3A-906-R-95-001
   U.S. ENVIRONMENTAL 1445 ROSS AVENUE EPA-906-R-95-001
   PROTECTION AGENCY SUITE 1200      MARCH 1995
   REGION 6         DALLAS, TX 75202
    oEPA
        WINTER SEASON AIR POLLUTION IN

           EL PASO - CIUDAD JUAREZ
                    ENVIRONMENTAL
                     PROTECTION
                      AGENCY
                    DALLAS, TEXAS

                         ARY
                      a te

-------
Online Library System (OLS) | Libraries US EPA
                                                          Page 1 of2
  http://cave.epa.gov/cgi/nph-bwcgis/BASIS/ncat/pub/ncat/DDW?W%3DREPNUM+PH+IS+%279o6/r-
         ^ORDER+BY+YR/Descend%26M%3Di%2$|^^agffiai<3^fifi^^y^ftH?fl3iJi4th,20o6.
                !»
                            L.     Display Kecoros as biDiiograpny     j [  item status
 RECORD NUMBER: 1 OF 1
  Main Title    Winter season air pollution in El Paso - Ciudad Juarez
 , Author
 Einfeld, Wayne
  Publisher
 U.S. Environmental Protection Agency, Region VI, Air Pesticides and Toxics
 Division,
 | Year
 1 Published

 ! Report
 Number

 ^OCLC
 Number
 t~""~	"~~—"""""•""
 j Subject
 ; Added Ent
(1995

 EPA-9o6-R-95-ooi;DW899334i9Oi;SAND95-O273
 33147100
 Air~Pollution~Texas~El Paso ; Air—Pollution—Mexico—Ciudad Juarez
 i Collation    j x, 119, {31} p.: ill., map ; 28 cm.
  Holdings
  Notes
  Author
  Added Ent
  LIBRARY    CALL NUMBER          LOCATION
  EMAD       EPA-906-R-95-001   Region 6 Library/Dallas,TX

 "March 1995." "EPA-9O6-R-95-OO1." "This work was conducted under EPA
 interagency agreement number DW89933419O1 with the U.S. Dept. of
i Energy." "SAND95-O273." Includes bibliographical references.
; Church, Hugh W.
  Corp Au
  Added Ent
  Place
 United States. Environmental Protection Agency. Region VI. Air, Pesticides
 and Toxics Division.; Sandia National Laboratories.; United States.
1 Environmental Protection Agency. Region VI. Air, Pesticides and Toxics
: Division.; Sandia National Laboratories.
; Dallas, TX:
http://cave.epa.gov/cgi/nph-bwcgis/BASIS/ncat/pub/ncat/DDW?W%3DREPNUM+PH+IS... 9/14/2006

-------
     WINTER SEASON AIR POLLUTION IN EL PASO - CIUDAD JUAREZ


                            March 1995
                                by
                          Wayne Einfeld
                          Hugh w.  Church
            Environmental Characterization Department
                   Sandia National Laboratories
               James W. Yarbrough, Project Manager
                           EPA-Region 6
                Air, Pesticides & Toxics Division
This work was conducted under EPA Interagency Agreement number
DW8993341901 with the U.S. Department of Energy.

-------
                                       Abstract

This report summarizes a number of research efforts completed over the past 20 years in the El
Paso del Norte region to characterize pollution sources and air quality trends. The El Paso del
Norte region encompasses the cities of El Paso, Texas and Ciudad Juarez, Chihuahua and is
representative of many US-Mexico border communities that are facing important air quality issues
as population growth and industrialization of Mexican border communities continue.  Special
attention is given to a group of studies carried out under special US Congressional funding and
administered by the US Environmental Protection Agency.  Many of these studies were fielded
within the last several years to develop a better understanding of air pollution sources and trends
in this typical border community. Summary findings from a wide range of studies dealing with
such issues as the temporal and spatial distribution of pollutants and pollution potential from
both stationary and mobile sources in both cities are presented.  Particular emphasis is given to
a recent study in El Paso-Ciudad Juarez that focussed on winter season PM10 pollution in El Paso-
Ciudad Juarez. Preliminary estimates from this short-term study reveal that biomass combustion
products  and crustal material are significant components of winter season  PM10 in this
international border community.

-------
                                  Acknowledgements

This report summarizes the efforts of many federal,  state and local agencies,  private-sector
contractors, universities and national laboratories who have collectively produced a sizeable body
of information dealing with air quality issues in the El Paso del Norte region.  In preparing this
review, the authors have attempted to give appropriate credit to participants through the use of
numerous citations in the text.

While personnel from the Texas Natural Resource and Conservation Commission (formerly the
Texas Air Control Board) and Sandia National Laboratories took the lead role in the design and
data analysis associated with the short-term winter season PM10 study, the diligent efforts of many
individuals from the organizations listed below enabled a satisfactory completion of the study.

       El Paso City-County Health Department
       US Environmental Protection Agency - Region 6
       US Environmental Protection Agency - Atmospheric Research and Exposure Assessment
              Laboratory
       Secretariat of Social Development of Mexico
       City of Juarez Health Department
       Science Applications International Corporation
       Radian Corporation
       Lawrence Berkeley Laboratory
       Sunset Laboratory

A number of Sandia and contract personnel made important contributions to the project. Gary
Brown, Mark Ivey and Monty Apple played  important roles in meteorological instrumentation
setup and maintenance along with field data collection during the short-term study.  Chris
Erickson and Laura McCarty were instrumental in making the necessary modifications to the
Diagnostic Wind Field Model from the Urban Airshed Model such that it would run on a Sandia
computer. They also designed and implemented the post-processor for graphical display of the
computed results.

This work was sponsored  by the US Environmental Protection Agency-Region 6.  Mr Jim
Yarbrough of the EPA-Region 6 Air Programs Branch was the project officer. A special recognition
goes Mr. Yarbrough who  provided  the  authors with much of the background information
describing early air pollution studies in the region.

Sandia  National   Laboratories  performed  this  work  under  Interagency  Agreement  No.
DW89933419-01-4. Sandia National Laboratories is a multi-program laboratory operated for the
US Department of Energy under Contract No. DE-AC04-94AL85000.
                                          ii

-------
                         List of Abbreviations and Acronyms


AIRS	Aerometric Information Retrieval System
ASARCO. . .  American Smelting and Refining Company
ASL	US Army Atmospheric Sciences Laboratory
CAA	Clean Air Act
CAMS	Continuous Air Monitoring Station
CMB	Chemical Mass Balance
CO	Carbon Monoxide
CPM	Coarse(>2.5 and <10 micrometer aerodynamic diameter) Particulate Matter
DWM	Diagnostic Wind Field Model (from the Urban Airshed Model)
EPA-EMSL  . Environmental Protection Agency - Environmental Monitoring Systems Laboratory
EPCCHD .  . El Paso City County Health Department
FAA	Federal Aviation Administration
FEMAP .... Federation of Private Health and Community Development Associations
FPM	Fine (<2.5 micrometer aerodynamic diameter) Particulate Matter
GC-MS .... Gas Chromatography-Mass Spectrometry
HC	Hydrocarbons
Hivol	High volume air sampler
IAQMD .... International Air Quality Management District
I/M	Inspection and Maintenance
MST	Mountain standard time
NAAQS  . .  . National Ambient Air Quality Standards
NMEID .... New Mexico Environment Improvement Division
NWS	National Weather Service
PAH	Polycyclic Aromatic Hydrocarbons
PC	Personal Computer
PM10	Particulate matter less than 10 micrometers aerodynamic diameter
PM15	Particulate matter less than 15 micrometers aerodynamic diameter
PUF	Polyurethane foam
RVP	Reid Vapor Pressure
SAI	Science Applications Incorporated
SCERP . . .   Southwest Center for Environmental Research Policy
SEDESOL  . Federal Secretariat of Social Development
SEDUE . .  . Secretariat of Urban Development
TACB  .... Texas Air Control Board
THC	Total hydrocarbons
TNRCC .... Texas Natural Resource and Conservation Commission
TSP	Total Suspended Particulate Matter
UTEP  .... University of Texas at El Paso
UTM	Universal Transverse Mercator
VMT	Vehicle Miles Traveled
VOC	Volatile organic compounds
XRF	X-ray Fluorescence
                                         iii

