EPA-450/3-75-049
JANUARY 1975
ESTABLISHMENT
OF A NON-EPA USER SYSTEM
FOR STATE IMPLEMENTATION
PLANS
li.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
-------
EPA-450/3-75-049
ESTABLISHMENT
OF A NON-EPA USER SYSTEM
FOR STATE IMPLEMENTATION
PLANS
by
PEDCo-Environmental Specialists, Inc.
Suite 13, Atkinson Square
Cincinnati, Ohio 45246
Contract No. 68-02-1001
Task No. 4
EPA Project Officer: Gerald J . Nehls
Prepared for
ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Air Quality Planning and Standards
Research Triangle Park, N. C. 27711
January 1975
-------
This report is issued by the Environmental Protection Agency to report
technical data of interest to a limited number of readers. Copies are
available free of charge to Federal employees, current contractors and
grantees, and nonprofit organizations-as supplies permit-from the Air
Pollution Technical Information Center, Environmental Protection Agency,
Research Triangle Park, North Carolina 27711; or, for a fee, from the
National Technical Information Service, 5285 Port Royal Road, Springfield,
Virginia 22161.
This report was furnished to the Environmental Protection Agency by
PEDCo-Environmental Specialists, Inc., Cincinnati, Ohio 45246 , in
fulfillment of Contract No. 68-02-1001. The contents of this report are
reproduced herein as received from PED.Co-Environmental Specialists,
Inc. The opinions, findings, and conclusions expressed are those of
the author and not necessarily those of the Environmental Protection
Agency. Mention of company or product names is not to be considered
as an endorsement by the Environmental Protection Agency.
Publication No, EPA-450/3-75-049
11
-------
ACKNOWLEDGMENTS
Task I of this report was prepared by PEDCo-Environmental Specialists
under the direction of Mr. Charles E. Zimmer. Mr. David W. Arraentrout served
as Project Manager. We would like to express our appreciation to the
following organizations for participating in this survey and for providing
technical information:
0 Arizona Division of Air Pollution Control
0 Argonne National Laboratory
0 Cleveland, Ohio, Division of Air Pollution Control
0 Connecticut Air Compliance Section
0 Indiana Division of Air Pollution Control
0 Lawrence Livermore Laboratory
0 Maryland Bureau of Air Quality Control
0 New Mexico Air Quality Section
0 Ohio Environmental Protection Agency
0 Washington State Department of Ecology
0 District of Columbia Bureau of Air and Water Pollution
Control
0 United States Department of Housing and Urban
Development
Mr. Gerald Nehls, National Air Data Branch, served as Project Officer
for the Environmental Protection Agency. We wish to thank Mr. Nehls for
his valuable assistance.
Task II of this report was prepared for PEDCo-Environmental Specialists
by PRC Systems Sciences Company under Subcontract No. 63-74. We wish to
express our appreciation for the contributions of Mr. Charles Bloomquist,
PRC Systems Sciences Company, who conducted the surveys to gather cost data,
made the cost projections, and wrote the report. The report was subsequently
edited by PEDCo-Environmental Specialists.
iii
-------
ABSTRACT
This report presents the results of a survey conducted
among selected state air pollution control agencies to deter-
mine their current practices and projected needs related to
accessing U.S. Environmental Protection Agency data bases.
Alternative methods for allowing non-EPA users to use the
data bases were introduced. A preliminary cost survey was
conducted for a projected method for allowing state agencies
to have direct access to the data bases.
IV
-------
PREFACE
This survey was conducted to identify requirements and quantify
costs for direct state access to air data bases. Nine state agencies
were surveyed and results were extrapolated to cover the nation by
the "State Air Data Information Survey" EPA-450/3-74-001.
The states to be surveyed were determined by examining their
data handling capability to have a mix of manual and computerized
systems. The regional office AEROS contacts also participated in
the selection.
The requirements for direct access to air data bases were not
overwhelming. Some states would use any system which is cost-
effective. Others would not use a system even if it were cost
effective. The costs of providing states with direct access are
estimated in this report. No attempt was made to estimate state usage
to fulfill their own requests or requirements other than for federal
reporting.
As a result of this survey the National Air Data Branch will
concentrate on reducing turnaround times for updating data and
retrieving data and implementing the Comprehensive Data Handling
System for interested state agencies. We will not plan to provide
direct state terminal access to the AEROS System.
An EPA user survey, soon to he released will identify user
requirements and requested enhancements to the /*EROS systems. These
will be developed as resources permit.
Gerald Nehls
-------
TABLE OF CONTENTS
TASK I. SURVEY OF STATE AND LOCAL AGENCIES Page
1.0 EXECUTIVE SUMMARY 1
1.1 Background 1
1.2 Approach 1
1.3 Survey Results 2
1.3.1 State Agency Applications for' 2
EPA Data Bases
1.3.2 Potential Users of EPA Data Bases 2
1.3.3 Alternative Systems for Accessing 3
NEDS/SAROAD Data
2.0 TECHNICAL REPORT 5
2.1 Introduction 5
2.1.1 Background 5
2.1.2 Purpose and Scope of the Survey 5
2.2 Survey Procedures 6
2.2.1 Background Preparation 6
2.2.2 -Content of the Survey 7
2.2.3 Organizations Surveyed 9
2.3 Survey Results 9
2.3.1 Interpretive Restrictions 9
2.3.2 Emissions Data 11
2.3.3 Air Quality Data 14
2.3.4 Compliance and Enforcement Data 16
2.3.5 SIP Regulation Data 16
2.3.6 Selected Non-EPA Data Tapes 17
3.0 CONCLUSIONS 18
3.1 User Classification 18
3.1.1 Basic Data Handling Systems 18
3.1.2 Intermediate Data Handling Systems 18
3.1.3 Advanced Data Handling Systems 19
3.2 Candidate Users for Alternative Systems 19
3.3 Projected Activity for Alternative Systems 20
3.4 Alternative Systems for NEDS/SAROAD Access 25
VI
-------
TASK II. COST ANALYSIS
1.0 EXECUTIVE SUMMARY • 28
1.1 System Overview . . 28
1.2 conclusions 30
2.0 TECHNICAL REPORT 33
2.1 System Overview 34
2.2 Candidate' Network 38
2.2.1 Assumptions 38
2.2.2 Network Evaluation Input 41
2.2.3 Cost Review 61
2.2.4 Evaluation of Cost Survey Results 63
Technical Report Data and Abstract 66
LIST OF FIGURES
Figure Page
1. Research Triangle Compute:: Center Input • 58
Vll
-------
LIST OF TABLES
Table - Page
1, Anticipated Use of NEDS/SAROAD 3
2. NEDS Reports for Meeting Minimum 12
Emissions Reporting Requirements of
State Agencies
3. SAROAD Reports for Meeting Minimum ' 15
Air Quality Reporting Requirements of
State Agencies
4. Estimates of Computer Status for State 20
E.I. (Emission Inventory) and A.Q.
(Air Quality) Data Systems
5, Minimum Summary Frequencies for E.I. 21
and A.Q. Data in State Agencies
6. Annual Range of Inquiries for E.I. and 21
A.Q. Data Reports
7. Minimum E.I. Report Volume for a Range of 22
7-29 States
8, Minimum A.Q. Report Volume for a Range of 22
3-18 States
9. Update Frequencies for Computerized E.I. 23
Data Handling Systems
10. Update Frequencies for Computerized A.Q. 23
Data Handling Systems
11. Volume of Update Data for E.I. Systems 24
12. Reporting Stations Volume of Update Data 25
for A.Q. Systems
13. Projected Stations Volume of Update Data 25
for A.Q. Systems
14. Network Benefit Comparison ' 31
15. Network First-Year Cost Comparison . 32
16. States with Manual Data Systems 35
17. Annual Data Volumes 40
18. Typical Report Processing Times 42
19- Terminal Network Cost Estimates 43
20. BCS Terminal Network Cost Estimates 45
21- EPA/OSI Terminal Network Cost Estimates 50
2~- GSA INFONET Terminal Network Cost Estimates 53
25- Length of Typical GSA INFONET Telephone Toll Call 54
-•*• RTCC Terminal Network Cost Estimates 59
-5- Typical Terminal Lease Charges 60
26. Typical CPU Time for CDHS Installation on an 60
IBM 360/50
-~. Network First-Year Cost Comparison 62
viii
-------
TASK I. SURVEY OF STATE AND LOCAL AGENCIES
1.0 EXECUTIVE SUMMARY
1.1 BACKGROUND
The Environmental Protection Agency (EPA) has developed
extensive computer-oriented systems for the storage and
retrieval of data on air quality measurements and air contam-
inant emissions. The EPA system for handling air quality
data is the Storage and Retrieval of Aerometric Data (SAROAD)
and the system for handling emissions data is the National
Emissions Data System (NEDS). EPA is developing systems to
create national data bases for state and local emission reg-
ulations, source tests, and hazardous pollutant emissions.
Each state and local control agency has access to software
included in EPA's Comprehensive Data Handling System (CDHS),
which gives state agencies the ability to establish their own
computerized air pollution data bases.
The purposes of the work performed under this contract
were to:
0 Survey state and local air pollution control
agencies and selected Federal organizations to
determine the extent to which their data re-
quirements are fulfilled by the various EPA
data bases.
0 In the event that existing EPA data systems
do not fulfill state and local agency require-
ments, perform a cost/effectiveness analysis
of alternative data systems required to meet
these needs.
1.2 APPROACH
A two-man interview team visited selected state and
-------
Federal agencies and reviewed the data systems now being used.
Personnel of each agency were questioned regarding the
reports and data analyses they prepare and the ways in which
management uses this information for decision making and
planning. To assist the interviewers in maintaining consis-
tency, a survey questionnaire was developed. Agencies to be
included in the study were selected in consort with NEDS/
SAROAD coordinators in EPA Regional Offices.
1.3 SURVEY RESULTS
The results of this survey are summarized here in terms
of current and projected use of the NEDS and SAROAD data bases
by state agencies. These results provide the basis for the
cost/effectiveness study to be performed in Task II. Informa-
tion obtained from local agencies and from the Federal research
and planning organizations is not included in these results and
projections but is analyzed separately in Attachment 1 (pro-
vided to the Project Officer).
1.3.1 State Agency Applications for EPA Data Bases
All of the agencies interviewed were familiar with NEDS
and SAROAD, but none were aware of the other data bases. None
of the agencies were familiar with all the report and retrieval
options of NEDS and SAROAD. All of the states currently using
manual or marginally computerized systems for emissions and air
quality data could benefit from increased use of NEDS and
SAROAD.
The agencies indicated interest in the other data bases
discussed, but did not identify specific applications or pro-
jected access frequencies. All agencies indicated that access
to these other EPA data bases by batch request through EPA
Regional Offices would be adequate.
1.3.2 Potential Users of EPA Data Bases
State agencies interviewed during this survey can be
classified as basic, intermediate, or advanced in terms of
data-handling systems. The basic category includes states
-------
with totally manual systems. The intermediate category in-
cludes states having some computerization of emissions data
or air quality data, but usually not both. The advanced
category includes states with both types of data computer-
ized.
The potential users of any alternative system for access-
ing NEDS/SAROAD data are primarily those states classified as
basic or intermediate. The numbers of potential users are
expressed as ranges, rather than absolutes. Based on the
results of EPA's "State Air Data Information Survey," (EPA-
450/3-74-001, January 1974) ranges have been calculated for
potential users of both NEDS and SAROAD. NEDS users would be
anticipated to include from 7 to 29 states. SAROAD users
would be anticipated to include from 3 to 18 states. Users
of NEDS cannot necessarily be anticipated to be users of
SAROAD. Anticipated ranges of output for NEDS and SAROAD are
summarized in Table 1.
Table 1. ANTICIPATED USE OF NEDS/SAROAD
Range of user states
Average output volume,
pages/year/state
Total volume range, pages
Total annual update
volume
NEDS
7-29
2,300
18,000-65,000
202-842
(Point sources)
SAROAD
3-18
1,900
56,000-34,000
484,000-1,356,000
(Data values)
1.3.3 Alternative Systems for Accessing NEDS/SAROAD Data
Several alternative systems could provide the states
classified as basic and intermediate with access to NEDS and
SAROAD data. Any system used would require a capability for
the users to submit directly to the computer rather than
through the EPA Regional Offices, as is the current practice.
