LEADS System Overview

1
METEOSTAR LEADS
2
WHAT IS LEADS

• Leading Environmental Analysis and Display System
• Development Began At Lockheed In Late 1989 To
  Produce A New Commercial Weather Analysis Product
• LEADS Grew Out Of Two Military Weather Contracts
• TESS(3) -- A Navy Program Which Uses Surface
  Observations And Numerical Forecast Models To
  Provide Tactical Weather Forecasts For Carrier
  Fleets
• Mark IVB -- An Air Force Program Which Uses
  Satellite Imagery And Analysis Tools To Provide
  Long-Range Strategic Forecasts
• In 1994 Lockheed Was Brought Under Contract With
  TCEQ To Add Pollution
• In 1997 Information Processing Systems (IPS)
  Acquired The MeteoStar Product Line

3
SYSTEM CAPABILITIES

• Near Real-time Data Acquisition
• Pollution Sensor Data
• Water Probe Data
• Conventional Meteorological Surface Data
• Weather Satellite Imagery (Geostationary)
• Automated Gas Chromatograph (AutoGC)
• Upper Air Data
• Radar Profiler And Acoustic Sounder
• Visualization
• Long-term Data Archival
• EPA Reporting
• Analysis
• Ozone Action Day Prediction
• Transport Studies
• Pollution Source Determination

4
SYSTEM PHILOSOPHY

• The MeteoStar LEADS At TCEQ Was Designed As An
  Upgrade To An Existing Network
• TCEQ Was Unable To Expand Their Existing Network
  Due To Hardware Constraints In Their Data
  Collection System
• Individual Instruments And Sites Worked Well, But
  No New Instruments Could Be Added
• Communications And Data Collection Needed
  Upgrading
• A New Communications Network Was Designed And
  Interfaced Into The Existing TCEQ Monitoring
  Network
• Changes At The Monitoring Sites Included A New
  Datalogger, An Automatic Gas Calibrator And
  Valves, And A New Modem
• TCEQ Specified The Gas Calibrator Separate From
  Other Equipment And Selected The Dasibi 5008
• The Zeno Datalogger Was Chosen Before The
  Calibrator Was Selected Based On Its Capabilities
  And The Willingness Of Coastal To Modify
  Datalogger Firmware To Interface To The Calibrator

5
SYSTEM INTEGRATION

• The Entire Data Collection System Is Closely
  Integrated
• The Datalogger And Automatic Gas Calibrator
  Communicate With Each Other Via A Serial
  Interface
• The Interface Was Completely Defined By IPS
• This Interface Results In Preliminary Data
  Flagging At The Monitoring Site Based On What The
  Calibrator Is Doing
• The MeteoStar LEADS Software That Collects Data
  From The Datalogger Uses The Native Datalogger
  Computer-To-Computer Language (cc-sail)
• The Datalogger Keeps Track Of Which Data Records
  Have Been Retrieved
• Simplifies Retrieval Software
• Retrieval Software Queries Datalogger For Latest
  Unsent Data Records
• Data Collection Is Very Forgiving Of Missed Phone
  Calls, Busy Signals, Network Outages, Etc.

6
POLLUTION DATA COLLECTION

• LEADS Uses A Multi-Tier Data Collection System
• Reduces Communications Load On Any Single
  Computer
• Able To Rapidly Reconfigure the Communications
  Network In The Event Of A Hardware Failure At A
  Monitoring Site Or One Of The Regional Hub
  Computers
• Multi-Stage Archival Of Raw Measurements
• Possible To Retrieve Data From Any Of The Archive
  Locations And Reintroduce It Into The System
• The Datalogger Installed At Each Site Is Capable
  Of Storing One To Three Weeks Of Data Based On
  Number Of Instruments
• Regional Hub Computers Are Located Across The
  State That Communicate With The Dataloggers And
  Retrieve Data From Them
• The Regional Hub Computers Forward Collected Data
  To Austin
• The Hub Computers Store Three Months Of Data
• Once Data Arrives In Austin, It Is Decoded,
  Checked For Errors And Stored In A Database - An
  Independent Archive Is Also Created

7
INSIDE A MONITORING SITE
NOx Analyzer
CO Analyzer
Data Logger/Modem
Calibration Valves
SO2 Analyzer
Ozone Analyzer
Gas Calibrator
Clean Air Supply
8
AMBIENT AIR MONITORING SITE
9
WATER MONITORING SITE
10
TCEQ REGIONS