-------
                                  Table of Contents

Executive Summary	   1

1.0 Introduction	   5

2.0 Background Information 	   7
       2.1  Local Topography	   7
       2.2  Local Winter-Season Climate  	   7
       2.3  Regional Air Quality Monitoring Programs	   10
       2.4  El Paso del Norte Airshed Pollution Historical Trends  	   10

3.0 Pre-1989 El Paso-Cd. Juarez Air Quality Studies	   15
       3.1  Carbon Monoxide Studies ~ University of Texas at El Paso. 1983 	   15
       3.2  Inhalable Partlculate Matter Assessment — Environmental Research and
             Technology Inc.. 1983	   16
       3.3  Air Quality Database Study — Radian Corporation, 1983	   17
       3.4  El Paso Quantitative Microscopy Study  — Energy Technology Consultants,
             1983  	   18

4.0 Regional Studies Conducted after 1988 	   21
       4.1  Airborne Lidar Study -- EPA Environmental Monitoring Systems Laboratory,
             1989  	   21
       4.2  Upper Air Winds and Visibility Study  - University of Texas at El Paso,
             1989  	   23
       4.3  Saturation PM10 Study - EPA Region 6,  1989	   24
       4.4  Cd.  Juarez Gasoline Vapor Pressure Study — SEDESOL and others, 1990 ...   27
       4.5  Juarez Vehicle Tampering Survey ~ Colorado State University,  1990	   27
       4.6  Cd.  Juarez Emissions Inventory — Alliance Technologies and others,  1990 ...   29
       4.7  Short-term Winter Partlculate Study ~ Texas Air Control Board, Sandla
             National Laboratories, 1990  	   29
       4.8   PM10 Modeling  Plan Assessment and Recommendations  —  Systems
             Application International, 1991	   29
       4.9   PM10 SIP  Development  — Texas  Natural  Resource and Conservation
             Commission,  1991  	   32
       4.10 Cd. Juarez Industrial Emissions Study  -- EPA-SEDESOL, 1992-93	   32
       4.11 Vehicle Emissions Remote Sensing Study ~ University of Denver, 1993  ....   33
       4.12 MOBILE5 Revisions — Energy and Environmental Analysis Inc., 1993	   37
       4.13 Cd. Juarez Vehicle Fleet Characterization — Texas Transportation Institute,
             1993  	   37
       4.14 Cd. Juarez Brickmaker Study — El Paso Natural Gas Co./FEMAP, 1993 ...   40
       4.15 Pollutant Emissions from Residential Heaters — University of Utah,  1993 . .   41
       4.16 Upper Air Wind and Temperature Data Collection — University of Texas at
             El Paso, 1993	   42
       4.17 Oxygenated Fuel Use in El Paso-Cd. Juarez -- SEDSOL and others,  1993 . .   42
       4.18 Other Activities - Paso del Norte Air Quality Task Force, 1993  	   42
       4.19 Comparison of Vehicle Emissions Inspection and Maintenance  Programs in
             Cd. Juarez and El Paso ~ Paso del Norte Task Force, 1994	   43
       4.20 Technology Transfer Session with I/M Technicians — Colorado State
             University, 1994	   46
                                         iv

-------
5.0 Short-term Winter Season PM10 Study	„	„  49
       5.1  Study Purpose and Scope	  49
       5.2  Measurement Techniques	  49
       5.3  Site Descriptions 	  50
             5.3.1  Site selection process (50)
             5.3.2  Air sampling enhancement at existing sites (50)
             5.3.3  General site descriptions (50)
       5.4  Study Results	  55
             5.4.1  Meteorological Measurement Results (55)
             5.4.2  Diagnostic Wind Field Model Runs (72)
             5.4.3  Summary of FPM, CPM and PM10 Measurements (81)
             5.4.4  Elemental Species Measurements (91)
             5.4.5  Aerosol and Semi-volatile Carbon Measurements (94)
             5.4.6  Other Measurements (95)
             5.4.7  Temporal PM10, Nephelometer and CO Data Analysis (96)
             5.4.8  Receptor Modeling Analysis (98)

6.0 Summary Conclusions and a Forward Look	 113
       6.1  Summary Findings	 113
       6.2  New Technologies	 114
             6.2.1  Lidar Systems (114)
             6.2.2  Model Development (116)
             6.2.3  Complementary Technologies (116)

7.0 References	 117

Appendix A Diagnostic Wind Model Results

-------
                                   List of Tables

Table 1 Fingerprint/Ratio Source Apportionment Results for TSP at Various Monitors in
       El Paso	  20

Table 2 Juarez Vehicular Fleet Tampering Survey Summary Statistics  	  28

Table 3 Summary CO and HC Tailpipe Emissions from the University of Denver Remote
       Sensing Study	  35

Table 4 Average Speed (MPH) by Roadway Class and Time of Day	  37

Table 5 Overall Cd. Juarez VMT Mix  	  38

Table 6 Registration Information by Vehicle Age	  39

Table 7 Cd. Juarez and El Paso Vehicle Use Comparisons  	  40

Table 8 Gas and Total Hydrocarbon Emissions Factors for Selected Waste Wood Fuels . .  41

Table 9 Air Pollution Parameters Measured in the 1990 Winter Season Study	  49

Table 10 Winter Season Study Site Locations and Monitoring Equipment	  51

Table 11  Tethersonde and Radiosonde Soundings at the Chamizal Site on Dec 8-10,
       1990	  71

Table 12 Summary of 12-hr FPM, CPM and PM10 Measurements	  81

Table 13  Estimated Number of Exceedences of the US PM10 Standard Based on  a
       Combination of 12-hour Dichotomous Sampler Results	  87

Table 14 Summary FPM Elemental Analysis Data for a Cd. Juarez and El Paso Monitoring
       Site	  92

Table 15  Summary CPM Elemental Analysis Data for a Cd. Juarez and El Paso
       Monitoring Site	  93

Table 16 Summary Aerosol Carbon Mass Fractions by Sampling Site	  94

Table 17  Goodness of Fit Results for a 5-Factor Solution to the Normalized FPM Data
       Set	  99

Table 18 Normalized Fine Particle Factor Analysis Results  	101

Table 19  Goodness of Fit Results for a 5-Factor Solution of the Normalized CPM Data
       Set	 102

Table 20 Normalized Coarse Particle Factor Analysis Results	  103

Table 21 Summary Source Strength Estimates for Fine and Coarse Aerosol Fractions  ...  110
                                        vi