Four proposed systems provide, the basis for the cost analyses
to be performed in Task II.
-------
Batch Processing Through EPA Regional Offices -
This is the mechanism currently used by states re-
quiring NEDS/SAROAD reports. Any state not using
this option would meet their data handling require-
ments by installing CDHS subsystems or by developing
systems in-house.
Access to NEDS/SAROAD Via Remote Terminal - With this
system data from both NEDS and SAROAD would be avail-
able from CDHS implemented on a central computer.
Access to NEDS Via Remote Terminal - With this system
NEDS data v/ould be accessible from the Emissions
Inventory Subsystem (EIS) of CDHS implemented on a
central computer. States with requirements for air
quality data would meet their requirements by re-
questing SAROAD reports through the EPA Regional
Offices., by developing their own systems, or by im-
plementing the Air Quality Data Handling Subsystem
(AQDHS) of CDHS on their own computers.
Access to SAROAD Via Remote Terminal - With this
system, SAROAD data would be accessible from AQDHS
implemented on a central computer. States with re-
quirements for emissions data would request NEDS
reports through the EPA Regional Offices, or they
could develop their own systems or implement EIS on
their own computers.
-------
2.0 TECHNICAL REPORT
2.1 INTRODUCTION
2.1.1 Background
The Storage and Retrieval of Aerometric Data System
(SAROAD) and the National Emission Data System (NEDS) were
developed by EPA to provide information on measurements of
air contaminants in the atmosphere and emissions of air
contaminants from area and point sources. The data bases
provided by these systems are used by state air pollution
control agencies and by EPA for testing of control strate-
gies, for evaluation of State Implementation Plans (SIP's),
and for SIP report and follow-up activities required by
Federal law. They are also being used to provide air quality
and emissions data needed for designations of Air Quality
Maintenance Areas.
EPA is currently developing a computer-oriented storage
and retrieval system to handle the information on air quality
regulations contained in the SIP's. Other associated data
bases are also available through or are being developed by
EPA. Output from SAROAD, NEDS, the SIP's, and the other EPA
data bases are available for use by state and local govern-
ments, research groups (universities, institutes, etc.),
regional planning groups, and industrial-commercial organiza-
tions. The EPA data bases are currently accessible through
requests made to EPA Regional Offices or made directly to
EPA's National Air Data Branch in North Carolina.
2.1.2 Purpose and Scope of the Survey
This survey was conducted to determine to what extent
-------
the EPA systems maintained by the National Air Data Branch
currently satisfy the data storage and retrieval requirements
of non-EPA users of the systems. To make this determination,
personal interviews were conducted to obtain answers to the
following questions:
0 What are the retrieval requirements, volumes
and request frequencies for information contained
in each system?
0 What are the requirements for using information
from multiple data bases to perform an analysis?
0 What are the query requirements of the user
(as opposed to long summaries and analyses)?
0 What requirements (if any) does the user have for
direct access to various data bases and software?
0 How does the user believe EPA should allow him to
access the information and how does he believe this
should be implemented?
The major emphasis of the survey was placed on determin-
ing requirements of selected state air pollution control
agencies. Recommendations for performing cost analyses for
development of additional or alternative EPA systems are
based primarily on the results of these state surveys.
2.2 SURVEY PROCEDURES
2.2.1 Background Preparation
The first step in preparation for the survey was to re-
view the following air data bases and the data handling
systems which are currently available through EPA:
0 NEDS - National Emission Data System
0 SAROAD - Storage and Retrieval of Aerometric Data
0 SOTDAT - Source Test Data System
0 EIS - Emission Inventory Subsystem (CDHS)
0 AQDHS - Air Quality Data Handling Subsystem (CDHS)
0 DQIS - Data Quality Information System
0 CDS - Compliance Data System
0 EMS - Enforcement Management Subsystem (CDHS)
0 HAPEMS - Hazardous Air Pollutants Enforcement
Management System
-------
0 HATREMS - Hazardous and Trace Emissions System
0 SIP Rules and Regulations
0 Non-EPA Selected Tapes (Bureau of Census, GSA
Property, Form FPC-67, and Federal Facilities Files)
To provide guidance in determining the form and content
of the survey format, PEDCo reviewed the "Survey of EPA User
Interest for Proposed State Implementation Plan Automated
Information System" (EPA-450/3-73-011). The format of the
formal survey questionnaire was designed to be compatible with
the format of the EPA user survey.
2.2.2 Content of the Survey
The PEDCo survey, known as the Non-EPA Users Survey, was
designed to meet several criteria:
0 Provide a multipurpose format suitable for obtaining
information on the air data requirements of organiza-
tions with diverse applications and needs.
0 Obtain answers to the five questions outlined in
Section 2.1.
0 Provide compatibility with the completed "Survey of
EPA User Interest for Proposed State Implementation
Plan Automated Information System."
0 Enable interviewers to spend minimal time discussing
areas of little concern to the user or potential user,
concentrating, rather, on areas of maximum interest.
0 Relate current and projected data requirements and
data handling practices to the data bases and data
systems currently available through EPA or being devel-
oped.
Attachment 2 (provided to the Project Officer) presents a
sample questionnaire used in this survey. The questionnaire
is divided into five major sections by data type:
0 Emissions Data
NEDS
SOTDAT
EIS
HATREMS (not shown on the questionnaire)
7
-------
° Air Quality Data
SAROAD
AQDHS
DQIS
0 Compliance and Enforcement Data
CDS
EMS
HAP EMS
0 SIP Regulation Data
0 Non-EPA Data
Bureau of Census Tapes
GSA Property Files
Form FPC 67
Federal Facility Files
INFONET models (not shown on questionnaire)
Each of these major sections of the questionnaire begins
with a detailed query regarding the agency's current require-
ments for various formats, turnaround times, mechanisms for
query, and the like. This general, user-oriented inquiry is
then followed by questions specific to the EPA data bases and
systems related to the major data category under discussion.
inquiry concerning the specific EPA systems was pursued only
if the agency indicated need for or interest in those applica-
tions. VThere appropriate, sample output forms for NEDS and
SAROAD were presented and discussed.
Only data bases of concern to a specific organization
were discussed in detail. Sample output formats for NEDS and
SAROAD were presented and discussed to determine how they
related to the organization's requirements. Several questions
provided information about the users and their data require-
ments, current formats, current operating mechanisms, and the
applicability of existing EPA data bases and systems. Infor-
mation concerning query requirements, request volume, desired
formats, and desired mechanisms for accessing data bases was
directed toward evaluating the requirements for accessing the
-------
data banks through remote terminals. Moreover, this line of
questioning provided information with which to relate current
requirements to data handling systems now in development but
not yet released by EPA.
2.2.3 Organizations Surveyed
Nine state agencies and three Federal organizations were
included in the survey; these organizations are listed in the
Acknowledgement. The nine state agencies were selected to pro-
vide representation of states operating three types of air data
handling systems:
0 Totally manual data handling systems.
0 Static computer systems developed to handle basic
reporting requirements.
0 Dynamic computer systems developed to handle multiple
reporting requirements.
NEDS/SAROAD coordinators in the EPA Regional Offices provided
recommendations concerning the state agencies to be included
in the survey. As an aid in evaluating the representativeness
of the agencies in terms of their data systems and in project-
ing their needs for direct access to the EPA data bases, PEDCo
used EPA's "State Air Data Information Survey" (EPA-450/3-74-
001). Since the PEDCo survey entailed only a small sample of
agencies, this earlier EPA survey provided the best available
means for extrapolating our results to a national scale.
In arranging for the interviews with state agencies, we
requested that they include persons representing the major
user groups or areas of responsibility (i.e. surveillance mon-
itoring, enforcement, permit issuance). In this way the sur-
vey could encompass experience from the working level as well
as from the management level of an agency.
2.3 SURVEY RESULTS
2.3.1 Interpretive Restrictions
2.3.1.1 Representativeness of the Survey Sample - The survey
sample included nine air pollution control agencies and three
Federal research and planning organizations involved with
environmentally related activities. Information obtained from
-------
the three Federal organizations provided some insight into the
data system requirements of organizations whose function is
other than control. Because no previous surveys of project
requirements or computer capabilities of such organizations
have been reported, we have no basis for extrapolating the
information obtained concerning their requirements to those of
other organizations of that type. Further comments and recom-
mendations concerning these organizations are given in Attach-
ment 1 (given to the Project Officer).
The survey plan anticipated that information obtained
from the state agencies would also represent the requirements
of local control agencies. This is not the case. The local
control agency structures are diverse in terms of both techni-
cal expertise and operational autonomy. Consequently, the
requirements of the state agencies for data storage and re-
trieval, systems- should-not-be interpreted, .as., representing the-
requirements of local agencies. It is emphasized, therefore,
that the conclusions and recommendations of this survey are
based solely on the requirements of state control agencies.
2.3.1.2 Alternative Federal Systems - EPA has developed the
Comprehensive Data Handling System (CDHS) for installation and
operation at the state and local levels. This system incorpor-
ates three subsystems:
0 Emission Inventory Subsystem (EIS)
0 Air Quality Data Handling Subsystem (AQDHS)
0 Enforcement Management Subsystem (EMS)
The subsystems correspond functionally to NEDS, SAROAD, and CDS
(Compliance Data System), respectively. Installation of any or
all of the CDHS subsystems by a state agency could signifi-
cantly reduce any requirement for the agency to access the
corresponding EPA data system. Recommendations resulting from
this survey are based to a great extent on states that have
already installed subsystems of CDHS or have firm commitments
to install them.
10
-------
2.3.1.3 Estimates of Frequency and Volume for Data Retrieval -
The state agencies surveyed indicated that they were more
aware of their requirements for providing input to NEDS and
SAROAD than they were of their options for retrieving reports
from those data bases. Most states were familiar with only
the "NEDS Point Source Listing," the SAROAD "Site Description
Inventory," and the SAROAD "Yearly Report by Quarters." Since
the other reports had not been requested, the frequency of
requests and volume of data were not discussed. The states
that had received the three basic reports indicated that turn-
around time for receiving reports through the EPA Regional
Offices is too long and that the data received are thus not
current enough for their applications. Most states were un-
aware of ways in which NEDS/SAROAD could help them do their
jobs more efficiently.
Since report retrieval estimates were not available from
the agencies, the projected system usage figures in this report
are based primarily on input and retrieval levels that would
allow the agencies to meet their formal reporting requirements
to EPA. Estimates of activity levels generated by the internal
requirements of the agencies were not attempted.
2.3.2 Emissions Data
2.3.2.1 Applications - The major applications identified for
emissions data were for use in air quality modeling and in
designation of AQMA's. Requirements for data on emissions
from specific point sources are usually met by accessing the
state permit file. Many states indicated an interest in area
source data, apparently the weakest portion of each state's
emission inventory. The applications identified for emissions
data did not indicate a need for scheduled accessing of an
emission data system to retrieve specific reports.
2.3.2.2 Retrieval Systems - Three states have computerized
emission inventory systems. Six states have totally manual
emission inventory systems, and two of these have initiated
plans to install EIS. All states indicated that they prefer
11
-------
to contact states directly for any required emissions data
from those states. The frequency of such requests is very low.
Existing systems, even in the states with computer retrieval,
do not meet the query requirements for emissions data. The
query types of requests for emissions data are usually ful-
filled by review of permit files.
Systems for the storage and retrieval of emission data
at the state level receive secondary development considera-
tion. Primary consideration for systems development is given
to air quality data.
2.3.2.3 NEDS Requirements - All state agencies interviewed
were familiar with NEDS. Their familiarity, however, came
primarily from EPA's requirements for states to update NEDS
data as part of their semi-annual reporting. The state agen-
cies had no significant experience with receiving and using
NEDS reports routinely. One state has never requested NEDS
output; the others have received at least one "NEDS Point
Source Listing." None were familiar with the other available
NEDS output formats.
