corrective actions	baghouse inspections

Record	Occurrence	upset operating condition, material spills

Record	inspection & corrective actions	building siding inspections

Record	average daily moisture content	concentrate, railcar and storage
pile fume

Record	appropriate ventilation rates maintained	fan amperages, outages,
ventilation rate monitoring data

Record	occurrence and corrective actions	kettle failures

Record	ambient monitor data	Continuous particulate samplers operated at
Broad St. and Main St.

record & report	ambient monitor data	Doe Run-operated monitors: Dunklin
High School, Main Street/City Hall, and Broad Street.

Record	meteorological data	meteorological information collection (see
Attachment 11)



Sanctions and Penalties

The Consent Judgment contains provisions for stipulated penalties and
sanctions should Doe Run fail to comply with provisions of the Consent
Judgment or Work Practices Manual.  EPA is not bound by the state’s
Consent Judgment penalties, and would enforce against violations of
these documents under section 113 of the CAA, rather than the Consent
Judgment, if EPA approves the Consent Judgment and Work Practices Manual
into the SIP. 

8.0	Process for Making Substitutions, Alterations, or Additions to the
SIP

There are a number of important requirements within this SIP revision
which are either currently unspecified (e.g., the parameters for
maintaining ventilation flow in the various buildings – condition A.20
of the Consent Judgment) or are subject to change upon approval from
MDNR (e.g., the concentrate delivery system – condition A.1, and the
substitute contingency measures – condition C.5).  While approving
these provisions, EPA notes that any subsequent requirements developed
pursuant to the provisions that could impact emission levels or required
monitoring must be submitted to EPA for approval as revisions to the
SIP, thereby also triggering opportunity for public input through the
state and federal processes.

9.0	Continual Improvement

	As part of the 2007 Consent Judgment, MDNR has included a number of
requirements beyond those required by the control strategy.  These
requirements were developed to provide additional assurances that Doe
Run will maintain emissions at or decrease emissions below those modeled
in the control strategy attainment demonstration.  Continual improvement
strategies include filter analysis requirements, an environmental
management system, and the “technological study for fugitive dust
control” contingency measure.  The Consent Judgment also contains
meteorological monitoring requirements and language concerning
alteration of Doe Run’s fence line.  A brief description of these
strategies follows.  For more detailed descriptions, please see the 2007
Consent Judgment.

Doe Run is required to submit to MDNR a list of sampling locations where
samples will be taken for updated chemical analyses (fingerprinting). 
These and other fingerprints will then be compared against filters
collected from the Broad Street and Main Street/City Hall continuous,
tapered element oscillating microbalance (TEOM) monitors in order to
identify specific emission sources.  Examples of emission sources
include blast furnace fugitives, in-plant roads, or fume handling. 
Filters from the Doe Run-operated Main Street/City Hall and Broad Street
monitors will be analyzed on any day that exceeds 5.0 µg/m3, and on any
day that exceeds 1.5 µg/m3 and falls on the every sixth day national
monitoring schedule.  Doe Run will conduct an analysis including the
filter analysis, identification of malfunctions, process upsets, process
throughputs, wind speed and direction data.  This analysis will be
submitted to MDNR on a quarterly basis.

The Consent Judgment requires Doe Run to implement an environment
management system (EMS).  In addition to other features, the EMS must
establish and require implementation and maintenance of procedures for
dealing with actual and potential nonconformities, and for taking action
to correct these nonconformities.  Doe Run must provide quarterly
reports to MDNR describing any nonconformities and preventive or
corrective actions related to a quarterly exceedance of the 1.5 µg/m3,
a daily average above 12 µg/m3, or any event increasing emissions above
levels related to normal operating processes.  Within 60 days of receipt
of the report, MDNR will notify Doe Run whether additional actions are
required.  If, for example, additional analysis indicates emissions from
roads are contributing significantly to days with

high-monitored lead concentrations, MDNR and EPA will engage Doe Run in
discussions to increase road controls and reduce emissions from
roadways.  Any resulting requirements will become part of the Consent
Judgment and fully enforceable.

Doe Run is also required to develop a work plan to study best practices
and technologies at three or more other smelting facilities and other
facilities with fugitive emissions control challenges.  The work plan
must be approved by MDNR, and the study and associated report must be
completed within 180 days of notification that a quarter has exceeded
90% of the 

1.5 µg/m3 standard (1.35 µg/m3).  The report must identify the best
practices and best available technologies found, identify those Doe Run
finds technically feasible and cost-effective for inclusion as
contingency measures, quantify associated emissions reductions, and
provide a time frame for implementation of each.  Sixty days after MDNR
receives the report, MDNR will notify Doe Run if the report assessment
is acceptable.  Agreed-upon practices and technologies, and associated
implementation timelines, will become contingency measures and
enforceable conditions of the Consent Judgment.  As discussed in the
Contingency Measures section of this document, EPA supports execution of
the Technology and Fugitive Dust Study, but does not consider it a
contingency measure under section 172(c)(9) as it does not contain
specific requirements or defined deadlines for implementation.