-------
                                   List of Figures

Figure 1  A contour map of the El Paso del Norte region. The grid shown is 30 km on a
      side and the axis units are In meters	  8
Figure 2  A perspective plot of the El Paso del Norte region as viewed from the southeast.
      The overall square grid domain is 30 km	  9
Figure 3 The four highest 24-hour PM-10 levels measured annually at the El Paso Tillman
      site from 1985 through 1993	  11
Figure 4  The two highest 8-hour CO averages measured annually at the El Paso Tillman
      site from 1985 through 1993	  11
Figure 5  The four highest 24-hour PM-10 levels measured annually at the Cd. Juarez
      Advance Transformer site from 1990 through 1993	  12
Figure 6  The two highest 8-hour CO averages measured annually at the Cd. Juarez
      Advance Transformer and Tecnologico sites from 1990 through 1993	  12
Figure 7  Fourth-quarter 1992, 24-hour average PM-10 measurements in Cd. Juarez and
      El Paso showing interpolated values over the region as contour lines	  14
Figure 8  A graph showing airborne flight tracks (circles) on the morning of February 23
      with near ground-level aerosol backscatter intensity, interpolated from the lidar
      data, shown as contour lines [from McElroy,  1990]	  22
Figure 9   A contour plot showing PM-10 values interpolated from measurements on
      December 11, 1989 during the Saturation PM-10 Study (from Kemp, 1990]	  26
Figure 10 Frequency histogram showing RVP measurements for 41 samples collected in
      Cd. Juarez [from Yarbrough, 19941	  27
Figure 11 Decile emissions in %CO for vehicles measured in El Paso and Cd. Juarez [from
      Stedman, 1993B]	  36
Figure 12 Decile emissions for HC in %propane for vehicles measured In El Paso and Cd.
      Juarez [from Stedman, 1993B]	  36
Figure 12  Estimate  of daily Cd. Juarez  traffic activity by hour of the day [from TTI,
       19931	  38
Figure 14 Population trends in El Paso and Cd. Juarez [from Rincon, 1994]	  44
Figure 15 Vehicle CO and hydrocarbon emission standards established In El Paso and
      Cd. Juarez I/M programs [from Rincon, 1994]	  45
Figure 16 Vehicle failure rate by vehicle year encountered during 1993 In the El Paso I/M
      program [data from Rincon, 1994]	  46
Figure 17 Topographical map of the El Paso del Norte region showing the sampling sites
       used In the study. See Table 10 for key to site abbreviations	  52
Figure 18 Wind speed and direction for December 7-11 at the Sun Metro site	  57
Figure 19 Wind speed and direction for December 7-11 at the Tecnologico site	  57
Figure 20 Wind speed and direction for December 7-11 at the Advance Transformer site.
         	  58
Figure 21 Temperature/humidity (A) and wind speed/wind direction (B) for December 8
       at the Chamizal site	  59
Figure 22 Sodar measured winds on December 8 at the Chamizal site	  60
Figure 23 Chamizal temperature sounding data from the morning of December 8 from
       radiosonde (A) and tethersonde (B)	  61
Figure 24 Scattering coefficient profiles during balloon-nephelometer ascent (upper) and
       descent (lower) during 0708-0732  hours on December 8 at the Chamizal site.  ...  63
Figure 25 Scattering coefficient profiles during 0753-0824  hours at the Chamizal site. . .  64
Figure 26 Scattering coefficient profiles during 0805-0920 hours on December 8 at the
       Chamizal site	  65
                                         vil

-------
Figure 27  Scattering coefficient profiles during 1045-1110 hours on December 8 at the
      Chamizal site	   66
Figure 28  Scattering coefficient profiles during 1235-1305 hours on December 8 at the
      Chamizal site	   67
Figure 29  Scattering coefficient profiles during 1408-1440 hours on December 8 at the
      Chamizal site	   68
Figure 30  Scattering coefficient profiles during 1615-1642 hours on December 8 at the
      Chamizal site	   69
Figure 31  Scattering coefficient profiles during 1718-1726 hours on December 8 at the
      Chamizal site	   70
Figure 32 Early morning (0400 hours) ground-level (13m agl) wind fields predicted by the
      Diagnostic Wind Field model on December 8	   75
Figure 33 Early morning (0400 hours) upper-level (1000 m agl) wind field predicted by the
      Diagnostic Wind Field model for December 8	   76
Figure 34  Midday  (1200 hours) ground-level (13 m agl) wind field predicted by the
      Diagnostic Wind Field model for December 8	   77
Figure 35  Midday (1200 hours)  upper-level (1000 m agl)  wind field predicted by the
      Diagnostic Wind Field model for December 8	   78
Figure 36 Afternoon (1600 hours) ground-level (13 m agl) wind field predicted by the
      Diagnostic Wind Field model for December 8	   79
Figure 37 Afternoon (1600 hours) upper-level (1000 m agl) wind field predicted by the
      Diagnostic Wind Field model for December 8	   80
Figure 38  A stacked barplot showing FPM and CPM for all sampling periods  at the
      Advance Transformer site	   82
Figure 39  A stacked barplot showing FPM and CPM for all sampling periods  at the
      CAMS6 site [note scale change]	   82
Figure 40  A stacked bar plot showing FPM and CPM for all sampling periods  at the
       Chamizal site	   83
Figure 41  A stacked barplot showing FPM and CPM for all sampling periods at the Sun
      Metro site [note scale change]	   83
Figure 42  A stacked barplot showing FPM and CPM for all sampling periods  at the
      Tecnologico site	   84
Figure 43 A scatterplot of FPM and PM-10 for all sampling sites and periods	   85
Figure 44 A scatterplot of CPM and PM-10 for all sampling sites and periods	   85
Figure 45 A scatterplot of FPM and CPM for all sampling sites and periods	   86
Figure 46 A contour plot of PM-10 levels interpolated from measurements at 15 sites on
       December 7 [from Dattner, 1993]	   88
Figure 47 A contour plot showing PM-10 levels interpolated from measurements at 15
       sites on December 13 [from Dattner, 1993]	   89
Figure 48 A contour plot showing PM-10 levels interpolated from measurements at 15
       sites on December 19 [from Dattner, 19931	   90
Figure 49  Total polycyclic aromatic hydrocarbon and selected species concentrations
      measured at the CAMS6 site during a portion of the study interval [from Dattner,
       1993]	   95
Figure 50 A graph of aerosol light scattering, CO and PM-10 measurements from the
       continuous monitors at the Chamizal monitoring site [from Dattner, 1993]	   97
Figure 51  Estimates of crustal source contribution to FPM for all sites and sampling
      periods	  106
Figure 52  Estimates of crustal source contribution to CPM for all sites and sampling
       periods	  106
Figure 53 Estimates of smelter source contribution to FPM for all sites and all sampling
      periods	  107

                                         viii

-------
Figure 54 Estimates of smelter source contribution to CPM for all sites and sampling
       periods	 107
Figure 55 Estimates of vehicular source contribution to FPM for all sites and sampling
       periods	 108
Figure 56 Estimates of the vehicular source contribution to CPM for all sampling sites
       and periods	 108
Figure 57 Estimates of biomass combustion source contribution to FPM for all sampling
       sites and periods	 109
Figure 58 Estimates of biomass combustion source contribution to CPM for all sampling
       sites and periods	 109
Figure 59 Estimates of source category contribution to total PM-10  during the winter
       season El Paso-Cd. Juarez	110
                                          ix

-------
                                  Executive Summary

The western portion of the  state of Texas meets the northern part of the Mexican state of
Chihuahua and the southern boundary of the state of New Mexico at a strategic mountain pass,
referred to by Spanish explorers as "El Paso del Norte" through which the Rio Grande courses its
way on its long journey from the high mountains in the state of Colorado to the Gulf of Mexico
many miles distant. This important trade route and center of commerce has been inhabited for
over 400 years and currently accommodates a population of nearly 1.4 million people on both
sides of the international border.  The twin cities of Ciudad Juarez (City of Juarez, hereafter Cd.
Juarez), Chihuahua and El Paso, Texas share a common international airshed that is partially
surrounded by mountains to the north and west of the population center. Both cities are
experiencing growth however Cd. Juarez, in particular, is experiencing striking growth as the
country of Mexico and the state of Chihuahua makes a transition from an agrarian economy to
an industrial economy centered in large urban zones such as Cd. Juarez.