A review of the formal- reports produced by the states in-
dicates that, although their format requirements vary slightly,
the formats available from NEDS are generally compatible with,
or adaptable to, their requirements. They suggested no format
changes or additions. The reports from NEDS that could help
most states meet their formal and informal reporting require-
ments are shown in Table 2, which indicates also the minimum
frequency with which each report would be requested.
Table 2. NEDS REPORTS FOR MEETING MINIMUM
EMISSIONS REPORTING REQUIREMENTS OF STATE AGENCIES
NEDS report
Point source listing
Point source listing (for
selected sources only)
Emission summary
Area source listing
Minimum annual
request frequency
2/state/year
14/state/year
2/county/year
I/county/year
12
-------
The estimates in Table 2 were derived from reviewing the cur-
rent reporting practices of the nine state agencies included in
the survey.
The current selection options for retrieval of each NEDS
report format are adequate for all states interviewed. Special
interest was expressed in the options for accessing the file by
Source Classification Code (SCC), since most applications for
emissions data relate to special studies entailing specific
fuel types, boiler types, processes, or other source, category.
No state in the survey has significant requirements for emis-
sions data from adjacent states.
For most of the states, turnaround time of 1 to 2 weeks
from time of request to receipt of a NEDS report is adequate.
One state requested a 1-day turnaround time. Most states are
discouraged with the current mechanism for batch requests
through the Regional Offices. They indicated that some turn-
around times have been as long as 6 weeks, which they believe
to be unreasonable. An additional problem with the NEDS mech-
anism is that the time lag between submittal of data to EPA and
the actual update of the computer file, which can be as long as
6 months, is intolerable. Most states indicated that they
could envision more uses for NEDS if they could be confident.
that NEDS reports represent current conditions. Turnaround
time represents the most significant objection to accessing
NEDS.
Five states surveyed would be potential users of a remote
terminal system for accessing NEDS: two of these indicating
that their use would be marginal. All potential users ex-
pressed two prerequisites for using a terminal system to access
NEDS.
0 The system should be cost/effective, i.e. it should
be less expensive to operate than EIS or an in-house
system.
0 An interactive, capability must be supplied to allow
the user to input emissions data or to update the
13
-------
data base directly. This capability would eliminate
the objection of the time lag currently encountered
when data are submitted through the Regional Offices.
Request levels for NEDS were not specified, since no state has
made extensive use of the system.
2.3.2.4 Other Data Bases - The other emissions data bases
discussed in the survey were SOTDAT (Source Test Data) and
HATREMS (Hazardous and Trace Emissions System). HATREMS gen-
erated no specific interest. SOTDAT was recognized as a means
of obtaining emissions estimates for specific sources. None
of the agencies interviewed is engaged in extensive factor
development work.
2.3.3 Air Quality Data
2.3.3.1 Applications - The major applications identified for
ambient air quality data were for control strategy testing,
designation of AQMA's, and urban planning (transportation,
land use, etc.). The pollutants of major concern are suspend-
ed particulates, SO , CO, and photochemical oxidants. One
state expressed interest in source-oriented air quality data.
2.3.3.2 Retrieval Systems - Five states have manual air qual-
ity data handling systems. Two states have sophisticated
computer systems to analyze data transmitted directly from
field monitors. Two states have adapted the original version
of AQDHS to meet their requirements. All states prefer to
contact other states directly to obtain any required air qual-
ity data. The frequency of those requests is low.
Systems for storage, analysis, and retrieval of ambient
air quality data receive development priority in all states in
the survey. The agencies were better prepared to discuss their
format requirements for air quality data than for emissions
data. Several states expressed interest in developing cap-
abilities for trends analysis and also in computer graphics
for display of air quality trends data.
2.3.3.3 SAR.OAD Requirements - All state agencies interviewed
were familiar with SAROAD. Experience with using SAROAD and
14
-------
acceptance of SAROAD reporting requirements is more widespread
than for NEDS. SAROAD users have,more confidence in the qual-
ity of the data than they have in the quality of NEDS emis-
sions data. As with NEDS, however, SAROAD users are generally
not familiar with the report formats available from SAROAD.
The formats most familiar are the "Quarterly Frequency Distri-
bution," and the "Yearly Report by Quarters."
A review of the formal reports produced by the states
indicates that summaries of air quality data are published
periodically from monthly to annually. All of the available
SAROAD formats are compatible with or readily adaptable to the
formal report requirements of the states. SAROAD reports that
could help most states meet their formal and informal report-
ing requirements are shown in Table 3, which also indicates
the minimum frequency with which each report would be request-
ed.
Table 3. SAROAD REPORTS FOR MEETING MINIMUM AIR
QUALITY REPORTING REQUIREMENTS OF STATE AGENCIES
SAROAD report
Raw data listing
Quarterly reports
Quarterly frequency distri-
bution
Yearly frequency distribution
Yearly report by quarters
Minumum annual reports
64/sensor
4/sensor
4/sensor
I/sensor
I/sensor
The selection options for retrieval of SAROAD report
formats are adequate.
Most agencies indicated that turnaround time of 1 week
from time of request to receipt of a SAROAD report is adequate.
One state requested 1-day turnaround time. The states ex-
pressed the same concerns with long turnaround for SAROAD pro-
cessing through the Regional Offices as they expressed for NEDS
processing. Lag time between data submittal and computer up-
date is more critical for SAROAD than for NEDS, because more
15
-------
data are submitted more frequently. Moreover, air quality
data have the potential of being used for analyses related to
air alerts.
Seven states indicated that they would be potential users
of a remote terminal system for accessing SAROAD: two of these
would be marginal users. All potential users expressed two
prerequisites for using a terminal system to access SAROAD:
0 The system should be cost/effective, i.e. it should
be less expensive to operate than AQDHS or an in-house
system.
0 An interactive capability must be supplied to allow
the user to input air quality data directly, or to
update the data base directly. This capability would
eliminate the objection of the time lag currently
encountered when data are summarized through the
Regional Offices.
Request levels for SAROAD reports were not specified in the
interviews, since no state has made extensive use of the
system.
2.3.3.4 Other Data Bases - The other data base discussed for
air quality data was DQIS for quality control data. No inter-
est was expressed in this system.
2.3.4 Compliance and Enforcement Data
The three systems discussed for compliance and enforce-
ment data were EMS, CDS, and HAPEMS. No one expressed inter-
est in HAPEMS. Only three states expressed interest in EMS or
CDS, and that interest is only in the passive segments of the
systems. One state has an operational system for tracking
compliance information for hazardous pollutant emission
sources. None of the other states could outline their require-
ments for compliance data.
2.3.5 SIP Regulation Data
Most states commenting on the utility of accessing regu-
lations data indicated that they have been satisfied to request
regulations data directly from the state involved. No states
cited past experience with reviewing regulations from other
states. All states indicated that batch requests through EPA
16
-------
Regional Offices for information in the developmental SIP Reg-
ulations data base would be adequate.
2.3.6 Selected Non-EPA Data Tapes
The selected Non-EPA data tapes discussed were for Fed-
eral Facility Files, FPC Form - 67, Bureau of Census Popula-
tion Data, and GSA Property Files. No interest was expressed
for accessing any of these data bases.
The INFONET system for air quality modeling was also dis-
cussed. Some states indicated a desire to obtain the model
tapes and install them on their own computers. One state
agency has had a formal system presentation and price proposal
from the contracting agency handling INFONET, but indicated
that the system is not cost-justified for them.
17
-------
3.0 CONCLUSIONS
3.1 USER CLASSIFICATION
The state agencies covered in this survey are classified
as basic, intermediate, or advanced in terms of their current
capabilities for^ storage, analysis, and retrieval of air pol-
lution data. Among the states included in the survey, three
are categorized as basic, four as intermediate, and two as
advanced. The important data handling characteristics of
state agencies in each category are summarized in the follow-
ing paragraphs. An extrapolation for categorizing all state
or equivalent agencies is explained in Section 3.2.
3.1.1 Basic Data Handling Systems
Agencies in this category are characterized as having no
operational systems for computerized storage and retrieval of
air pollution data. All calculations, trends analyses, and
report summaries are performed manually. These states have
made no commitments to convert to computer-based systems.
NEDS and SAROAD could fulfill the data handling require-
ments of these agencies. The reporting formats available from
NEDS and SAROAD represent a potentially significant improvement
in the data-handling capabilities of these agencies, if the
agencies will make frequent use of those data bases. So that
states may make optimal use of the data banks, the updating
mechanism should be changed to ensure that the data are more
current than data presently available from NEDS/SAROAD.
3.1.2 Intermediate Data Handling Systems
Agencies in this category are characterized as having one
or more computer software systems operational for handling
emissions data and/or air quality data. These states are
18
-------
generally committed to the concept of total computer system
development for emissions and air quality data. Generally at
least one computer systems analyst is assigned full-time to
the development and operation of internal systems. The mix of
in-house systems and CDHS subsystems appears to be about equal.
Increasing interest is being shown for EIS or AQDHS by states
in which systems for emissions data or air quality data, but
not both, have been developed. This category includes agen-
cies at all stages of systems development.
NEDS and SAROAD report formats generally parallel the
formats required by these agencies. Installation of in-house
computer software systems precludes extensive use of NEDS or
SAROAD, except where NEDS and SAROAD can provide the data
bases for the projected systems. Those state agencies that
have not yet made firm commitments for system development to
handle either air quality data or emissions data represent
potential users of NEDS or SAROAD.
3.1.3 Advanced Data Handling Systems
Agencies in this category have most of their systems for
both emissions and air quality data computerized and opera-
tional. They often use extensive telemetry systems for trans-
mitting data from ambient air sampling networks.
These states have designed and implemented their own report
formats, which may or may not be compatible with those avail-
able for NEDS and SAROAD. The data bases for these systems are
established. Access to NEDS and SAROAD is not an advantage to
these agencies. Any access made would be to fulfill query re-
quirements or to obtain out-of-state data. Requests of these
types are infrequent.
3.2 CANDIDATE USERS FOR ALTERNATIVE SYSTEMS
The major candidates for alternative systems for access-
ing NEDS and SAROAD data are projected to be the state agencies
classified as basic or intermediate. States classified as ad-
vanced in systems development may access NEDS and SAROAD data,
but their requirements would be marginal, since most use only
19
-------
data generated within their own state boundaries. States with
computerized systems for handling air quality data but not for
emissions data might be candidates for a NEDS access system
but not for a SAROAD access system, and vice versa. Conse-
quently the candidate users of NEDS data may not number the
same as the candidate users of SAROAD data.
Table 4 shows current estimates of the status of manual
and computerized systems for handling emissions and air quality
data in 55 state or equivalent agencies. The figures repre-
sent the results from the EPA "State Air Data Information
Survey." The CDHS figures represent states that have made
direct requests to EPA for software to implement either the
Emission Inventory Subsystem (EIS) or the Air Quality Data
Handling Subsystem (AQDHS) of CDHS.
Table 4. ESTIMATES OF COMPUTER STATUS FOR STATE E.I.
(EMISSIONS INVENTORY) AND A.Q. (AIR QUALITY) DATA SYSTEMS
Computer
CDHS
Manual
Total
Emissions data
No.
20
6
29
55
% of total
36
11
53
Air quality data
No.
31
6
18
55
% of total
56
11
33
The states in the manual category of Table 4 are the
potential users of alternative systems for accessing NEDS and
SAROAD data. Projections from the "State Air Data Information
Survey" indicate that at least 7 states will not computerize
emissions data and 3 states will not computerize air quality
data. The potential users of NEDS data would be 7 to 29
agencies, and the potential users of SAROAD data would be 3 to
18 agencies.
3.3 PROJECTED ACTIVITY FOR ALTERNATIVE SYSTEMS
The minimum frequencies with which state agencies summar-
ize emissions data and air quality data are shown in Table 5.
20
-------
Table 5. MINIMUM SUMMARY FREQUENCIES FOR
E.I. AND A.Q. DATA IN STATE AGENCIES
Frequency
Weekly
Monthly
Quarterly
Semi- annually
Annually
E.I.