• TCEQ Divides Texas Into Sixteen Regions
• Regions Maintain And Operate Monitoring Sites
• TCEQ Collects Data From Fifteen Of The Sixteen
  Regions
• Data Collection System Based Around These Regions
• Communication Processing Is Distributed Across
  The State
• Green Highlighted Regions Have Hub Computers That
  Service Monitoring Sites
• Yellow Highlighted Regions Have Active Monitoring
  Sites But No Hub Computer
• Sites In Regions Without Hub Computers Are
  Serviced From Other Regions

11
MONITORING SITES

• There Are 177 Active Monitoring Sites - Including
  1 Site In Indiana And 2 In Juarez
• Sites Measure A Wide Variety Of Parameters
• Local Meteorological Data
• Air Pollutants
• Water Parameters
• Anything Else That Can Be Connected To The
  Datalogger
• Specific Parameters Measured At A Site Are
  Determined By
• Site Location
• What Is Located Nearby
• What Is Measured At Other Sites In The Area
• Public Concerns
• Federal Requirements

4
1
33
1
5
2
19
2
18
12
47
12
9
5
6
12
GETTING DATA FROM THE SITES

• Monitoring Sites Serviced By Regional Hub
  Computers
• Communications Is A Balance Of Cost And Data
  Retrieval
• TCEQ Collects Data From Every Site In The System
  Every 15 Minutes
• That Means 96 Connections To Every Site Each Day
• Phone Lines And Modems Are The Least Expensive
• Used When Local (Non-Toll) Calls Can Be Made From
  The Hub Computer To The Site
• Major Metropolitan Areas Have Extended Dialing
  Areas
• Wireless Service Is Nice, But Limited To
  Metropolitan Areas
• Costs More Than Phone Lines But Less Than
  Satellite Service
• Satellite Service Can Be Established Anywhere
• Data Is Generally Available Within 20 - 22
  Minutes From The Time It Is Collected
• Data Available To The Public Within An Hour
• Public Data Has Gone Through Automatic QA/QC
  Checks, But May Change Later Based On Human Review

13
LEADS QA/QC PHILOSOPHY

• Standardize Data Handling
• Insure Consistency Across The Entire Network
• Minimize Subjective Data Quality Judgements
• Automate As Much As Possible
• Instrument Calibrations
• Periodic Instrument Challenges
• Data Quality Flagging
• Rigorously Check The Instruments In Each Station
  And The Calibration System
• Perform Automatic Instrument Challenges On A
  Schedule
• Augment With Periodic Station Visits To Perform
  Independent Audits
• Provide A Set Of Tools For Operators, Validators,
  And Other Data Customers To View Not Only The
  Data But QA/QC Information And System Performance
  As Well

14
LEADS QA/QC IMPLEMENTATION

• TCEQ, IPS, Coastal (Datalogger), And Dasibi (Gas
  Calibrator) Developed Interfaces Between The
  Various Components And Software
• The Gas Calibrator Communicates Its Status To The
  Datalogger Which In Turn Sets The Initial Data
  Quality Flag
• IPS-Developed Software Examines The Flags Coming
  From The Datalogger And Determines Whether Or Not
  To Interpret The Data As Ambient Or Some Type Of
  Calibration Or Instrument Challenge
• TCEQ Specified The Methods For Challenging The
  Various Instruments
• Five-Point Calibration
• Three-Point Span Check
• Span-Zero Check
• Limit Checks
• TCEQ Modified Their Existing QAPP To Incorporate
  LEADS

15
OTHER QA/QC CONSIDERATIONS

• The Introduction Of LEADS Into The TCEQ System
  Did Not Affect Some Things
• Documentation Of Procedures And Processes
• Traceability Of Measurements Back To A Basic
  Standard
• Independent Instrument Calibration And
  Certification
• Independent Station Audits
• Instrument Maintenance And Upkeep Procedures
• Major Impacts Of LEADS
• Ability To Expand Monitoring Network Without
  Increase In Manpower
• Higher Data Return
• High Confidence Level In Quality Of Data
• Ability To Rapidly Furnish Data To The General
  Public