While the country of Mexico and the state of Chihuahua are in  principal committed to the
preservation and maintenance of clean air resources, at a practical  level the degree of pollutant
source control on the  Mexican side  of the  border lags  considerably behind  air pollution
characterization and control  programs in place on the US side. The existence of sister cities on
either side of the border poses special problems for those cities faced with compliance with the US
and Mexican clean-air statutes. Local, state and federal agencies have struggled for many years
for a  mutually satisfactory  approach to the problem.  The Cd. Juarez-El  Paso  airshed has
historically exceeded US Clean Air Act National Ambient Air Quality Standards for such pollutants
as PM10 (particles less than 10 pm diameter), CO (carbon monoxide) and ozone.  In an effort to
address this problem, air pollution officials in the State of Texas and the City of El Paso have
lobbied the US EPA for many years to provide a Federal mechanism by which air pollution
characterization and control strategy development could be achieved  by international cooperation
in the area. In recent years,  the Mexican Federal Secretariat of Social Development (SEDESOL)
and its predecessor the Secretariat of Urban Development and Ecology (SEDUE) the agency that
is analogous to the US EPA, has taken a more active interest in U.S-Mexico border air pollution
problems. Factors such as these figured prominently in the successful negotiation and approval
of the so-called Annex V that is part of the 1983 US-Mexico Border Environmental Agreement
dealing in particular with air pollution issues along the US-Mexico  border.  Finalization of this
agreement in 1989, and the  appropriation  of funds by the US Congress, also in 1989, laid the
groundwork for a series of joint US-Mexico air pollution studies designed to more fully understand
pollution sources and their impacts on both sides of the border.

A number of research efforts to characterize pollution sources and air quality trends completed
over the past 20 years in the El Paso del Norte region are summarized in this report. Early studies
concentrated on definition of stationary sources such as local copper smelters and petroleum
refineries. Later studies have focussed upon PM10 pollution in particular since the region shows
non-attainment for current US PM10 standards when both US and Mexican sources are taken into
account. Summary findings from the collection of studies reviewed  in this report are more fully
summarized below.

       Industrial stationary sources do not contribute significantly to airborne particulate
       matter in the El Paso del Norte region -  Studies in the  1980's focusing on total
       suspended particulate matter and more recent studies dealing with PM10 both reveal that
       large industrial  stationary sources  such as copper smelters in the region are minor
       contributors to winter season airborne particulate matter.

-------
Winter season PM10 levels are highest in the Cd. Juarez-El Paso downtown areas and
in general show a concentration gradient increasing toward Mexico -  A number of
studies, among them the EPA-EMSL airborne lidar study, EPA-6 saturation PM10 study,
and the short term winter PM10 study, all reveal higher PM10 levels as one moves toward
Cd. Juarez. Further indirect evidence for this concentration gradient comes by way of the
TNRCC PM10 SIP analysis which reveals that no exceedences of PM10 are predicted in El
Paso if only El Paso particulate matter sources are taken into account.  In actual studies,
however, PM10 concentration levels in excess of air quality standards are encountered on
both sides of the border.

Emissions from the average  vehicle in Cd. Juarez are about three-fold higher than
from the average vehicle in El Paso - The University of Denver remote sensing studies
for tailpipe CO and hydrocarbons show that while the amount of pollutants emitted from
a high-polluting car is the same in both Cd. Juarez and El Paso, there are more of these
high-polluting cars operating in Cd. Juarez than in El Paso.  Other studies undertaken to
characterize the age of the vehicular fleet in Cd. Juarez reveal that the average age of the
fleet is older in Cd. Juarez than in El Paso.  This observation is consistent with the
measured higher tailpipe emissions in Cd.  Juarez.

Vehicle miles traveled in El Paso are about three-fold higher than in Cd. Juarez -
Results from a series of studies by the Texas Transportation Institute on vehicle usage in
Cd. Juarez reveal about 3.4 million anuual VMT in Cd. Juarez as compared to 9.9 million
annual VMT in El Paso. Per capita mileage in El Paso is about six-fold higher than in Cd.
Juarez revealing very different vehicle usage patterns in the two cities. Should Cd. Juarez
residents adopt US driving habits, vehicular emissions could significantly increase in the
Paso del Norte region.

Winter stagnation events and complex terrain significantly limit pollutant dilution
within the region -Meteorological data taken during the PM10 short term study confirm
the presence  of very shallow vertical mixing heights in the evening and early morning
hours of the winter season.  Wind flow at various meteorological monitoring locations
throughout the El Paso del Norte region reveal local terrain influence during winter
stagnation periods.  Wind fields predicted by the Diagnostic Wind Field Model show
general agreement with observed  winds. A rigorous comparison was not carried out
between measured and  predicted wind flow as a part of this study however.  In some test
cases abnormal discontinuities in the predicted wind fields were observed suggesting that
optimization of the model may be required to obtain representative results.

Winter  season PM10 in the Paso del Norte region reaches levels in excess of the
National Ambient Air Quality Standard - Analyses completed by the TNRCC as a part
of the short term winter PM10 study revealed numerous exceedences of the 24-hour PM10
standard at both Cd. Juarez and El Paso sites during the 1990-91 winter season study
period.  In some cases, average PM10 levels three-fold higher than the US PMi0 air quality
standard were observed.

Aerosol carbon is a major constituent of fine, coarse and PM10 aerosol in the region -
 Carbon species measurements of collected particulate matter reveal aerosol  carbon
composition much like that observed in other urban areas. High levels of both elemental
and volatile carbon are  observed and originate from vehicular, biomass combustion and
other local combustion sources. Limited measurements of individual toxic aerosol species
such as  benzo(a)pyrene show that, in general, these vary proportionately with overall

-------
aerosol carbon levels. The relatively high levels of elemental or soot carbon have important
implications for visibility in the region as well since elemental carbon is an important
contributor to visibility reduction.

Crustal and biomass combustion sources together constitute nearly 80 percent of the
winter season total PM10 measured in the Paso del Norte region - Preliminary estimates
of source category strength  using  tracer elements reveal that  crustal and biomass
combustion sources are  the major  contributors to winter season  PM10.  Average
contribution of the crustal source to PM10 is about 40%. The crustal source is understood
to be closely linked to vehicular sources which are estimated to contribute no more than
20% to total PM10. Vehicle traffic on many unpaved roadways, prevalent in Cd. Juarez
residential areas,  results  in the suspension of both fine and coarse fraction crustal
material, thus  linking  the vehicular and crustal  source categories.  The use of soil-
corrected potassium as a tracer for biomass  combustion reveals about  40% average
contribution of this source to PM10 as well. The use of biomass fuel for the brckmaking
industry and for residential heating in Cd. Juarez is suspected to contribute significantly
to the overall biomass particulate source category.

The pie chart below illustrates the average degree of source contribution to winter-season
PM10 pollution from all El Paso-Cd. Juarez sites and sampling periods as measured during
the Winter 1990 PM10 Scoping Study.
              Crustal (38.2%)—T£
                   Smelter (2.2%)
                                                  Vehicular (18.7%)
Biomass (40.8%)
                        El Paso-Cd. Juarez Winter Air Quality
                       Source Category Contribution to PM-10

-------
                                   1.0 Introduction

The western portion of the  state of Texas  meets the northern part of the Mexican state of
Chihuahua and the southern boundary of the state of New Mexico at a strategic mountain pass,
referred to by Spanish explorers as "El Paso del Norte" through which the Rio Grande courses its
way on its long journey from the high mountains in the state of Colorado to the Gulf of Mexico
many miles distant. This important trade route and center of commerce has been inhabited for
over 400 years and currently accommodates a population of nearly 1.4 million people on both
sides of the international border.  The twin cities of Ciudad Juarez (City of Juarez, hereafter Cd.
Juarez), Chihuahua and El Paso, Texas share a common international airshed that is partially
surrounded by mountains to the north and west of the population center.  Both cities are
experiencing growth however Cd. Juarez, in particular, is experiencing striking growth as the
country of Mexico and the state of Chihuahua makes a transition from an agrarian economy to
an industrial economy centered in large urban zones such as Cd. Juarez.