X
X
X
A.Q.
X
X
X
X
Correlating the frequencies in Table 5 with the numbers
of reports shown in Tables 1 and 2 allows us to project ranges
of annual requests for output among anticipated users of NEDS
and/or SAROAD data. Table 6 shows the range of inquiries that
might be made by all the candidate users of a NEDS/SAROAD
access system.
Table 6. ANNUAL RANGE OF INQUIRIES FOR
E.I. AND A.Q. DATA REPORTS
Frequency
Weekly
Monthly
Quarterly
Semi-annually
Annually
Total per year
E.I. (7-29 states)
84-348
28-116
21-87
133-551
A.Q. (3-18 states)
156-936
36-216
24-144
6-36
222-1332
The projected volumes of output for each NEDS report
needed to meet minimum state reporting requirements are shown
in Table 7. The assumptions are that point source listings are
one page each, and emissions summaries and area source listings
are five pages each. Values for both the emissions summary and
area source listing must be multiplied by the total number of
counties. The average number of counties for all states is
taken to be 70.
The average number of point sources in NEDS is assumed to
be 576/state; this value is based on the EPA report "Status of
21
-------
Table 7. MINIMUM E.I. REPORT VOLUME
FOR A RANGE OF 7-29 STATES
E.I. report
Point source listing
Point source listing for
selected sources
Emission summary
(70 counties/state, 5 pg/county)
Area source listing
(70 counties/state, 5 pg/county)
Total
Volume, pages/year
8,000-33,000
2,900-12,000
4,900-20,000
2,500-10,000
18,000-75,000
the National Emissions (NEDS) and Air Quality (SAROAD) Data
Banks as of November 1973." States with computer systems for
emissions data were eliminated from consideration.
The projected volumes of SAROAD output needed to meet
minimum state reporting requirements are^shown in Table 8,,
which gives values for two ranges. The first range is based
Table 8. MINIMUM A.Q. REPORT VOLUME
FOR A RANGE OF 3-18 STATES
A. Based on reporting stations (Nov. 1973), average 25/state
A.Q. Report Volume/year, pages
Raw data listing 4,800-29,000
Quarterly reports 300-1,800
Quarterly frequency dist. 300-1,800
Yearly frequency dist. 75-450
Yearly report by quarters 75-450
Range: volume/year 5,600-34,000
B. Based on projected stations, average 40/state
A.Q. Report Volume/year, pages
Raw data listing ; 7,700-46,000
Quarterly reports 480-2,900
Quarterly frequency dist. 480-2,900
Yearly frequency dist. 120-720
Yearly report by quarters 120-720
.Range: volume/year 8,900-53,000
22
-------
on the number of reporting stations designated in the "Status
of the National Emissions and Air Quality Data Banks as of
November 1973," and the second range represents the number of
stations projected for each state in its State Implementation
Plan. States with known computerized air quality data systems
have been eliminated from consideration. The figures shown
are based on the assumption that one page of output per sensor
for each report is required, and for the states known to have
manual systems an average number of stations was assumed. The
actual range of volumes should be somewhere between the two
ranges shown.
Tables 9 and 10 show the frequencies with which states
with computerized emissions inventory or air quality systems
update their data bases. Again, the figures are based on the
"State Air Data Information Survey".
Table 9. UPDATE FREQUENCIES FOR
COMPUTERIZED E.I. DATA HANDLING SYSTEMS
(FROM "STATE AIR DATA INFORMATION SURVEY")
Frequency
Weekly
Monthly
Semi-annual ly
Annually
Total
No. of states
8
2
3
3
16
% of total
50
13
19
19
Table 10. UPDATE FREQUENCIES FOR
COMPUTERIZED A.Q. DATA HANDLING SYSTEMS
(FROM "STATE AIR DATA INFORMATION SURVEY")
Frequency
Weekly
Monthly
Quarterly
Total
No. of states
12
14
3
29
% of total
41
48
10
23
-------
The percentage figures in Tables 9 and 10 can be applied to
estimates for the rate of data generation by states with manual
systems to derive an indication of probable data input activi-
ties. For deriving the input volumes for emissions data and
distributing those volumes according to anticipated update fre-
quency, the following formula was used:
No. of sources updated per update interval =
NX 576 x 0.05 x F
I
where:
N = number of potential users (7-29)
576 = average number of emission sources
0.05 = estimated update volume
F = percent/update frequency (from Table 8)
I = update intervals (e.g. weekly = 52/year,
monthly = 12/year, etc.)
The anticipated updates for emissions data on alternative
systems are shown in Table 11.
Table 11. VOLUME OF UPDATE DATA FOR
E.I. SYSTEMS (7-29 STATES)
Frequency
Weekly
Monthly
Semi-annual ly
Annually
Total updates/year
No. of sources
2-8
2-9
18-79
38-158
Annual total
100-417
26-109
38-158
38-158
202-842
For deriving the input volumes for air quality data and
distributing those volumes according to anticipated update
frequency, the following formula was applied:
A.Q. input volume =7 V x S x F
1 I
24
-------
where:
V = values generated/sensor
S = number of sensors/pollutant averaging time
F = percent update frequency (from Table 9)
I = update interval (weekly - 52/year, monthly = 12/year
etc.)
The anticipated updates for air quality data on alternative
systems are shown in Tables 12 and 13.
Table 12. REPORTING STATIONS VOLUME OF
UPDATE DATA FOR A.Q. SYSTEMS (18 STATES)
Frequency
Weekly
Monthly
Quarterly
Total update values/year
No. of
input values
4,000
19,000
12,000
Annual total
208,000
228,000
48,000
484,000
Table 13. PROJECTED STATIONS VOLUME OF
UPDATE DATA FOR A.Q. SYSTEMS (18 STATES)
Frequency
Weekly
Monthly
Quarterly
Total update values/year
No. of
input values
11,000
54,000
34,000
Annual total
572,000
648,000
136,000
1,356,000
3.4 ALTERNATIVE SYSTEMS FOR NEDS/SAROAD ACCESS
Four alternative systems can be considered for allowing
state agencies with manual systems for handling emissions or
air quality data to access NEDS/SAROAD.
1. EPA Regional Office Requests - In this system,
states without computer systems would meet their
minimum data requirements by making batch requests
through the EPA Regional Offices. State agencies
are currently not using this system extensively,
25
-------
and the projected impact on Regional Office manpower
if they were to use this option may be an important
cost consideration. This system would probably not
be used by the states unless an update mechanism to
provide more current output could be implemented.
Access to NEDS/SAROAD Data via Remote Terminal -
With this system NEDS and SAROAD data would be avail-
able from EIS and AQDHS implemented on a central
computer. States would have access to the computer
through terminals installed in the state agency
offices. Capability would be provided for the states
to submit data directly to the system for edit and
update. The EPA Regional Offices could access the
systems quarterly or semi-annually to obtain for
review data submitted during each interval from any
of the states. Several computers could be used for
implementing this system:
0 EPA computer at National Environmental
Research Center, Research Triangle Park,
North Carolina.
0 OSI (Optimal Systems, Inc.) Washington, D. C.
This company is currently a government con-
tractor providing computer services to EPA.
0 GSA INFONET System
0 BCS (Boeing Computer Services, Inc.)
Access to NEDS Data via Remote Terminal - The states
with manual air quality data systems would fulfill
their data handling requirements by making batch re-
quests for SAROAD reports through EPA Regional
Offices, or they would implement AQDHS, or they could
design their own systems. The NEDS users would ac-
cess the data bank through remote terminals to a
central computer on which EIS had been implemented.
The candidate computer systems would be EPA at
Research Triangle Park, OSI, BCS, or GSA INFONET.
26
-------
Access to SAROAD Data via Remote Terminal -
The states with manual emissions data systems
would fulfill their data handling requirements
by making batch requests for NEDS reports
through EPA Regional Offices, or they could
design their own systems. The SAROAD users
would access the data banks through remote
terminals to a central computer on which AQDHS
had been implemented. The candidate computer
systems would be EPA at Research Triangle Park,
OSIf BCS, or GSA INFONET.
27
-------
TASK II. COST ANALYSIS
1.0 EXECUTIVE SUMMARY
This is a cost/benefit analysis of four possible systems
to provide state agencies with computerized data processing
of their air quality and emissions data.
Each of the four alternative systems is a computer net-
work with one central computer and remote terminals at the
participating state agencies. The four networks analyzed are:
(1) the GSA INFONET (Computer Sciences Corp.); (2) the EPA/OSI
system (Optimal Systems, Inc.); (3) a Research Triangle-based
(EPA) Network; and (4) BCS (Boeing Computer Services, Inc.),
an independent commercial network. Each of these networks
is assumed to use the EPA-developed Comprehensive Data Han-
dling System (CDHS), and in particular the Air Quality Data
Handling Subsystem (AQDHS) and the Emissions Inventory Sub-
system (EIS).
1.1 SYSTEM OVERVIEW
There are 55 states and territories in the United States
which are required to collect, use, and submit air quality
and emissions data to the National Air Data Branch (NADB).
Two computerized data banks have been set up at NADB to store
these data. The first is the Storage and Retrieval of Aero-
metric Data (SAROAD) and the second being the National Emis-
sions Data System (NEDS). Many states process data at their
own facilities using their own computers. These states set up
and maintain their own data files, interpret and analyze their
data, and possess a practical means of transmitting their data
to regional and national levels of the EPA for inclusion in
NEDS and SAROAD. Many states, however, still use manual data
processing systems.
28
-------
Any state agency currently without an automated system for
handling air quality and emissions data may want such a capa-
bility and would like it to have the following characteristics.
It must be under the direct control of the state agency insofar
as its own data are concerned. Access to the system for both
data input and retrieval should be via teletype terminal re-
quiring only minimal operator training. Terminals must have
card readers attached. Terminal speeds and turnaround time on
retrievals are not critical. Batch retrieval of lengthy re-
ports mailed from the central facility with overnight turn-
around on the central processor is adequate for virtually all
purposes. Some interactive capability is desirable for editing
data inputs and querying the central facility regarding report
retrievals. A capability to store historical data and various
summaries of the data from the state at the central processor
is required. Data summaries should be updated at least quar-
terly with more frequent updates being desirable. Payment for
system utilization should fall entirely on the user, although
initial setup costs and periodic updating of the official data
base would be the responsibility of the Environmental Protec-
tion Agency (EPA).
The evaluation of the four networks proceeded in the
following manner. First, a standard set of assumptions was
provided to a representative from each potential network in
an informal interview. The inputs required from each candidate
network were similarly defined, and a list was provided to the
various representatives for their response. A fifth interview
was conducted with a terminal supplier to get independent
inputs on this critical aspect of the system, and Mr. L.
Hedgepeth of the EPA was consulted regarding the implementation
costs of CDHS.
The data resulting from these interviews were then trans-
formed to a common base of costs and benefits to enable a
direct comparison to be made. Various tradeoffs were made and
system sensitivities investigated.
29
-------
1.2 CONCLUSIONS
The general benefit of the proposed system is to provide
state agencies with computerized access to and processing of
their air quality and emissions inventory data. This facili-
tates meeting internal reporting requirements and those man-
dated by the EPA. The access and processing is provided by
remote terminals tied into a shared central computer facility.
While each of the four networks vary in precise operating
procedures, costs, and perhaps some other factors, they are all
capable of performing the required function and all are gener-
ally technically equivalent.
With the exception of the Research Triangle Computer
Center (RTCC), all candidates are established nationwide net-
works. All of them have been in business for a considerable
time, and all of them are or have recently been contractors
to agencies of the U.S. Government and to the EPA itself.
The DCS cost estimate is the only one that includes the
capability for invoicing user states directly. The other can-
didate networks could invoice directly, but direct invoicing
would increase the cost of the system. Current government dis-
counts, on which the GSA (GSA discount) and OSI (EPA discount)
estimates are based, would not apply if the states were in-
voiced directly. The alternative to direct invoicing would be
the provision of user accounting data to the EPA with distri-
bution billing of the states performed by the EPA. This mech-
anism is unacceptable, because it increases the cost to the
EPA, and because it would neccessitate multiple handling of
accounting data.