16
QUALITY CHECKS

• There Are Several Types Of Automatic Quality
  Checks
• Calibration Sequences Are Run Periodically Or
  When Equipment Is Changed To Establish A New
  Slope And Intercept
• Span Check Sequences Are Run Weekly To Ascertain
  Whether Or Not A Particular Instrument Is
  Drifting Or Malfunctioning - May Also Set A New
  Intercept
• Span-Zero Check Sequences Are Run Nightly To
  Ascertain Whether Or Not A Particular Instrument
  Is Drifting Or Malfunctioning
• Any Of These Sequences Can Be Automatically
  Scheduled On Either The Dasibi Calibrator Or The
  Zeno Datalogger
• Any Of These Sequences Can Be Manually Initiated
  As Necessary

17
AUTOMATIC TESTS
18
OTHER DATA CHECKS

• Automatic Met Data Checks
• Optional Upper And Lower Limits
• Optional NEG Test
• Optional MUL Test

19
OPERATIONAL DOCUMENTATION

• There Is An Electronic Operator Log On Each
  Datalogger
• Web Interface For Viewing Operator Logs
• Can Also Add Log Entries Via The Web
• Certain Web Pages Also Trigger Automatic Operator
  Log Entries
• Any Change Made By A Data Validator Results In An
  Automatic Validator Note Entry
• Web Interface For Viewing Validator Notes
• Certain Web Pages Also Trigger Automatic
  Validator Note Entries
• There Is A Web-Based Hardware Tracker That Can Be
  Used To Record And Report Instrument Changes At
  Each Site
• There Is A Web-Based Contact Manager That Can Be
  Used To Record Who Operates Sites And/Or
  Particular Instruments As Well As Who Validates
  The Data
• Web-Based Data Reports

20
DATA DISPLAY
21
WEB INTERFACE

• Web Pages Written In Perl And Most Are Dynamic
• Web Pages Run Off A Combination Of Derived Files
  And The Master Pollution Database
• Derived Files Automatically Updated By Background
  Processes
• Derived Files Generally Contain Data Averages Or
  Information That Is Not Stored In The LEADS
  Pollution Database
• Some Derived Files May Take Hours To Update - The
  Derived Files Speed Delivery Of Information To
  Web Users
• Many Of The Web Pages Display System Health And
  Status Information
• Public URL
• http//www.tnrcc.state.tx.us/air/monops
• Internal URL
• http//dsr.tnrcc.state.tx.us
• The Web Pages Grow And Expand Based On User Input
  And Feedback

22
EXTERNAL WEB INTERFACE
23
INTERNAL WEB INTERFACE
24
POLLUTION RETRIEVAL/DISPLAY

• Also Known As Manual Validation
• Lets You Examine 5-Minute Data As Well As Hourly
  Averages
• Data Graphically Displayed
• Up To One Month Of Data Can Be Displayed
• Data Validators Use This Interface To Edit Data
• Requires Specific Privileges In Order To Make Any
  Changes To Data
• Can Change Data Flags
• Can Change Slope And Intercept
• Wind And Pollution Roses Included In This
  Interface

25
MANUAL VALIDATION WINDOW
Pull-Down Menus
Activate Panel
Horizontal Scroll Controls
Vertical Scroll Controls
Time Zone Select
Status Information
26
STATUS INFORMATION

• Information About Each Data Point Is Displayed At
  The Bottom Of The Window
• The Information Displayed Is Dynamic And Is Based
  On The Data Point Located Under The Cursor
• Available Information Includes
• Date And Time Sample Was Collected
• Sample Value In Engineering Units
• EPA Data Flag
• The Slope And Intercept Applied To The Measured
  Voltages
• Whether Or Not The Data Point Has Been Validated
  (Two Indicators)
• Whether Or Not There Are Any Validator Notes
  Attached To The Data
• There Are Also Data Entry Areas For Manually
  Entered Slopes And Intercepts

27
WHATS IN THE DISPLAY
Site ID
Missing Data
Parameter Name
5-Point Cal
3-Point Span
Data Curve
Validated Bar
28
ROSE TYPES

• There Are Four Types Of Roses Available
• Each Of The Roses Summarizes Data Over The Time
  Span You Select
• Data Is Accumulated Into Each Of The Direction
  Bins
• Wind Rose
• This Is A Standard Wind Rose Plot Which
  Illustrates Wind Speed As A Function Of Direction
• Average Concentration Rose
• This Pollution Rose Will Display The Average
  Concentration Of A Particular Pollutant As A
  Function Of Wind Direction
• Maximum Concentration Rose
• This Pollution Rose Will Display The Maximum
  Concentration Of A Particular Pollutant As A
  Function Of Wind Direction
• Percentile Rose
• This Pollution Rose Allows You To Filter
  Pollution Concentrations In Order To Reject
  Abnormally High Or Low Readings - The Smaller The
  Percentile You Select, The Less Data Is Included
  In The Pollution Rose
• Selecting A Percentile Of 100 Is The Same As
  Selecting A Maximum Concentration Rose