While the US-Mexico border denotes a physical separation between the two countries, it also
demarcates significant disparities in the standards of living of inhabitants in El Paso and Cd.
Juarez.  Evidence of these economic disparities are particularly evident in the area of air pollution
control. While the country of Mexico and the state of Chihuahua are in principal committed to
the preservation and maintenance of clean air resources,  at a practical level the degree of
pollutant source control on the Mexican side of the border lags considerably behind air pollution
characterization and control programs in place on the US side. The existence of sister cities on
either side of the border poses special problems for those cities faced with compliance with the US
and Mexican clean-air statutes. Border cities such as Tijuana-San Diego, Nuevo Laredo-Laredo,
Matamoros-Brownsvllle and Ciudad Juarez-El Paso, to name a few, share air resources that either
city has the potential to significantly alter. Such issues as trans-border pollutant flux, disparity
of pollution monitoring and controls and enforcement require considerable inter-governmental
cooperation and understanding, particularly in locales, such as El Paso where federal punitive
measures are imposed for non-compliance with air quality regulations.

The El Paso-Cd. Juarez airshed is no exception in regard to such Issues. Local, state and federal
agencies have struggled for many years for a mutually satisfactory approach to the problem. The
Cd. Juarez-El Paso airshed  has historically exceeded US Clean Air Act National Ambient Air
Quality Standards for such pollutants as PM10 (particles less than 10 um diameter), CO (carbon
monoxide) and ozone. In an effort to address this problem, air pollution officials in the State of
Texas and the City of El Paso have lobbied the US  EPA for many years to provide a Federal
mechanism by which air pollution characterization and control strategy development could be
achieved by international cooperation in the area. In recent years, the Mexican Federal Secretariat
of Social Development (SEDESOL) and its predecessor the Secretariat of Urban Development and
Ecology (SEDUE) the agency that is analogous to the US EPA, has taken a more active interest
in U.S-Mexico border air pollution problems. Factors such as these figured prominently in the
successful negotiation and approval of the so-called Annex V that is part of the  1983 US-Mexico
Border Environmental Agreement dealing in particular with air pollution issues along the US-
Mexico border. Finalization of this agreement in 1989, and the appropriation of funds by the US
Congress, also in 1989, laid the groundwork for a series of joint US-Mexico air pollution studies
designed to more fully understand pollution sources and their impacts on both sides of the border.

Specifically, Annex V commits the US and Mexico to cooperate in the further development of air
monitoring programs, pollutant emission inventories  and air modeling approaches for the trans-
border regions, with the ultimate purpose of assessing and implementing various pollution control
strategies for the entire region.

-------
This report summarizes much of the work that has been completed under the Annex V agreement
by various regulatory agencies and research institutions.  Taken as a whole, the work reviewed
in this report represents a significant level  of effort that has been advanced in the  region over
nearly 20 years.  Collectively,  the studies have advanced the technical knowledge from which
further decisions regarding pollutant characterization and pollutant source control can be based.
Many of the study efforts have also enabled productive working relationships to develop between
US and Mexico counterparts in the various regulatory and research agencies.

Six major sections constitute this report. The first section, of which this paragraph is a part,
introduces the topic at hand.  A second section serves to briefly introduce the reader to some
generalities about the  El Paso-Cd. Juarez area such as local topography,  climate and a brief
historical perspective on regional air pollution trends over the past decade. Reviews of a number
of special air pollution investigations carried out in the region by various institutions prior to 1989
are presented in Section 3.  The work summarized in these reviews clearly illustrates  that
investigations into border air pollution issues  were by no means initiated in the late 1980's.
Summary results from these early investigations are presented in this report since they provide
the technical basis upon which many of the later studies followed. A fourth section of the report
summarizes much of the work funded by US Congressional grant monies as a part of the Annex
V agreement and carried out between 1989 and 1992. These recent studies represent investigative
work on both sides of the border and fill some of the information gaps revealed by those studies
conducted prior to 1989. A fifth section of the report goes into considerable detail in describing
a short-term, intensive, winter-season study carried out in 1990-91 to assess PM10 pollution levels
and to  investigate PM10  sources using receptor modeling techniques.  Analytical work and
conclusions  drawn by the project team are discussed with a particular emphasis on their
implications for understanding PM10 sources and pollution levels in the region. A final section of
the report presents an overall summary of work accomplished along a forward look at some of the
air quality issues likely to be of importance  as well as discussing new technologies that may be
available for further characterization and control strategy assessment of air pollution sources in
the region.

-------
                             2.0  Background Information

2.1 Local Topography

The local topography In the El Paso-Cd. Juarez region falls into the category of complex terrain
as a result of moderately sized mountain ranges that lie both to the north and west of the central
El Paso-Cd. Juarez urban area. Local elevations range from 1150 m at the river to 1850 m at the
top of Franklin Peak as shown in Figure 1, a contour map of the region. A 3-D perspective view
of the region with an exaggerated scale in the vertical direction, shown in Figure 2. reveals the
Franklin mountains dividing suburban areas of El Paso into eastern and western halves with the
downtown El Paso area lying just beyond the southern extent of the Franklins. The Rio Grande
flows southward out of central New Mexico through the broad Mesilla Valley on the west side of
the Franklin Mountains. At the bottom of this valley, a large obstacle, the Sierra de Cristo Rey,
lies directly in the path of the river's southward course.  The river bends around Cristo Rey, cuts
through the pass between the Franklin and Sierra Juarez mountain ranges, courses in a generally
southeasterly direction between the downtown districts of El Paso and Cd. Juarez and finally flows
out into an even broader valley to the southeast.  To the west of the Juarez city center lie the
Juarez Mountains, rising to 1650 m above the valley floor.  To the northeast of the downtown El
Paso region lies the extreme southern portion of the Tularosa Basin, bordered on the east by the
Sacramento Mountains and on the west  by the northern reaches of the Franklin Mountains.
Much of the recent suburban growth in El Paso is occurring in this northeastern portion of the
city. To the south of Cd.  Juarez lies the Chihuahuan Plateau characterized by relatively flat
terrain, gradually increasing in altitude above the valley floor as one moves toward the south.

2.2 Local Winter-Season Climate

During winter, atmospheric stagnation conditions marked by stable  air masses and low wind
speeds, are often encountered in the region. These frequently observed calm conditions result in
pollutant buildup over time in the stable air mass.  Pollutant dilution is further minimized during
the evening and early morning hours during stagnation events by the formation of radiation
inversions. As the sun approaches the horizon near the evening hours, the radiant energy flux
from the sun to the surface of the ground decreases, resulting in a cooling of the earth's surface
by the process of radiant heat loss. During the evening hours, the earth, which is warm relative
to the cool winter air aloft, radiates thermal energy to the atmosphere. Air in contact with the
ground is, in turn, cooled by convective heat transfer with the cool  surface.  Air aloft, not in
contact with the ground, experiences no significant change in temperature.  The resulting state
of the atmosphere is characterized by a cold ground surface, a cool air layer near the ground and
an overlying relatively warm air layer. The cool air layer possesses no buoyant rise relative to the
warmer air layer aloft and thus vertical mixing of the atmosphere is suppressed. A radiation
inversion is enhanced by clear skies, since overlying clouds will radiate heat back to the surface
and limit surface cooling. The inversion depth will characteristically build throughout the evening
and early morning hours, finally dissipating when the early morning sun once again begins to heat
the ground.  During a radiation inversion,  the air behaves much like water in the bottom of a flat
valley. Cold air tends to flow into low lying areas and puddle there until forced to move by the
frictional effects of overlying warmer air in motion or when solar heating of the ground occurs.
In the El Paso-Cd. Juarez region,  the valley is  roughly a hundred meters deep below the
surrounding plateau while the mountains protrude into the upper flow some 700 m above the
plateau.  Neither the mountains nor the valley are as simple as described—however the overall
topographical picture is one that features complex mountain ridges and valleys that serve to trap
airmasses thereby limiting the dispersal of pollutants.