An additional consideration with the candidate systems is
the contractors' current hardware capacity. Both DCS and GSA
INFONET have the core capacity and the network capability re-
quired. The RTCC would need additional terminal lines to ser-
vice the proposed system. OSI might be expected to reach a
core capacity peak in the near future, and increased core would
need to be added (this input comes from a representative of the
30
-------
Management Information and Data Systems Division of the EPA).
Current cost estimates do not include these possibilities.
A third problem with the analysis is the inordinately
high on-line storage costs of the GSA INFONET system. These
estimates are based on the total data base being on-line at
all times. This would not be an acceptable way to operate
the system, and the CSC representatives have indicated that
with a more in-depth system definition they could anticipate
revised storage estimates that would be more competitive.
The capabilities of each of these networks are summarized
in Table 14.
Table 14. NETWORK BENEFIT COMPARISON
Benefit
Adequate experience
with networks
Experience with EPA
Established nation-
wide networks
Can invoice states
directly at rates
quoted
Sufficient capacity
Network
BCS
yes
yes
yes
yes
yes
EPA/OSI
yes
yes
yes
no
maybe
GSA- INFONET
yes
yes
yes
no
yes
RTCC
yes
yes
yes
no
maybe
The contrasting costs of the four networks are derived as
follows. Each network is assumed to service 25 "typical"
states distributed throughout the contiguous 48 states. The
estimated costs are first-year network costs including both
recurring and non-recurring costs. Table 15 summarizes the cost
figures.
Any final recommendation as to which system would be most
cost effective to implement should be deferred until all of the
problems discussed here have been resolved. Any feasibility
study should address these problems.
31
-------
Table 15. NETWORK FIRST-YEAR COST COMPARISON
Cost factor
Recurring Costs
(Per month)
Connect time
CPU time
Storage time
Terminal rental
Central system updates
State agency operations
Subtotal
(Per month)
Non-Recurring Costs
Software and data base
conversion
Initial familiarization
Subtotal
Total First-Year Cost
Total EPA Costs
Total State Costs
Typical State Costs
Network
BCS
2,600
4,300
2,400
5,800
700
5,000
21,000
7,000
25,000
32,000
284,000
21,000
263,000
9,400
EPA/0 SI
1,500
1,700
2,480
5,800
700
5,000
17,000
7,000
25,000
32,000
236,000
20,000
216,000
10,000
GSA-INFONET
4,700
1,600
21,000
5,800
350
5,000
38,000
330
25,000
25,000
480,000
10,000
470,000
37,000
RTCC
4,700
940
-
5,800
350
5,000
17,000
—
25,000
25,000
229,000
9,500
220,000
8,700
32
-------
2.0 TECHNICAL REPORT
This is a cost/benefit analysis of four possible systems
to provide state agencies with computerized data processing
of their air quality and emissions data.
Each of the four alternative systems is a computer network
with one central computer and remote terminals at the partici-
pating state agencies. The four networks analyzed are: (1)
the GSA INFONET (2) the EPA/OS1 System (Optimal Systems, Inc.);
(3) a Research Triangle-based (EPA) network; and (4) BCS
(Boeing Computer Services, Inc.) an independent commercial
network. Each of these networks is assumed to use the EPA-de-
veloped Comprehensive Data Handling System (CDHS), and in
particular the Air Quality Data Handling Subsystem (AQDHS) and
the Emissions Inventory Subsystem (EIS).
The first subsection (2.1) gives an overview of the sys-
tems being considered, delineates the requirements of each
system, and generally provides the background and framework
within which the cost/benefit analysis is conducted.
Subsection 2.2 contains the cost/benefit analysis itself,
including a standardized list of assumptions and the inputs
required for each network considered. The results of informal
interviews with representatives from each network are reported
as is the result of an interview with a terminal supplier and
the result of another interview regarding CDHS conversion.
Using these inputs each network is evaluated for its cost and
suitability. Various cost tradeoffs are made and sensitivities
in the results are explored.
33
-------
2.1 SYSTEM OVERVIEW
These are 55 states and territories in the United States
which are required to collect, use, and submit air quality and
emissions data to the National Air Data Branch. Two computer-
ized data banks, SAROAD and NEDS have been setup at NADB to
manage these data. Many states process their data at their own
facilities, using their own computers. These states setup and
maintain their own data files, interpret and analyze their
data, and possess a practical means of transmitting their data
to regional and national levels of the EPA for inclusion in
NEDS and SAROAD. Many states, however, still use manual data
processing systems.
Table T. 16 lists those states, by EPA region, which have
manual emissions inventory systems and air quality data han-
dling systems. Note that a total of 31 states are involved;
16 have neither data base automated, 13 have an automated air
quality data handling system but lack an automated emissions
inventory system, and 2 have an automated emissions inventory
system but lack an automated air quality data handling system.
The statistical demands in terms of data input and output
that these states are expected to place on a terminal network
are given in Section 3.3 of the Task 1 report (Projected
Activity for Alternative Systems) and are further summarized in
Section 2.2 of this report. More generally, however, a state
agency currently without an automated system for handling air
quality and emissions data might want such a capability and
would like it to have the following characteristics. It must
be under the direct control of the state agency insofar as its
own data are concerned. Access to the system for both data
input and retrieval should be via a teletype terminal requiring
only minimal operator training. Each terminal must be equipped
with a card reader. Terminal speeds and turnaround time on
retrievals are not critical. Batch retrieval of lengthy re-
ports mailed from the central facility with overnight turn-
around on the central processor is adequate for most purposes.
-------
Table 16. STATES WITH MANUAL DATA SYSTEMS
EPA Region
Region I
Region II
Region III
Region IV
Region V
Region VI
Region VII
Region VIII
Emission inventory
State
Maine
New Hampshire
Rhode Island
Vermont
Puerto Rico
Virgin Islands
Washington, D.C.
West Virginia
Florida
Georgia
North Carolina
Tennessee
Minnesota
Arkansas
Louisiana
Oklahoma
Kansas
Colorado
Montana
North Dakota
South Dakota
Utah
Wyoming
Number of
point sources
380
290
165
145
340
85
110
550
2150
1020
3075
2030
690
695
1250
825
345
445
340
180
115
100
185
Air Quality
State
Maine
New Hampshire
Rhode Island
Vermont
Puerto Rico
Virgin Islands
Washington, D.C.
West Virginia
Louisiana
Iowa
South Dakota
Wyoming
No. of sensors
Currently
reporting
16
33
64
2
13
6
13
52
34
30
4
8
Projected
47
65
83
22
64
10
27
64
34
67
10
13
en
-------
Table 16 (Continued). STATES WITH MANUAL DATA SYSTEMS
EPA Region
Region IX
Region X
Totals
Averages
Emission Inventory
State
American Samoa
Arizona
Guam
Hawaii
Alaska
Idaho
Number of
point sources
260
12
475
100
350
16707
576
Air Quality
State
American Samoa
Arizona
Guam
Hawaii
Nevada
Alaska
No. of sensors
Currently
reporting
0
55
5
39
48
19
441
25
Projected
2
72
6
34
55
38
713
40
-------
Some interactive capability is desirable for editing data in-
puts and querying the central facility regarding report re-
trievals (job status). A capability to store historical data
and various summaries of the data from the state at the central
processor is required. Data summaries should be updated on at
least a quarterly basis with more frequent updates being de-
sirable. Payment for system utilization should fall entirely
on the user, although initial setup costs and periodic updating
of the official data base would be the responsibility of the
EPA.
Two approaches are considered for data storage. The first
would transfer the entire NEDS and SAROAD data banks to the
central facility for storage and access by the individual
states. This approach is not satisfactory, because it requires
that large quantities of data be stored which are never (or
rarely) utilized, such as the data from states that already
have computerized systems and, therefore, would not be part of
the proposed network. The other alternative is to place in
storage only those data which have been accumulated for a given
state as that state enters the network. Storage costs could
thus be reduced and the states charged a more equitable storage
fee. This latter alternative is more reasonable.
The computer software to be used has already been devel-
oped. It will handle air quality data in SAROAD format and
emissions data in NEDS format. It is referred to as the Com-
prehensive Data Handling System(CDHS) and was originally devel-
oped for use by state and local air pollution control agencies
as an air quality management and technical data handling sys-
tem. It includes three subsystems, but only two are of direct
interest in this study. The first, Air Quality Data Handling
Subsystem (AQDHS) is software for processing air quality data
in SAROAD format; the other, Emissions Inventory Subsystem
(EIS) is software for processing the emissions data in NEDS
format. In each of the networks considered, the central com-
puter will function as the processor for each state agency
37
-------
through remote terminal access. Direct invoicing of the states
would be expected to increase all of the cost estimates except
the .BCS estimates.
2.2 CANDIDATE NETWORK EVALUATION
A standard set of assumptions was prepared and provided to
a representative from each potential network in an informal
interview. These assumptions are given in subsection 2.2.1
very much as they were given to each of the candidates. The
inputs required from each candidate network were similarly
defined and provided to the various representatives for their
response. The general input requirements are listed in sub-
section 2.2.2 with the specific response from each candidate
summarized in subsections 2.2.2.1 through 2.2.2.4. A fifth
interview was conducted with a terminal supplier to get inde-
pendent inputs on this critical aspect of the system, and
Mr. L. Hedgepeth of the EPA was consulted regarding the imple-
mentation costs of CDHS. The results of these interviews are
given in subsections 2.2.2.5 and 2.2.2.6.
The actual comparison of network costs and the discussion
of results are in subsections 2.2.4 and 2.2.5.
2.2.1 Assumptions
There are up to 31 state agencies that might be interested
in a terminal network for air quality data, emissions inventory
data, or both. These 31 agencies may be subdivided into three
groups. Group A (16 states) requiring both air quality and
emissions inventory data. Group B (13 states) requiring only
emissions inventory data, and Group C (2 states) requiring only
air quality data.
These agencies, for the most part, are located in the more
remote and less densely populated areas of the contiguous 48
states. Although Alaska, Hawaii, and the several island terri-
tories were used to derive average usage figures, the special
problems associated with these agencies are excluded from this
38
-------
analysis. These problems are basically communications network
problems.
Each agency will utilize a low-speed terminal with punched
card input. The agency will be able to interactively request
a job but the job itself will be remotely run and its output
mailed to the agency later. Turnaround time is assumed to be
an overnight run plus mail transit time. Data entry will
require an edit listing back on both the air quality and emis-
sions inventory systems.
Annual report generation for a typical Group A state to
meet minimum requirements are:
Emissions Inventory
Air Quality
Totals (Approximate)
a
No . of j obs
19
74
100
b
Pages of computer printout
2400
1850-3000
5000
a The emissions inventory jobs are described in Table 2; the air quality
jobs are described in Table 3.
Pages of computer printout are derived from Tables 19 and 20 by
transforming the figures given there to a per state basis.
39
-------
Table 17. ANNUAL DATA VOLUMES
Annual data input volumes for a typical Group A state are:
j, V^
No. of jobs Cards of input
Emissions Inventory 15 175
Air Quality 68 10000-17000
Totals (Approximate) 100 14000
Data storage volumes of currently existing data bases
are as follows:
Volume0 Equivalent No. of data cards0
Emissions Inventory 88,000 500,000
Air Quality
(data values) 31,000,000 11,250,000
Totals 32,000,000 12,000,000
(Approximate)
Annual additions to the entire data bases are estimated
by NADB to be approximately:
Volume Equivalent No. of data cards
Emissions Inventory
(records) 10,000 60,000
Air Quality
(data values) 11,000,000 4,000,000
Totals 11,000,000 4,000,000
(Approximate)
a For emissions inventory 12 monthly updates plus 2 semi-annual
inputs plus one annual update give 15 jobs; air quality as-
sumes weekly, monthly and quarterly updates.
Cards of input for emissions inventory come from Table 10 of
the Technical Report on Task 1 by determining the average up-
dates per year per state and multiplying by 6 cards per update,
Air quality input volume refers to Tables 11 and 12, and uses
an average to get annual volumes per state.