29
WIND ROSE
30
POLLUTION ROSE
31
COMMUNICATIONS OPTIONS
32
COMMUNICATION OPTIONS -MODEM-TO-MODEM
RS-232 Serial I/F
RS-232 Serial I/F
Phone Company
Datalogger
Modem
Modem
Hub Computer

• IPS-Developed Software On Hub Computer Initiates
  Phone Calls
• Control Software Written In Perl
• Comms Software Written In C
• Comms Software Driven By Configuration Files
• Phone Number Of Site
• Number Of Records Requested
• Comms Software Communicates Directly With
  Datalogger
• Each Hub Computer Services Multiple Monitoring
  Sites

• Datalogger Comes With A Computer-To-Computer
  Communications Interface
• Maintained By Datalogger Vendor
• Hub Comms Software Written To Use This Interface
• Datalogger Keeps Track Of Which Records Have Been
  Previously Sent To The Hub Computer

33
COMMUNICATION OPTIONS -WIRELESS HANDHELD PC
Phone Company Hub
Ethernet
Hub Computer

• IPS-Developed Software On Hub Computer Initiates
  TCP/IP Network Connection To Handheld PC
• Hub Software Communicates Directly With
  Datalogger Over Network Connection
• Hub Computer Can Use Modems To Contact Other
  Sites
• CDPD Availability Limited To Larger Metropolitan
  Areas

• Handheld PC Runs Windows CE
• Connection To Cellular Network Via Wireless Air
  Card
• Type II PCMCIA Card
• CDPD Protocol (Cellular Digital Packet Data)
• IPS-Developed Software On Handheld PC Establishes
  Network Connection To Hub Computer
• Can Also Use A Regular PC Or Laptop Instead Of
  Handheld

34
COMMUNICATION OPTIONS -Satellite
Provider Hub
RG-59 Cable
RS-232 Serial I/F
Ethernet
Ethernet
Internet
Hub Computer
Datalogger
PC
Satellite Modem

• IPS-Developed Software On PC Establishes A
  Network Connection To Hub Computer
• Approximate 6 Second Time Lag For Satellite
  Transmission

• IPS-Developed Software On Hub Computer Initiates
  TCP/IP Network Connection To Monitoring Site PC
• Functionally Identical To Wireless Solution

35
COMMUNICATION OPTIONS -CELLULAR MODEM
RS-232 Serial I/F
RS-232 Serial I/F
Phone Company
Modem
Hub Computer
Datalogger
Cellular Modem

• IPS-Developed Software On Hub Computer Initiates
  Phone Call To Cellular Modem
• Functionally Identical To Modem-To-Modem Solution
• Still In Test Phase

• Cellular Coverage Limits Usability
• Service Providers Trying To Phase Out This
  Service In Favor Of CDPD And Other Protocols

36
COMMUNICATION OPTIONS -PACKET RADIO MODEM
RS-232 Serial I/F
RS-232 Serial I/F
Packet Radio Modem
Datalogger
Hub Computer
Packet Radio Modem

• IPS-Developed Software On Hub Computer Triggers
  Packet Radio Modem To Initiate A Connection
• Once The Connection Is Made, This Is Functionally
  Identical To The Modem-To-Modem Solution
• Still In Test Phase

• Packet Radio Modem Operating At 5 Watts Output
• Designed For Line-Of-Sight Installations
• Highly Dependent Upon Terrain And Other Obstacles
• First Installations Will Be Approximately 5 Miles
• Other Configurations Can Use Repeaters Or Relays
  To Extend Distance
• Increased Radio Power Will Also Extend Distance

37
AUTOMATIC QUALITY CHECKS
38
QC SEQUENCES

• Sequences Are Composed Of Levels
• A Level Consists Of A Set Concentration From The
  DASIBI 5008 Calibrator Introduced Into A Monitor
  For A Set Number Of 5-Minute Sample Periods
• Each Level Is Assigned A Letter Code (M, R, S, T,
  Or G) By The Datalogger -- These Correspond To
  Various Percentages Of The Instrument Full-Scale
• A Set Number Of 5-Minute Samples (Usually 1 - 3)
  In Each Level Are Allowed For Instrument
  Stabilization
• The Remaining Samples (Usually 3 - 4) Are
  Processed By LEADS.