-------
   3525000-
   3520000-
 S
 a
 01
 (3
   3515000-
   3510000-
   3505000-
                                   1900
                                   1850
                                   1800
                                   1750
                                   1700
                                   1650
                                   1600
                                   1550
                                   1500
                                   1450
                                   1400
                                   1350
                                   1300
                                   1250
                                   1200
                                   1150
                                   1100
                                                                                    M-MSL
                  355000      360000      365000
                                     EASTING, meters
370000
375000
Figure 1  A contour map of the El Paso del Norte region.  The grid shown is 30 km on a side
and the axis units are in meters.
                                           8

-------
Figure 2  A perspective plot of the El Paso del Norte region as viewed from the southeast.
The overall square grid domain is 30 km.

-------
2.3 Regional Air Quality Monitoring Programs

Air quality monitoring programs have been in place within the El Paso del Norte region for many
years. As a result of the original Federal Clean Air Act of 1970, funding mechanisms through the
US EPA have made possible continuing development of air quality networks on the El Paso side
of the border.  In the early days of the program, monitoring sites usually consisted of total
particulate samplers and gas sampling systems employing wet chemical techniques at a few sites.
Over the past several decades the  development of reliable continuous air quality monitoring
instrumentation has been a key factor in the  further development of air quality monitoring
networks in El Paso.  Currently,  numerous US monitoring sites are scattered throughout the
region.  The coverage ranges from the Moon City clinic at the  extreme southeast portion of the
valley, about 25 km distant from downtown El Paso, to Las Cruces, New Mexico, 70 km to the
northwest of El Paso.  Monitoring capabilities require considerable cost and manpower for
instrument acquisition and maintenance however, so  not all sites are similarly equipped with
monitoring instruments.  At the present time the US monitoring  sites within the region are
maintained and operated by the Texas Natural Resource and Conservation Commission (TNRCC),
the El Paso City-County Health Department, and the New Mexico Environment Department.

Monitoring capabilities in Cd. Juarez are less developed than those on the US side of the border.
In general, the stage of development at Cd. Juarez monitoring sites is at a point similar to the
condition of US sites 10 or 15 years ago. At many of the sites, the monitoring capabilities consist
of a single total particulate or PM[0 hivolume sampler. While many of the US sites are networked
to a common data processing location, this is not true for Cd. Juarez sites. The operating budgets
of SEDESOL, are much smaller when compared to US  agency budgets and they typically do not
include  provisions for the purchase of relatively expensive air quality monitoring equipment.
While one can convincingly argue that pollution monitoring agencies  in the US are under-staffed,
Mexican agencies find themselves in a comparatively worse  situation  significant  resource
limitations. As a part of the Annex V agreement however, some US funding was made available
to the Mexican agencies for the modernization of the air monitoring network in Cd. Juarez.  Two
Cd. Juarez sites in particular were extensively developed and networked to a  common data
repository along with several US sites as a part of the intensive winter season particulate study
carried out in Winter 1990-91 discussed in more detail In Section 5 of this report.

2.4 El Paso del Norte Airshed Pollution Historical Trends

An in-depth review of pollutant levels in the El Paso-Cd.  Juarez area over the past several decades
is beyond the scope of this report however, graphical summaries of the four highest PM10 levels
and the two highest CO levels at selected sites in both  Cd. Juarez and El Paso over the past
decade are given in Figures 3 through 6. The PM10 data  represented in these graphs are compiled
from "quick look" summaries taken from the EPA Aerometric Information Retrieval System (AIRS)
data base. Numerous exceedences above the US air quality 24-hour average PM10 standard of 150
ug/m3 are observed at some of the  more highly polluted monitoring sites  on both sides of the
border.  Of interest, however, is the fact that PM,0 pollution, as represented by the four highest
readings in a year for the US sites, shows a decreasing trend, particularly in the years 1990
through 1993.  Highest and second highest 8-hour averages for CO are shown that exceed the
air quality standard of 9 ppmv.  The  CO trend for the US site, shown in Figure 4, is less clear and
indicates either stable or increasing levels.  Pollution trends  for PM10 and CO data from Cd.
Juarez,  as shown in Figures 5 and 6, cannot be discerned since only four years of data are
available in the AIRS database.
                                          10

-------
          CO
          <

          I

           o"
          I
           u.

           o


          4
          CM
          O
3SU-




150-

100-

50-









....
1 E







:
j
i
!
;
:
'
I
1

1
1
I
1
1
1
1
I
1
1
1
I

1
I
1
I

1



_ 1
I
I
I
1
I


	 . (
I
1
1
1
~ \

\


\
1
I
1
1
!



:
R!

; !



*. D
: !
- \
- I
\ \
• \
: (

: \


I
I
1
1
i
1
I
_ 1
1

: !
;
i
ai






1—8
1 .
1 .
•
.


"


!






t
*
i
*
«
!
«

V
*
>




i
1
1
I
1
!
1
i

!





i_
i
i
i
i
i
i

i
i
i
i
: »







:

-

-
                     1985  1986  1987  1988  1989  1990  1991   1992  1993


                                      Year
Figure 3 The four highest 24-hour PM-10 levels measured annually at the El Paso Tillman

site from 1985 through 1993.
             14-
           11*

           Q.
           0.

           Q)" 10-


           5
           <5
           _


           O



           CO


           O

           O
111
1
   i!
                                              I
n
X
x
N.
N.
X
X
X
X
x
x


1
II
                    1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993


                                      Year
Figure 4 The two highest 8-hour CO averages measured annually at the El Paso Tillman site

from 1985 through 1993.
                                    11

-------
350-
300-
E
"Bi 250-
3
of
O)
S. 200-
>


o 15°"
1
o 100'
5
Q.
50-



H
i
i
i
i
i
i
i
i
i
i
i
i


|
i
I
NO DATA AVAILABLE j
I
!

1


I
" 1
I
|

, _ '
I j
1
1
1
1
1
I
!


1 * i
1 " i
i
i
i
i
i
i
1

i
i

-
:


:
•


• '

i
3 i
> _ i
i
: !



i
: |
; !

: i
!


; !





-

;
;
\

•



'-
                        1985  1986  1987  1988   1989   1990   1991   1992  1993
                                            Year
Figure 5  The four highest 24-hour PM-10 levels measured annually at the  Cd. Juarez
Advance Transformer site from 1990 through 1993.
               16-
               14-
a.
o.

-------
A summary of average 24-hour PM10 levels measured during the 1992 fourth quarter prepared by
the EPA Region 6 Geographical Information Center is shown in Figure 7. Average PMi0 data from
all PM10 sites reporting into the AIRS data base are shown along with interpolated values over the
entire region as contour lines. The highest level is near 150 ug/m3 and is observed in central Cd.
Juarez. In general, the PM10 contour levels show an increasing trend as one moves south from
the US into Mexico.