The number of emissions inventory records currently in NEDS
and the number of air quality data values in SAROAD were
provided by NADB.
40
-------
The current storage requirements for a "typical" state
are approximately:
Emissions Inventory
(records)
Air Quality
(data values)
Totals
(Approximate )
Volume
1,000
560,000
560,000
Equivalent No. of data cards
7,000
200,000
210,000
In addition to the above basic data storage requirements,
nine 2314 disc files of data summaries would have to be moved
initially. The entire data base will be updated at least
quarterly and, perhaps, monthly. It is assumed that with mini-
mal programming only that portion of the overall data base and
associated summaries which apply to states utilizing the data
network need be transferred to a central computer location and
stored. All necessary programming would be in COBOL.
Preliminary estimates indicate that about 10 hours of
computer connect time per agency per month will be required.
Typical report processing times, on an IBM 360/50 CPU, are
given in Table 17 as provided by MADE.
2.2.2 Network Evaluation Input
Inputs required for each network evaluation are basically
those listed in Table 19. During an informal interview, a rep-
resentative of each network was asked to complete this table
and/or to provide the inputs from which the values could be
estimated. Any other cost factors or benefits that the network
representatives felt should be included in the analysis were
invited.
It was specifically requested that the following factors
be considered:
1. Is the assumption of overnight turnaround reasonable?
Is there an extra cost for assuring overnight turnaround? What
is the maximum turnaround time without priorities? What are
41
-------
Table 18. TYPICAL REPORT PROCESSING TIME
NO. Of
jobs
annually
EMISSIONS INVENTORY
Point Source Listing
Selected Source Listing No. 1
Selected Source Listing No. 2
Emissions Summary
Area Source Listing
Totals
a
AIR QUALITY
2
12
2
2
1
19
No. of
pages of
computer
printout
per job
600
5
30
350
350
Minutes of
360/50 CPU
time/page
0
0
0
0
0
No. of stations
Oper- Pro-
.02
.06
.06
.09
.03
Minutes
per job
12
0.
1.
31.
10.
.3
,8
.5
.5
No. of stations
Oper- Pro-
ating jected
Raw Data Listings
Quarterly Summary Reports
Quarterly Frequency
Distributions
Yearly Frequency Distributions
Yearly Summary By Quarters
Totals
64
4
4
1
1
74
25
25
25
25
25
40
40
40
40
40
. 0.
0.
0.
0.
0.
OS
055
08
02
07
ating
1.
1.
2.
0.
1.
25
375
0
5
75
jected
2.0
2.2
3.2
0.8
2.8
Total
360/50
CPU time
Minutes
24
3.6
3.6
63
10.5
104.7
No. of stations
Oper- Pro-
ating
80.
5.
8.
0.
1.
95.
0
5
0
5
75
75
jected
128.0
8.8
12.8
0.8
2.8
153.2
Operating refers to average number of stations currently reporting; Projected refers
to average number of stations projected in the State Implementation Plans.
-------
Table 19. TERMINAL NETWORK COST ESTIMATES
Cost Element
Expected Value
1. Computer Connect Time
Entry Time
Print Time
2. Processing (CPU) Time
3. Storage Time
Total Data Base - On-line
Off-line
Typical State Data Base - On-line
Off-line
4. Terminal Rental (including maintenance)
5. Maintenance of Duplicate Systems
Conversion of Data Base
Monthly Maintenance
6. Impact on State Agencies Management
Initial Training and Familiariza-
tion
Monthly Manpower Costs
43
-------
your particular network priorities, what do they provide, and
what do they cost?
2. What are the costs for quarterly (monthly) data base
updates? What are the costs for program changes (up to 50
small changes per year)?
3. What are the costs for initial conversion from the
RTF Univac 1100 system? This will involve at least 100 pro-
grams (mainly retrievals) to be compiled and an interactive
query capability to be implemented, as well as the initial
file conversion and a program to incorporate file updates.
4. Is the capability to invoice states directly for usage
of the system available? Could the capability be implemented?
Responses to the interviews are summarized in the follow-
ing sections.
2.2.2.1 BCS System - One way to implement a non-EPA user net-
work is to contract with an existing commercial network. A
number of possibilities exist, but Boeing Computer Services,
Inc. (BCS) was selected as being representative. BCS was pre-
sented with the information outlined in the previous two sub-
sections of this report, and they provided the following infor-
mation.
BCS suggests the use of their MAINSTREAM-TSO/RJE. This
system offers on-line conversational editing and program devel-
opment; batch processing and a choice of output media: printed
copy, punched cards, or microfiche. The MAINSTREAM-TSO is the
BCS adaptation of the Time Sharing Option of the IBM Operating
System. It utilizes a purchased two-million-byte IBM System/
370 Model 168 computer system located at the BCS data center in
McLean, Virginia. It allows terminal users to enter, store,
modify, and retrieve data in a fully 'conversational edit mode,
and to submit programs (coded in any OS-supported programming
language) for compilation and execution in a remote batch
environment.
1) This subsection is based on information provided by Donald
D. Davenport and Larry E. Parish of the Los Angeles Offices
of BCS on 30 May and 16 July 1974.
44
-------
Table 20. BCS TERMINAL NETTORK COST ESTIMATES
Cost element
Expected value
2,
3,
Computer Connect Time
Entry Time
Print Time
Processing (CPU) Time
Storage Time
15% of Total Data Base - On-line
Total Data Base - Off-line
Typical State Data Base - On-line
Off-line
$10 0/user/month
$167/user/month
$2,.400/month
$780/month
$280/user month
$14/user/month
4,
5,
Terminal Rental (including maintenance) $225/user/month
Maintenance of Duplicate Systems
Conversion of Data Base
Monthly Maintenance
Impact on State Agencies Management
Initial Training & Familiarization
Monthly Manpower Costs
$7,000
$350/month
$l,000/user
$ 2 0 0/user/month
Cost estimates are summarized in Table 20. The derivation
of each estimate is explained below:
1. Computer connect time is defined as entry time and
print time. Each is charged at $10 per hour, and assuming 10
hours connect time per user per month, the $100/month/user
figure results. In the worst case, BCS estimates that the
total entry and retrieval cost would be $150/month/user.
2. Using the equivalent 360/50 CPU time from Table 17 and
given an estimated core size of 125K-bytes per program and 2000
disc accesses per emissions inventory program, 3500 accesses
for air quality raw data listings and 2500 accesses for the
other air quality programs, BCS estimated a monthly processing
cost of $167 per user. This is based on their pricing algo-
rithm and internal assumptions regarding relative processing
times. (The pricing algorithm is shown as Equation 1 below.
Many of the terms of this equation have their own formula for
deriving the quantative factor to be entered.)
45
-------
Equation 1
BCS Price = Service Factor x (CPU time + disc and tape I/O
+ core usage + tape drives + disc spindles +
job schedular) + lines printed + multiple
parts + cards punched + cards read + connect
charge + storage.
3. Storage costs for raw data and data summaries are de-
rived by first assuming that the total NADB data bases were on-
line full time. Then twelve million cards of raw data, or 960
million characters, would require ten Model 3330 disc packs.
Three disc packs of data summaries would be required. With an
on-line storage cost of $l,200/month/3330 disc pack, the month-
ly charge for 13 packs would be $15,600.
A more reasonable assumption would be to transfer only
that portion of the NADB data base to the system that pertains
to the user states. Specific volume estimates are not readily
available, however. A second assumption, one that can be
applied here, is that only 15 percent of the files would need
be on-line at any time. The more practical estimate then
becomes 0.15 x 13 x $1,200, or approximately $2,400/month for
on-line storage costs.
Off-line storage costs are $60/month/disc/pack, or $780/
month for a data base of 13 disc packs. Both on-line and off-
line costs would be expected to be somewhat lower than the
figures derived here, since the entire NADB data base would
not be transferred to this system. (Estimates of how many
disc packs are required for the 25 projected user states are
not available. Therefore, storage costs based on the total
data base were used.) A typical state using this system might
require only 1/6 of a disc pack for raw data and 1/15 of a
disc pack for summary data. Projected costs for on-line stor-
age would then be $280/month. Off-line costs for a typical
state would be $14/month.
4. A simple 30 character per second terminal is felt by
BCS to be adequate. The rental cost for such a terminal, in-
cluding installation and maintenance, is estimated by BCS to
be $100 per month per user. No terminal installation costs are
anticipated. The required card reader would cost an extra $125
per month, for a total terminal cost of $225 per month.
46
-------
5. The cost of maintaining duplicate systems consists of
an initial component and a monthly component. BCS estimates
that to train one of their people (including suitable backup)
will cost about $5,000. This person would be the BCS contact
for all activities involving the network. It is also estimated
to cost from $1,000 to $5,000 to install the data base on the
BCS machine; expected value is $2,000. This includes any
requisite input software development. The person assigned to
the proposed terminal network would be used about 1 to 3 days
each month at an anticipated cost of $200 to $500.
6. BCS suggested that each participating agency send one
man and an alternate to BCS for a 1-day training session.
Monthly operation for each state agency is estimated to be on
the order of 10 hours per month per user at an average burdened
salary of $20 per hour or $200 per month. Training costs are
estimated as $500 travel and per diem plus $320 for salaries
plus $180 for administrative and other costs; i.e., $1,000
per user.
7. Additional considerations included the following
items:
(1) Overnight turnaround is no problem. It is in fact
the lowest priority and no extra cost is associated
with assuring it. Network priorities range from
overnight to 10 minutes turnaround, with multipli-
cative factors on the computer charging unit price
ranging from 0.85 to 3.
(2) BCS suggests that EPA retain programming control and
enter data base updates and programming changes
directly, thus obviating all costs except those in-
cluded in BCS personnel costs (for coordination) and
those generated by EPA in simply entering the changes.
(3) Costs of initial conversion of the CDHS from the RTP
Univac system are included in the initial component
of the maintenance of duplicate systems given in
Item 5 of Table 20.
47
-------
(4) The BCS terminal network has the capability to bill
the states, and any other users such as NADB, directly
for their utilization of the system at the rates
quoted.
(5) A change of contractor would incur at least the
costs associated with starting up the BCS system,
or approximately $7,000.
(6) BCS estimates that there would be no significant
impact on overall costs from reducing input or
output frequencies so long as the total volume of
data input and retrieval remains constant. Some
reduction in CPU time could be anticipated but total
cost savings are felt to be essentially negligible.
2.2.2.2 EPA/OSI System - Optimal Systems, Inc.(OSI) currently
is an EPA contractor. EPA Regional Offices nationwide access
the OSI system for various applications. Some Regional
Offices, for example, edit SAROAD data on the OSI system
prior to making data submittals to NADB. Regional Offices
also are using the OSI facility for modeling applications
and for MARK IV applications written within the Regional
Offices. Under the EPA contract OSI has the capability to
provide states with access to emissions inventory and air quali-
ty data. OSI uses IBM 360/65 computers connected to a national
communications network for input, output, and the proposed
network. It does not, however, include a mechanism for direct
billing to state users at rates that reflect current EPA discounts.
This mechanism would have to be developed. The subject is
treated in greater detail in the conclusion of this report.
* This subsection is based largely on information provided by
Susan Falkson of the Management Information and Data Systems
Division of the EPA in an interview on 11 June 1974.
48
-------
Cost estimates are summarized in Table 21. The deriva-
tion of each estimate is explained below:
1. Connect time, consisting of entry time and print
time is assumed to be 10 hours per month. From a printout
of EPA/OSI Contract History for Cost, Usage, and Statistics,
the unit cost of current hours (including toll-free telephone
service) is given as $5.74 for April 1974. Thus, 10 hours
of connect time would cost $57.40.
2. Processing time charges come from the same report
alluded to above and again are for April 1974. The overnight
charges are given as $375 per hour. Using the annual process-
ing time given in the assumptions, Table 17, and assuming
processing rates are twice as fast as for the IBM 360/50, the
monthly charge would be from $52/month to $67/month using
operating and projected air quality stations, respectively.