39
FIVE-POINT CALIBRATION
40
THREE-POINT SPAN CHECK
41
TWO-POINT SPAN-ZERO CHECK
42
POSSIBLE QC CHECK FLAGS

• Incomplete
• The Data Did Not Arrive In The Proper Sequence Or
  There Were Either Too Few Or Too Many Samples At
  Each Level
• Invalid
• The Calibrator Did Not Deliver The Challenge Gas
  At Correctly Spaced Intervals
• Failed
• One Or More Of The Automatic QC Checks Exceeded A
  Pre-Set Failure Limit -- Data Is Automatically
  Rejected Accordingly
• Warning
• One Or More Of The Automatic QC Checks Exceeded A
  Pre-Set Warning Limit -- The Calibration Or Span
  Check Is Still Valid And Is Still Used
• Passed
• All Automatic QC Checks Were Passed

43
AUTOMATIC TESTS

• Completeness Test
• Tests To See That Data Arrives In Correct Order
  And The Correct Number Of Updates For Each Level
  Are Present
• Voltage Outlier
• Tests Are Based On The Measured Voltages From The
  Instrument
• Tests To See That No Single Voltage Measurement
  For A Level Is More Than An Allowed Deviation
  From The Average Voltage
• Concentration Outlier
• Tests To See That No Single Concentration
  Measurement For A Level Is More Than An Allowed
  Deviation From The Average Concentration
• This Test Was Specifically Designed To Catch
  Calibration System Problems

44
AUTOMATIC TESTS (Cont.)

• NO2 Concentration Outlier
• Based Entirely On NO Channel Performance On
  Previous Non-Titration Level (M-Level) And
  Current Titration Level (M-Level) -- NO2
  Voltages Are Not Part Of This Test
• Tests To See That No Single Concentration
  Measurement For A Level Is More Than An Allowed
  Deviation From The Average Concentration
• Concentration Spacing
• Tests To See That The Calibrator Is Delivering
  Evenly Spaced Concentrations Throughout The
  Instruments Range
• No Concentration For A Level Can Be More Than An
  Allowed Deviation From The Ideal Concentration

45
AUTOMATIC TESTS (Cont.)

• NO2 Concentration Spacing
• Based Entirely On NO Channel Performance On
  Previous Non-Titration Level (M-Level) And
  Current Titration Level (M-Level) -- Differs
  From Normal Concentration Spacing Test
• No Concentration For A Level Can Be More Than An
  Allowed Deviation From The Ideal Concentration
• Slope
• Tests To See That The Instrument Is Responding
  Correctly. All Instrument Responses Should Be
  Linear And The Calculated Slope Should Be Near
  The Ideal
• Ideal Slopes
• CO 20
• Other Pollutants 1000 (0 - 1000 ppb Instrument
  Range)
• Other Pollutants 2000 (0 - 500 ppb Instrument
  Range)

46
AUTOMATIC TESTS (Cont.)

• Intercept
• Tests To See That The Instrument Is Responding
  Correctly And That The Calculated Intercept Is
  Near The Ideal
• Ideal Intercept 0
• Precision
• Tests To See That The Instrument Has A Stable,
  Linear Response
• The Voltage Average For Each Level Is Compared To
  The Calculated Regression Line And Must Be Within
  Limits
• Linearity
• Tests To See That The Instrument Is Holding A
  Stable, Linear Response
• The Voltage Average For The T-Level Is Compared
  To The Line Drawn Between The M-Level And The
  G-Level And Must Be Within Limits

47
AUTOMATIC TESTS (Cont.)

• Zero
• Tests To See That The Instrument Calibration Is
  Holding Since The Last Good Calibration Or Span
• The Voltage Average For The G-Level Is Compared
  To The G-Level From The Last Good Calibration Or
  Span And Must Be Within Limits
• Span
• Tests To See That The Instrument Calibration Is
  Holding Since The Last Good Calibration Or Span
• The Voltage Average For The M-Level Is Compared
  To The M-Level From The Last Good Calibration Or
  Span And Must Be Within Limits