The Consent Judgment also includes meteorological monitoring
requirements.  These requirements were established to ensure that
appropriate meteorological parameters are measured, that the information
is recorded and stored, and that proper quality assurance and quality
control occurs.  See Attachment 11, Meteorological Data Collection
Requirements, for a listing of these requirements.

	Alteration of Doe Run’s fence line is also addressed in the Consent
Judgment.  The dashed yellow line shown in Attachment 13 depicts the
fence line that was used for this SIP development.  The Consent Judgment
specifies that, if Doe Run moves the fence line, Doe Run cannot relocate
any existing processes or construct new lead emission sources in the
area between the existing (SIP-modeled) fence line and the new fence
line.  Any expansion of the fence line will not affect the control
requirements of the SIP.  Doe Run is also required to continue
monitoring at Dunklin High School, Main Street/City Hall, and Broad
Street regardless of relationship to the fence line.  Data from monitors
located inside of Doe Run’s fence line, and thus not located in
ambient air, will be used for informational purposes and not for
determining compliance with the NAAQS for lead.

On August 25, 2008, Doe Run completed a new fence line to the west of
the facility (see Attachment 13).  New monitors have been installed to
determine lead concentrations near the fence line (see Attachment 13). 
The Broad Street monitor is inside this new fence line.  As outlined
above, Doe Run will continue reporting data from Broad Street but the
values will not be used for determining compliance with the NAAQS for
lead.  MDNR and EPA continue to evaluate the adequacy of the monitoring
network as changes occur in emission sources, the town of Herculaneum,
the fence line, and the lead NAAQS itself.

	The additional measures described above are included in the SIP in
order to provide further assurances that the area will attain and
maintain the 1.5 µg/m3 standard.  EPA encourages implementation of any
measures which would reduce lead emissions and supports implementation
of the continual improvement activities.  However, because the continual
improvement activities do not contain specific requirements and/or
associated deadlines for implementation, EPA does not consider them to
be part of the control strategy or contingency measures for the 2007
SIP.

Attachments



List of Attachments

Attachment 1: 	Map of Herculaneum City Boundaries.

Attachment 2: 	Map of Herculaneum Monitors, 1982 – 2005.

Attachment 3: 	Source Chart, 2002 SIP versus 2007 SIP.

Attachment 4: 	Process Flow Diagram Without Source Numbers.

Attachment 5: 	Process Flow Diagram With Source Numbers.

Attachment 6:	Individual Day Concentrations: Monitors versus Model.

Attachment 7a:	2005 Filter Analysis Laboratory Results

Attachment 7b:	2005 Filter Analysis Source Profile Key

Attachment 8a:	Source Contribution Analysis: Modeled versus Monitored. 
As provided by MDNR.

Attachment 8b:	Source Contribution Analysis: Modeled versus Monitored. 
As clarified by EPA.

Attachment 9: 	Quarterly Average Concentrations: Modeled versus
Monitored

Attachment 10: 	Missouri Rule 10 CSR 10-6.120, Restriction of Emissions
of Lead from Specific Lead Smelter-Refinery Installations

Attachment 11:	Meteorological Data Collection Requirements

Attachment 12:	Clarification Letter from MDNR to EPA, March 19, 2008.

Attachment 13:	Map of Fence Line and Current Monitors.

Attachment 14:	Ventilation Study Commitment Letter from MDNR to EPA,
August 18, 2008

 The December 7, 2000 document, “2000 Revision of the State
Implementation Plan for Lead for the Doe Run Resources Corporation
Primary Lead Smelter Herculaneum, Missouri,” pg 21, incorrectly states
Broad St. as one of the three monitors used to calculate background. 
The September 15, 2000 memorandum “Doe Run- Herculaneum State
Implementation Plan Dispersion Modeling Review” from Jeffry Bennett,
P.E. to John Rustige, P.E., pg 9, accurately discusses how the
background calculations were completed in 2000 using the Ursaline,
Bluff, and High School monitors.  

 The Consent Judgment and Work Practices Manual both require analysis of
filters from the Doe-Run operated Main Street/City Hall and Broad Street
monitors.  In Attachment 12, “Clarification Letter from MDNR to EPA,
March 19, 2008,” page 5, MDNR states that, “Doe Run is required to
conduct an analysis every day that an individual monitor records a value
of greater than five micrograms per cubic meter or when 1.5 micrograms
per cubic meter or more is recorded on days that are part of the every
sixth day national monitoring schedule.”  While Doe Run may analyze
filters from additional monitors, the enforceable Consent Judgment and
Work Practices Manual only require Doe Run to analyze filters from the
Doe-Run operated Main Street/City Hall and Broad Street monitors.

Attachment 14.  Ventilation Study Commitment Letter from MDNR to EPA,
August 18, 2008.