The higher figure is used in Table 21.
3. Storage time costs are estimated from the following
direct inputs. On-line storage costs were given independently
as 17 cents per 3330 track per month or $1,240 per month for
an entire disc pack. It was previously estimated that ten
such packs would be required for the total data base and
three additional packs for data summaries. Again, these fig-
ures reflect the total NADB data base storage requirements.
Assuming 15 percent of the files on-line at any given time
would give a projected on-line storage cost of $2,480 per
month as opposed to $16,120 per month if all files were on-
line continuously. Off-line storage costs are $48 per month
per pack; assuming 13 packs, this is $624 per month. A typ-
ical state in the network would require about 1/6 disc pack
to satisfy current requirements for raw data - an on-line
storage cost of $207 per month and an off-line cost of $5 per
month. On-line storage of data summaries of a typical state
in the network would require about 1/15 disc pack for a total
of $13 per month. Off-line storage for data summaries for
the same state would be $3 per month.
49
-------
Table 21. EPA/OSI TERMINAL NETWORK COST ESTIMATES
Cost element
Expected value
1. Computer Connect Time
Entry Time
Print Time
2. Processing (CPU) Time
3. Storage Time
15% of Total Data Base - On-line
Total Data Base - Off-line
Typical State Data Base -
On-line
Off-line
4. Terminal Rental (including
maintenance)
5. Maintenance of Duplicate Systems
Conversion of Data Base
Monthly Maintenance
6. Impact on State Agencies Management
Initial Training and Familiar-
ization
Monthly Manpower Costs
$ 57.40/user/month
$67/user/month
$2,480/month
$624/month
$220/user/month
$8/user/month
$225/user/month
$7,000
$350/month
$l,000/user
$200/user/month
4. Terminal rental figures are the same as the BCS
estimates since the EPA/OSI interview agreed with BCS that the
states (or at least someone else) must supply the terminals,
modems, maintenance, etc. The EPA/OSI interview did not pro-
vide any direct estimates, hence those provided by BCS are
used here.
5. The EPA/OSI interview did cover the conversion problem
but did not provide any direct estimates of conversion costs.
No problem was anticipated in the conversion process and, from
what was said qualitatively, one would have no reason to assume
that the EPA/OSI costs would exceed the BCS estimates. The
monthly maintenance figure comes unmodified from BCS.
6. The impact on the state agency's manpower requirements
also comes from the BCS estimate since none was ventured in the
previously referenced interview and there seems to be no reason
SO
-------
why the BCS and EPA/OSI systems should result in a signif-
icantly different cost for this factor.
7. Additional considerations include the following items:
items:
(1) The assumption of overnight turnaround is
reasonable and there is no added cost for assuring that it is
received. There are 7 priority levels in this system ranging
from 15 minutes to overnight turnaround. The intermediate
.turnarounds are 30 minutes, 1 hour, 2 hours, 3 hours, and 4
hours. Unit cost per resource hour as of April 1974 was $1,181
for 15 minute turnaround and $375 for overnight turnaround.
(2) EPA/OSI, like BCS, suggests that EPA retain
programming control and enter programming changes and data
base updates directly thus obviating all costs except those
generated by EPA in simply entering the changes. No estimate
of this cost was attempted.
(3) The costs associated with initial conversion
of the system and subsequent changing of contractors are
taken from the BCS interview since no quantitative estimates
of these costs were made. It was the opinion of the EPA/OSI
representative that this was not a major problem, however,
thus lending credence to the BCS estimate.
(4) As discussed earlier, there is currently no
capability to invoice the state directly at EPA discount
rates. This will be a major consideration for any final
decision on a preferred network. If the capability to in-
voice states directly were implemented, then the processing
rates quoted here may no longer apply, since these rates
are based on current EPA discounts.
2.2.2.3 The GSA INFONET. System1 - General Services Administra-
tion (GSA) is authorized to establish and operate data pro-
cessing centers providing services to all Federal agencies.
This subsection is based largely on information provided by
Patrick Hodges of the Information Network Division of the
Computer Sciences Corporation in an interview on 27 June 1974,
51
-------
There are two time-sharing systems available through GSA - the
conversational GE-440 Time Sharing System (RAMUS) and the
conversational/remote batch CSC (Computer Science Corp.)
National Teleprocessing System (INFONET). The second of these
systems is more suitable for providing states the required
access to emissions inventory and air quality data. The GSA
INFONET system uses Univac 1108 central processors, remote com-
munications concentrators (RCC's), and accommodates either
low- or high-speed terminals at the network user agencies.
Provision of the technical capability required by the pro-
posed network is straightforward; however, the users cannot be
charged directly for their utilization of the system at rates
that reflect normal GSA discounts. If the states were invoiced
directly, commercial CSC INFONET rates would apply. To qualify
for the discounted rates quoted here requires that EPA be in-
voiced for services rendered. This invoicing can be itemized
by states so that EPA could pass the charges on to the states
by deducting matching funds or through some other process.
This alternative would, however, place a burden on the EPA.
Cost estimates are summarized in Table 22. The derivation
of each estimate is explained below:
1. Connect time, consisting of entry time and print
time is assumed to be 10 hours per month. The current billing
rates for low speed terminals are as shown in the following
table:
Characters
per second
10
15
30
Prime Time,
Dollars/hr
5.60
7.44
9.30
Non-prime time (after
Dollars/hr
3.41
4.03
4.65
7 PM)
The differing quantities of data, operating policies, bud-
gets, time zones, etc., for the state agencies which will form
52
-------
the network could easily result in use of any of the six com-
binations of terminal speeds and prime or non-prime time trans-
mission. In order to get a single figure for use in Table 21,
the six figures given above are averaged, using equal weights,
resulting in a figure of $5.74 per hour. This is only coinci-
dentally precisely the same figure given for the EPA/OSI net-
work. There was no indication of a monthly surcharge independ-
ent of use. Communications charges to the RCC are separate,
however, and are invoiced at 54 cents per month per mile. If
an average of 277 miles to the RCC is assumed (see Table 23) another $123
per month is required for connect time giving the figure in Table 22.
Table 22. GSA INFONET TERMINAL NETWORK COST ESTIMATES
Cost element
Expected value
1. Computer.Connect Time
Entry Time
Print Time
2. Processing (CPU) Time
3. Storage Time
15% of Total Data Base-On-Line
Total Data Base-Off-Line
Typical State Data Base-
On-Line
Off-line
4. Terminal Rental
(including maintenance)
5. Maintenance of Duplicate Systems
Conversion of Data Base
Monthly Maintenance
6. Impact on State Agencies
Management
Initial Training and
Familiarization
Monthly Manpower Costs
$18 0/user/month
$60/user/month
$21,000/month
$200/month
$2,350/user/month
$4/user/month
$ 22 5/user/month
$330
-0-
$1,000
$200
53
-------
Table 23. LENGTH OF TYPICAL GSA INFONET TELEPHONE TOLL CALL
EPA
Region
I
III
IV
V
VI
VII
VIII
IX
X
State
Maine
New Hampshire
Rhode Island
Vermont
District of Columbia
West Virginia
Florida
Georgia
North Carolina
Tennessee
Minnesota
Arkansas
Louisiana
Oklahoma
Kansas
Iowa
Colorado
Montana
North Dakota
South Dakota
Utah
Wyoming
Arizona
Nevada
Idaho
State
Capital
Augusta
Concord
Providence
Montpelier
Washington, D.C.
Charleston
Tallahassee
Atlanta
Raleigh
Nashville
St. Paul
Little Rock
Baton Rouge
Oklahoma City
Topeka
Des Moines
Denver
Helena
Bismarck
Pierre
Salt Lake City
Cheyenne
Phoenix
Carson City
Boise
Concentrator
city
Boston
Boston
Boston
Boston
Washington, D.C.
Washington, D.C.
Atlanta
Atlanta
Washington, D.C.
Memphis
Chicago
St. Louis
Ft. Worth
Ft. Worth
Kansas City
Kansas City
Denver
Seattle
Denver
Denver
Denver
Denver
Los Angeles
San Francisco
Seattle
Estimated
mileage
158
69
40
148
0
257
237
0
237
207
346
296
375
188
59
188
0
494
543
405
375
99
356
178
415
en
Total
5670
Average
227
-------
2. Processing time charges are given as standard
rates per Standard Resource Unit (SRU), which is approximately
one-half of a CPU second. Priority 1 (72 hour turnaround) is
$0.0186/SRUf Priority 4 (24-hour turnaround) is $0.0279/SRU and
Priority 9 (2-hour turnaround) is $0.0992/SRU. Assuming a 24-
hour turnaround and assuming that 360/50 CPU time is 20 percent
slower than Univac 1108 CPU time, the monthly cost per user
ranges from $47 (using operating monitoring stations) to $60
(using projected monitoring stations). The latter figure is
used in Table 22.
3. Storage time costs for on-line storage were
estimated directly to be $163 per day for emissions inventory
data and $3,600/day for air quality data. Using 30 days per
month this yields a monthly charge of $115,000. The formula
for estimating on-line costs was not provided. Off-line
charges were estimated by taking the total size of the data
base, 12 million cards, multiplying by 80 characters per card
to get 960 million characters, dividing by 100 million char-
acters per disc pack to get 10 disc packs and multiplying by
a daily charge of $0.496 per disc pack to get a daily charge
of $4.96. Multiplying by 30 days per month gives the annual
monthly off-line storage charge of $150. Added to this is the
charge for storing three disc packs of data summaries at a
charge of $0.496 per disc pack/day or $45 per month. On-line
costs for the data summaries are ($13.95 per disc pack per
hour) 720 hours per month (3 disc packs) = $30,000. A more
realistic estimate of 15 percent of the data base on-line
would cost approximately $21,000/monthly. An individual state
projected for the network would require about 1/6 of a_3300
type disc pack to satisfy current requirements. This would be
about $1,700 per month on-line and $2.50 off-line. The data
summaries which are pertinent to the states being considered
should require no more than two 3330 disc packs. Prorated
costs would be about $650 per month for on-line storage and
about $1 per month for off-line storage. The figures for the
55
-------
typical states in this network are based on any one state
leaving its data on-line continually. This assumption allows
a worst case estimate.
4. Terminal costs were given as $80 per month for
a Teletype Model 33 terminal using GSA discount. Since GSA
discount may not apply for the proposed network, the terminal
cost for this network is revised to be $225 per month per user
to parallel the estimates for the other networks.
5. No problems are anticipated with transferring or
maintaining the data base. Under repeated questioning it was
estimated that $330 would be required to convert the data base
to the CSC machine, and no subsequent costs would be charged.
6. No estimate of the impact to the states' manpower
requirements was attempted. The impact should be the same as
for the other networks.
7. Added considerations are given in the following
items.
(1) There is no problem with overnight turn-
around. Priorities range from 1 (72 hours) to 9 (2 hours).
The cost for overnight (24 hour) turnaround is 1.5 times
greater than the cost for a 72-hour turnaround. Two-hour
turnaround is 3.5 times more expensive than a 24-hour turn-
around .
(2) Again, it was suggested that EPA retain
programming control and enter programmatic changes and data
base updates directly, thus obviating all costs except those
generated by EPA in simply entering the changes. No estimate
of this cost was attempted.
(3) No costs other than those given in para-
graph 5 above were associated with initial system conversion.
(4) As discussed earlier, there is currently no
capability to bill the states directly at GSA discount rates.
(5) The inordinate on-line storage costs (rel-
ative to BCS and OSD were pursued with the CSC INFONET repre-
sentative, but no resolvable cause was found for the signifi-
cant difference.
56
-------
2.2.2.4 Research Triangle Computer Center - A basic way to
implement a non-EPA user network is to provide the users with
direct access to the data banks stored on the Univac 1100 Com-
puter system at the National Environmental Research Center
(NERC) in Research Triangle Park.
It was, unfortunately, not possible to pursue this option
with personnel of the Data Systems Division at the NERC to the
extent that would have been desirable for the purposes of this
analysis. The basic input from the Research Triangle Computer
Center to this study is shown in Figure 1.