48
AUTOMATIC TESTS (Cont.)

• NOx Balance Test
• Tests The Electronic And Flow Balance Between The
  NO and NOx Channels. Compares The Ratio Of The
  Average NO2 Span Response To The Average NO Span
  Response During Non-Titration Steps. Sensitive
  To NO2 Impurities In NO Cylinder. Impurities
  Give Positive Errors.
• Only Affects The Validity Of The NO2 Channel
  Calibration
• NO2 Converter Efficiency Test
• Tests The Efficiency Of the NO2 ? NO Converter
• Affects The Validity Of The NO2 And NOx Channel
  Calibrations
• H2S Converter Efficiency Test
• Tests The Efficiency Of the H2S ? SO2 Converter
• Based On The T-Level And The T1-Level. Test Is
  Sensitive To Agreement Of The Calibrations Of The
  H2S Cylinder To The SO2 Cylinder.

49
AUTOMATIC TESTS (Cont.)

• SO2 Scrubber Efficiency Test
• Tests The Efficiency Of the SO2 Scrubber

50
TEST LIMITS

• All QC Tests Performed On Calibration,Span Check,
  Or Span-Zero Data Have Both Warning And Failure
  Limits Except The Outlier Tests
• Each Outlier Test Uses Only One Limit But The
  Test Is Repeated If An Outlier Is Detected.
• Each Warning Limit Is Chosen Statistically To
  Represent The 3rd Standard Deviation Value About
  The Mean Error Of A Test
• There Should Be Only A 0.27 Probability Of
  Exceeding A Warning Limit If The Monitoring
  System Is Working Properly
• Each Failure Limit Is Intended To Represent The
  Maximum Error That Will Be Tolerated Without
  Invalidation Of The Affected Data

51
DATA VALIDATION RULES

• Automatic Invalidation Of Data Is Based Only On
  Whether A QC Test Passes Or Fails
• Warnings Are Not Considered In This Processing
• Failure Of A Concentration Outlier Test Or A
  Concentration Spacing Test Indicates A Problem
  With The Calibration System But Not With The
  Monitor (Except NO2)
• The Calibration Or Span Check Event Involved Is
  Considered Invalid And The Ambient Pollution Data
  Is Unaffected
• In The Event Of A Test Failure, Data Is
  Automatically Rejected (Flagged LIM)
• To Stop Rejecting Data Forward, A Successful
  Five-Point Calibration Must Be Run
• Data Is Rejected Back To The Last Successful
  Calibration, Span Check, Or Span-Zero Check

52
SPAN SOURCE AUDITS

• Automatic Quality Control Is Augmented At TCEQ By
  Quarterly Manual Span Source Audits
• Span Source Audits Evaluate The Accuracy Of The
  CAMS DASIBI 5008 Calibrator
• Corrective Action Is Required If Audit Limits Are
  Exceeded
• If the Audit Passes, Then the Pollutant Monitor
  Should Be Adjusted to Agree With the M-level
  Concentration Produced by the CAMS DASIBI
  Calibrator
• This Sets the Slope of the Monitors Response to
  the Ideal
• If a Slope Test Warning Is Reported Thereafter,
  Then Action Must Be Taken to Determine If the
  Error Was Caused by Monitor Drift or Span Source
  Drift
• A Span Source Audit May Be Needed If There Are No
  Obvious Instrument Problems

53
OTHER DATA CHECKS
54
UPPER AND LOWER LIMITS

• Limits Can Be Defined For A Specific Instrument
  At A Site
• Any Parameter Can Have Limit Checks Defined
• Can Define Either An Upper Or Lower Limit Or Both
• System Automatically Flags Any Data That Falls
  Outside Specified Limits As LIM
• LIM Data Is Automatically Excluded From Averaging
  And Automatic Alerts

55
AUTOMATIC MET CHECKS
56
NEG AND MUL TESTS

• Only Performed On Hourly Averages - Does Not
  Affect Five-Minute Averages
• NEG Test Can Be Defined For Any Parameter That Is
  Rolled Up Into An Hourly Average
• Test Will Automatically Set Slightly Negative
  Averages To 0.0
• If Data Is Too Negative, It Will Be Flagged NEG
• Data Flagged NEG Is Automatically Excluded By
  Other Displays And Analysis Algorithms
• MUL Test Designed Specifically For NOx Family
• Tests To Make Sure That Sum Of NO And NO2
  Measurements Are Consistent With NOx Measurement
• If Data Is Not Consistent, All Three Parameters
  Will Be Flagged MUL
• Data Flagged MUL Is Automatically Excluded By
  Other Displays And Analysis Algorithms