These summary historical data illustrate that pollution levels in excess of applicable air quality
standards occur with some regularity in the area. Although graphical summaries for ozone are
not shown here,  air quality standard  exceedences  are  observed  for this  pollutant as well.
Exceedences of the US NAAQS PM10 standard occur  on both sides of the border raising many
issues concerning the degree of responsibility borne  by each country and the extent to which
control measures need be applied in the respective US and Mexican urban areas.
                                           13

-------
      DONA ANA; MM
      10 ug/m3 Contour Interval
      Base Contour 20 ug/m3
                      Roads
                      Contour Lines
                      PM-10 Monitoring Site
.  .
                  40.000 Contour Level, ug/m3
                  22.5  PM-10 Mean Value. ugftn3
Figure 7  Fourth-quarter 1992, 24-hour average PM-10 measurements in Cd. Juarez and El
Paso showing interpolated values over the region as contour lines.
                                           14

-------
                  3.0 Pre-1989 El Paso-Cd. Juarez Air Quality Studies

3.1 Carbon Monoxide Studies - University of Texas at £1 Paso, 1983

A report published by the Center for Inter-American and Border Studies at the University of Texas
at El Paso reviews a number of studies that were carried out in the early 1980's in the Cd. Juarez
- El Paso region [Bath, 1983]. A cursory review of each of several chapters included in the report
follows.

       A brief review of topography, climatology and traffic are presented by H. G. Applegate. In
       this short summary, the author notes that correlation studies comparing simultaneous
       wind speed and direction measurements at the El Paso International Airport, local TACB
       monitoring sites and Cd. Juarez Aeropuerto Federal failed to show any correlation among
       the various sites.

       M. Aguirre and others summarize various local and Federal air quality standards enacted
       at the time when this report was published along with a description of various measured
       CO exceedences and the meteorology associated with them. The authors note that El Paso
       was determined to be out of compliance for both CO and O3 federal air quality standards.

       Price and Applegate present a discussion of international bridge crossings along with
       estimates of CO release from vehicles crossing the  bridges based on estimates of CO
       emissions from US cars. The authors did not attempt to estimate the CO emission levels
       from many older, poorly maintained Mexican vehicles but rather applied US vehicle
       emission factors thus very likely underestimating the amount of CO from the Mexican
       vehicle fleet.

       Olmstead reports on predictions of CO levels near the University of Texas at El Paso using
       the CALINE 3 Dispersion Model. The results of this particular study reveal that violations
       of the NAAQS for CO (35 ppm one-hour average) were not to be expected at any of five
       intersections surrounding UTEP  for the period 1982 -  1990.  The highest  predicted
       concentration (16.8 ppm at the Mesa-University intersection) was less than 50% of the
       one-hour standard.

       Applegate reports on the history of CO measurements in El Paso prior to 1980.  An initial
       study was carried out by Tillman in 1959.  This effort was followed by studies in 1968 at
       San Jacinto Plaza and in  1973 at the Americas and International Bridges.

       Reynoso and Gonzales report on a region-wide CO  grab sampling study conducted in
       November of 1981. Grab samples of urban air were collected by volunteers at numerous
       sites throughout the region at precisely the same time.  Samples were subsequently
       analyzed for CO content by a portable sampler.  CO "hot spots" were observed along a
       corridor extending from east of downtown Cd. Juarez in a northerly direction to a point
       east of downtown El Paso.  The International  Bridge was also located  in this corridor.
       Mean CO values throughout both cities were 13, 15 and 11 ppm at 1800, 2000 and 2200
       hours respectively.

       McDonald and Rab report on CO in the vicinity of Fort Bliss.  The authors note the
       formation of a rideshare program to reduce CO emissions in this particular area of El
       Paso.

       Quiz reports on CO concentrations in Cd. Juarez (written in Spanish).

                                           15

-------
3.2  Inhalable Particulate Matter Assessment - Environmental Research and Technology
Inc., 1983

The Texas Air Control Board commissioned Environmental Research and Technology Inc. to
conduct a study to assist the agency in an assessment of the significance of primary particulate
matter emissions in the PMi0 and PM1S size ranges In the Houston-Galveston and El Paso areas
[ERT, 1983].  The study was carried out prior to the enactment of a PM10 or PM15 standard for
airborne particulate matter and was intended to Investigate how the new standard might Impact
these two urban areas. The primary tasks carried out during this study are briefly summarized
below.

       Identify the major stationary sources of primary particulate matter In the Houston-
       Galveston and El Paso areas and determine the extent to which these sources contribute
       to PM10 and PM15 emissions.

       Categorize and characterize the emissions of certain sources Identified In the first task
       with respect to such qualities as size distribution, chemical and elemental composition,
       control devices in use etc.

       Review selected source categories to determine physiologically significant properties of
       PMj0 and PM15 that might serve as appropriate bases for alternate regulatory standards

       Identify and evaluate the effectiveness of control technologies currently available for PM10
       and PM1S emissions from the major source categories.

       Determine the cost effectiveness of the various available particulate matter control systems
       along with estimates of emissions reductions likely to result from the application of these
       technologies.

       Review and evaluate the effects of PM,0 and PM15 emissions on ambient air quality with
       an Intent to identify gaps In the current level of understanding and to suggest additional
       studies to fill these Information gaps.

In the interest of brevity only a few salient points, most relevant to the current Issues faced in the
El Paso-Cd. Juarez airshed, are summarized from this broad effort.

       The study examined emissions from only those stationary sources compiled into theTACB
       1980 emissions inventory. Based on this Inventory, stationary sources accounted for 15%
       of the total emitted particulate matter in the El Paso area and 55% of the total In the
       Houston-Galveston area.

       Thirty-nine source categories are listed for the El Paso region as compiled from the TACB
       1980 Emission Inventory.  Most of these are  related to non-ferrous (copper, lead, zinc)
       metal production, cement production and petroleum refining operations. The two highest
       non-fugitive point source categories were copper reverberatory furnaces at 107 tons of
       PMi0 per year and fluid catalytic crackers at 108 tons per year.

       The study produced an estimate of the degree to which overall PM15 emissions would be
       reduced if all non-fugitive point sources were eliminated. For the case of El Paso, a total
       reduction of  6% was  determined.  In contrast, a PM15 reduction of 25 to 37% was
       estimated for the Houston-Galveston area. The authors note however, that no attempt
       was made to take sources In either Mexico or New Mexico Into account in these estimates.

                                          16

-------
       The authors speculate that based on limited PM15 measurements completed at the time
       of the study, exceedences of a likely 24-hour average PM10 standard in the range of 55-120
       ug/m3 would not be expected to occur.

       The authors also point out the need for the chemical and physical speciation of ambient
       air samples to provide appropriate data for the application and use of receptor models
       which they suggest would help to quantify the contributions of various source types to the
       overall PM10 in the region.

3.3 Air Quality Database Study — Radian Corporation,  1983

A study was commissioned by the Texas Air Control Board in the early 1980's to describe El Paso
air quality and to identify significant pollution sources.  A database for selected pollutants and
meteorological parameters was compiled for use in the study from sampling networks, on the US
side of the border only, operated by TACB, EPCCHD, NMEID, NWS and ASARCO over the years
1977 through 1981.  Specific pollutants investigated were TSP, CO, SO2 and Pb. The database
drew from 22 TSP monitoring sites, 17 Pb sites, 10 SO2 sites and four CO sites and included
quality assurance criteria for all pollutant  and meteorological parameters accepted into the
database. The study is segmented into two volumes, the first [Radian, 1983A] presents analyses
results and conclusions and the second [Radian, 1983B] presents the data validation criteria used
for compilation of the database.

The authors of the study conclude that exceedences of national ambient air quality standards for
TSP, CO and Pb  occurred during the study interval. Sulfur dioxide was not observed to exceed
the standard however values close to the standard were observed.

Some of the more important conclusions from this study are summarized as follows:

       One-hour CO exceedences were not observed at any sites however the 8-hour CO standard
       was exceeded at three of the four sites used in the study.  Eight-hour CO averages
       exceeded the 9 ppm standard less than 1 percent of the time for three of the four sites.
       Peak 8-hour levels at these same three sites averaged 10.9 ppm.