Cost estimates are derived using this input and making a
number of assumptions. The cost estimates are summarized in
Table 24; the derivation of each estimate is explained below:
1. Connect time is again assumed to be 10 hours per
month per user. From Figure 1 this would generate a charge of
$40 per month... To this must be added communications charges.
At least two ..WATS lines (at $2,200 per month per line) would be
required.* These charges are prorated among all 31 potential
users. Total monthly connect time, then, is $182 per user.
2. From Figure 1 processing charges are given as $300 per
hour. Using the annual processing time given in the assump-
tions Table 17, and assuming processing rates are three times as
fast for the IBM 360/50, the monthly charge would be from $28
to $36 per month using operating and projected air quality
stations, respectively. The higher figure is used in Table 23.
3. Since the data bases must be maintained at the RTCC in
any event it could be argued that no additional storage costs
would accrue to serve as the computer center for the non-user
terminal network, hence no entry is made for this cost category
in Table 24.
4. Terminal rental costs were not estimated, so the $225
figure is assumed here.
* To assure reasonable service 4 to 6 lines would be required.
57
-------
SUBJECT:
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Response to PRC's Recent Inquiry DATE: August 28, 1974
FROM: J. Michael
Chief, Systems Analysis and Programming Branch
TO: Gerald Nehls
National Aerometric Data Branch
Through: Harold B. Sauls
Director, Data
ems Division
We have some problems responding to all of PRC's questions.
Understandably, their comparison is designed for evaluating
commercial supplies of network services. We are not really able
to cost-various areas until we receive guidelines on any proposed
charge-back system for RTP.
In the absence of a costing yardstick, we are only able to
provide the-following:
1. An assumption of overnight turnaround is certainly
reasonable and does represent the maximum turn-
around time without priorities.
2. Our facility for billing state users would at present
be strictly limited to blanket billing of NADB.
3. Cost per hour of computer connect time = $4.00.
4. Cost per Univac 1110 computing hour would be
approximately $300.00.
Figure 1. Research Triangle Computer Center Input.
58
-------
Table 24. KTCC TEEMINAL NETWORK COST ESTIMATES
Cost element
Expected value
1. Computer Connect Time
Entry Time
Print Time
2. Processing (CPU) Time
3. Storage Time
15% of Total Data Base - On-line
Total Data Base - Off-line
Typical State Data Base -
On-line
Off-line
4. Terminal Rental (including
maintenance)
5. Maintenance of Duplicate Systems
Conversion of Data Base
Monthly Maintenance
6. Impact on State Agencies Management
Initial Training and Familiar-
ization
Monthly Manpower Costs
$182/user/month
$ 3 6/user/month
$225/user/month
$l,000/user
$ 2 0 0/user/month
5. Using the RTCC as the focal point of the non-EPA user
network would require no data conversion or maintenance of
duplicate systems.
6. The impact on the state agency's manpower requirements
also comes from the BCS estimates since none are available from
the RTCC.
7. The information provided in Figure 1 contains all
other available information from this potential network.
2.2.2.5 Terminal Devices - Representatives from each candi-
date network were asked for estimates of monthly terminal lease
costs. None, however, were prepared to discuss the subject in
detail. An independent terminal supplier was consequently
asked for input on terminal costs.
Terminals generally are available on a 1 to 3 year lease
basis. The terminal costed here includes a built-in acoustic
59
-------
coupler, an auxiliary keyboard, a form feed, and a card reader,
Typical lease terms are shown in Table 25.
Table 25. TYPICAL TERMINAL LEASE CHARGES
Terminal Printer
Acoustic Coupler
Auxiliary Keyboard
Form Feed
Card Reader
TOTAL (Cost per month)
One year
$125.00
16.50
9.00
2.50
115.00
$268.00
Two years
$110.00
14.50
8.00
2.00
100.00
$234.50
Three years
$100.00
12.50
7.00
1.75
90.00
$211.25
For the configuration discussed here, a $50 installation
fee is required. An additional $30 per year maintenance fee
would be assessed after the first year.
2.2.2.6 CDHS Conversion - This network would use the Compre-
hensive Data Handling System software already developed by the
EPA. Mr. Lloyd Hedgepeth of the EPA provided input for CDHS
conversion costs.
The longest conversion time required to install CDHS on
a state agency computer has been 4 man-weeks. The shortest
time has been 20 man-hours. An experienced programmer/analyst
should be able to install CDHS within 1 man-week on any of
the computers considered for this system. Typical CPU time
for compilation and test runs on CDHS are shown in Table 26.
Table 26. TYPICAL CPU TIME FOR CDHS
INSTALLATION ON AN IBM 360/50
Operation
Installation
Copy Procedures
Compilations
(11 programs)
Testing (programs
plus sorts)
Total
CPU time, minutes
4
1
16
66
60
-------
CDHS installation costs, then, assuming 1 man-week profes-
sional effort ($1,000) and approximately 1 hour CPU time
($1,000) could be expected to be approximately $2,000.
2.2.3. Cost Review
The contrasting costs for the four networks are shown in
Table 27. Each of the four networks is assumed, in turn, to
service 25 "typical" states distributed throughout the remoter
parts of the contiguous 48 states. This figure excludes
Alaska, Hawaii, and the four island territories from the 31
jurisdictions listed in Table 16 and assumes that each of the
remaining states is typical or average. The costs which are
estimated are total first year network costs including both
those which are recurring and non-recurring.
Table 27 summarizes the cost figures. The derivation of
<\
the individual entries are explained below. All values are
rounded to three significant figures.
The connect time charges are the entries in Tables ,20, 21,
22 and 24 multiplied by 26 to account for the number of assumed
states plus one additional terminal at RTF connected 10 hours
per month, to enable program and data base updates to be
entered. The processing time charges are similarly derived.
Storage charges present a special problem. First, the size
of the data base to be stored must be determined - toal or
only that attributable to the 25 states - and then the propor-
tion to be stored on-line must be decided. For this analysis
only that portion of the data base and data summaries attribu-
table to the 25 states is assumed to be stored. On-line versus
off-line storage costs are shown in the following array and are
derived by multiplying the appropriate values in Tables 20, 21,
and 22 by 25.
Network
BCS
EPA/OS I
GSA INFONET
On-line
7,000
5,500
58,750
Off-line
375
300
100
61
-------
It is obviously very desirable to keep as much of the
data bank off-line as much as possible because of the high
on-line storage costs. Assume that, on the average, 15 percent
of the data base is on-line and the rest is stored off-line.
Some programming will be required to achieve this but its
cost should be relatively invariant and insignificant. Then,
Table 27. NETWORK FIRST-YEAR COST COMPARISON
Cost factor
Recurring Costs
(Per month)
Connect time
CPU time
Storage time
Terminal Rental
Central system updates
State agency operations
Subtotal
(Per month)
Non-Recurring Costs
Software and data base
conversion
Initial familiarization
Subtotal
Total First-Year Cost
Total EPA Costs
Total State Costs
Typical State Costs
Network
BCS
2,600
4,300
2,400
5,800
700
5,000
21,000
7,000
25,000
32,000
284,000
21,000
263,000
9,400
EPA/OS I
1,500
1,700
2,480
5,800
700
5,000
17,000
7,000
25,000
32,000
236,000
20,000
216,000
10,000
GSA-INFONET
4,700
1,600
21,000
5,800
350
5,000
38,000
330
25,000
25,000
480,000
10,000
470,000
37,000
RTCC
4,700
940
-
5,800
350
5,000
17,000
—
25,000
25,000
229,000
9,500
220,000
8,700
62
-------
the figures of Table 27 result. No storage costs attributable
/i . . . • "
to the non-EPA user network are assumed for the RTCC,
Terminal rental costs are taken from subsection 2.2.2.5
and assume the additional of a card reader. The cost is then
$225 per terminal with 26 terminals in the network.
Central system update costs are taken directly from Tables
20, 21, 22, and 24 with an additional $350 per month added to each
figure to account for EPA personnel charges incurred in enter-
ing the changes. The state agency monthly operations charge is
taken directly from Tables 20, 21, 22, and 24, multiplied by 25 to
/
account for the assumed number of states in the network.
Software and data base conversion is taken from the
entries in Tables 20, 21, 22, and 24 as is the initial training and
familiarization cost of $1,000 per state.
The subtotals and totals are direct additions with the
monthly subtotals multiplied by 12 to get total recurring first
year costs. The cost of one terminal, its connect and CPU time
charges and the central system update costs were attributed to
EPA as were the non-recurring software and data base conversion.
All other costs were attributed to the states. Table 27 shows
these costs and the per state cost of the various networks.
2.2.4. Evaluation of Cost Survey Results
For all candidate networks surveyed, except GSA INFONET,
the operating cost estimates are sufficiently close to assume
that they represent the typical costs that would probably be
incurred. The discrepency with the GSA INFONET costs has not
been resolved. We can only conclude that differences in CSC
pricing algorithms and those used by the other candidate net-
works might require that a different approach in pricing would
be necessary. A second explanation might be that the system
was defined as being on-line continuously. Defining the system
as having portions of the data base on-line only as needed
would probably result in more realistic estimates for data
storage. This approach is probably the one that would be used
in any system feasibility study. Since no previous explanation
63
-------
of the problem was given prior to the interviews with the net-
work representatives, and because the problem was only defined
in general terms, the cost estimates given can be expected to
reflect only the cost range, but not specific costs for the
proposed system. To derive more specific costs would require
a more refined definition combined with a feasibility study.
In further pursuing the proposed system costs, the considera-
tions discussed here will be important.
lo For defining the system, the following information
would be required:
(1) a system flow diagram
(2) definition of the specific portion of the data files
to be on-line at any one time.
(3) a projection, by potential number of users in a re-
gion, of the prime time and non-prime time require-
ments .
(4) complete file descriptions.
2. The capacity to invoice user states directly is impor-
tant. Estimates in this survey, except BCS estimates, are
based on current Federal government discounts. Estimated
network operation costs would be expected to increase if states
were billed directly for the use of the system. If a breakdown
of usage by state were provided to the EPA with subsequent
billing of the states by the EPA, the EPA would incur an in-
creased administrative cost.
3. The mechanism and frequency of transferring updates
from NADB to the central system should be defined.
4. The cost of compilation for each program in the system
should be supplied to obtain a better estimate of the system
maintenance costc Univac to other systems conversion costs
should be considered.
5o Core storage or communications network capacity is an
important consideration, since some of the representatives
interviewed in this survey alluded to possible capacity
problems„
64
-------
6. The problem of on-line and off-line storage require-
ments should be better defined. The results of this prelimin-
ary survey illustrate the confusion that can result from
failure to specify on-line storage requirements accurately.
On-line storage would be expected to be low, since priority
processing and time-sharing mode processing is not a system
requirement.
65
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-450/3-75-049
2.
3. RECIPIENT'S ACCESSIOI*NO.
4. TITLE AND SUBTITLE
Establishment of a Non-EPA User System for State
Implementation Plans
5. REPORT DATE
January 1Q75
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
PEDCo Environmental Specialists, Inc
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
PEDCo Environmental Specialists, Inc
Suite 13, - Atkinson Square
Cincinnati, Ohio 45246
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-1001
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final Rpnnrt
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report presents the results of a survey conducted among selected
state air pollution control agencies to determine their current practices
and projected needs related to accessing U. S. Environmental Protection
Agency data bases. Alternative methods for allowing non-EPA users to use
the data bases were introduced. A preliminary cost survey was conducted
for a projected method for allowing state agencies to have direct access
to the data bases.
This is a preliminary analysis of expected costs for operating the
Comprehensive Data Handling System on a centralized computer accessed
through remote terminals located in state air pollution control agency
offices. The analysis was compiled from data obtained in informal dis-
cussions with four (4) computer centers that might be expected to be
candidate contractors for such a system. System usage inputs were obtained
from Task I of this contract.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. cos AT I Field/Group
EPA Data Bases
NEDS
SAROAD
Emissions Data
Air Quality Data
Compliance and Enforcement Data
SI P Regulation Data .
3. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
72
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
66
------- |