       High TSP levels in El Paso are  associated with  dry conditions which favor the re-
       suspension of urban or rural particulate material. Monitoring data show that high winds
       can increase TSP levels as a result of entrainnient of material from exposed soil.  High TSP
       levels were recorded on so-called "dust storm" days as well as on days characterized with
       moderate or light wind speeds.

       The most important sources of TSP in El Paso are re-suspension of local particulate matter
       by urban traffic and by winds gusting over 15-20 mph. These situations produce at least
       50  to 80 percent of the measured  TSP at most monitoring sites as indicated  by  a
       companion microscopy study discussed in a following section of this report.

       Sources of lead emission that contribute to high ambient lead levels include transportation
       sources using leaded gasoline, ASARCO smelter emissions, and re-suspension of lead-
       containing particles by wind or vehicles. Lead levels are observed to be highest during
       periods of air stagnation in the region.
                                          17

-------
As a part of this study, meteorological conditions were reviewed for days on which high and low
TSP measurements were observed.  High TSP days were classified as days on which at least one
site recorded its highest reading for the period 1979-1981. An analogous approach was taken to
identify a collection of low TSP days. The authors observe that peak sustained winds in excess
of 20 mph were reported on six of 16 "high" TSP days encountered in the period  1979-1981.
Winds of 11-20 mph were recorded on another six of the days with the remaining four days having
peak wind speeds less than 10 mph.  The daily resultant wind  directions did not show any
consistent patterns for the high TSP days leading the authors to conjecture that non-local sources
of TSP accounted for the measured levels.

The authors go on to point out that most of the low TSP days were characterized by rainfall
occurring within the previous two days of the measured level. Nine of the 12 low TSP days had
rain on the same  day or during the preceding two days. Ten of the 12 days  had moderate wind
speeds with peak sustained winds from 11-20 mph.

A correlation study of so-called "persistent wind" days with TSP concentration was also conducted
in this study.  Persistent wind days were classified as those days in which the  ratio  of resultant
wind speed (the vector average) to mean wind speed was 0.7 or greater.  The authors conclude
that for the sites investigated on days with the wind blowing from a persistent  direction, there was
no  significant dependencies of measured TSP on wind direction.  These results are taken to
suggest TSP originated from either multiple local sources or a regional source.

Finally, the authors note that the statistical analyses carried out on the database do not resolve
the question of the relative contributions to high pollution levels made by sources in El Paso and
Cd. Juarez. The absence of any clear correlations of TSP with wind direction is taken to suggest
that Cd. Juarez is not necessarily the predominate source of TSP.

3.4  £1 Paso Quantitative Microscopy Study ~ Energy Technology Consultants, 1983

Another study was commissioned by the Texas Air Control Board in parallel with the Radian Air
Quality Database Study.  This study, known as the El Paso Quantitative Microscopy Study, was
conducted by Energy Technology Consultants with a goal of identifying and quantifying the
sources of TSP and particulate lead in the El Paso airshed using a combination of microscopy
analysis and a receptor modeling approach. The report is organized into two volumes with the
first [Energy Technology Consultants, 1983A], giving analysis results and conclusions, and the
second [Energy Technology Consultants, 1983B] presenting the analytical methods in detail along
with quality assurance procedures and results of statistical analyses.

For this study, a number of ambient TSP samples collected during 1981  were selected  from
archives for detailed analysis using a computer controlled scanning electron microscopy-energy
dispersive x-ray method. The study incorporated receptor modeling techniques to the extent that
source contributions were inferred from the elemental content of particulate material on the filters
subjected to analysis. A total of 70 ambient TSP samples on cellulose or glass filters were selected
from six monitoring sites to represent air quality in El Paso.  Filters were selected to correspond
to low TSP (< 59 ug/m3) mid TSP (59 - 125 yg/m3) and high TSP (>125 ug/m3) occurrences.  In
the analysis an emphasis was placed on the high TSP samples in an effort to identify sources
contributing to exceedences of the air quality standards.


A number of bulk soil, industrial particle, and vehicular emission samples were also collected in
the El Paso area and were analyzed with the same apparatus as used for the  filter samples. The
samples were used to quantitatively describe known major sources of particulate material in the

                                          18

-------
El Paso airshed and were the references against which the filter sample results were compared
using receptor modeling techniques. Analyses of numerous soil samples resulted in considerable
overlap of the fingerprints of each sample.  To simplify the source profile, the following general
source categories were defined:

       Soil - a blend of the various soils encountered In El Paso

       Industrial Fugitives - emissions from brick manufacturing, rock quarrying, slag crushing,
       and cement operations

       Highway - exhaust emissions from lead fueled vehicles along with entrained soil, road
       wear and tire wear particles

       Smelter - various fugitive and stack emissions from the ASARCO primary lead and copper
       smelter

       Unknown - material for which no fingerprint could be found (this source classification
       likely Includes emissions from various Mexican sources not measured in the study)


Table 1 summarizes the TSP source apportionment results from this study using the computer
controlled scanning electron microscopy method for seven sites throughout the El Paso city area.
These summary results reveal that urban soil contributes in the range of 50 - 85% of the total TSP
on the filters at all sites surveyed in the study. The next highest category is Industrial Fugitives
followed by the Highway category.

The authors compared receptor modeling  results obtained by microscopic analysis to receptor
modeling results carried out by Radian Corporation and the Texas Air Control Board using
elemental tracer techniques and found good agreement between all three approaches for a site
near the ASARCO smelter.
                                          19

-------
I
at
C
4> ^
18
0.3
1*
i
|£
s_
(0
s-
1 ?
X
•1 §
IF
c _
i CO &*
« $
•** «- "5.
z 1
Q.
| g
to
0 O
0 OC
i <
5 co
.2 <
o o
•8 o
e >

5
C
O
I
CO

o
co


^.


en
CM
U)
o


co
00
•*






LU

m
o
i

0>
CD


CO
T-


*~
•
O)


2
00
"*






LU
co



UTillman,
EPCCHD

O)
at


00


*~
8



S
s







UJ



Q.
UJ

CO
CO


co


CO
ID
co



§
CVJ
•^






LU
CO
UJ



I Ascarate P

^
T-


CM


&
*
CM


CO
co
o






1
1



00
m

co
CO


CO


o>
R
co


co
en

tO






|


— s.
5 "5
v* O
O -C
CO O
N 05

h.
^


T-



N.
IO
CO


R
en
°






LU



Ivanhoe
                            c
                            o
                            •£3

                            s
                            o
                            O
                            c
                            s>

                            8
                            LU
                                                       o
                                                       CN

-------
4.0 Regional Studies Conducted after 1988

4.1 Airborne Lidar Study - EPA Environmental Monitoring Systems Laboratory, 1989

Region 6 EPA commissioned the EPA Environmental Monitoring Systems Laboratory to conduct
an airborne lidar study of the El Paso-Cd. Juarez area in February 1989 [McElroy, 1990]. The
objectives of the study were three-fold and are summarized as follows:

       Develop an aerometric database in order to ascertain the spatial extent and magnitude of
       the suspended particulate problem in the El Paso-Cd. Juarez region.

       Plan a  comprehensive monitoring program to  aid  in the development of pollution
       attainment strategies  in the El Paso-Cd.  Juarez region.  This study was carried out
       preliminary to an intensive ground-based  PM10 monitoring study that is described in a
       later section of this report.  Information from this study was  used in the selection of
       numerous PMJO monitoring sites.

       Determine, if possible, the magnitude and direction of particulate pollution flux across the
       international boundary between the cities of El Paso and Cd. Juarez.

The principal feature of this study was the deployment of a downward-looking airborne lidar
system used to map aerosol in the region on both sides of the border by measuring the aerosol
backscattered intensity of laser pulses from the lidar. The lidar system measured backscatter
intensity from a light pulse as a function of time after exiting the lidar.  The elapsed time during
which the laser pulse travels out and is scattered back from a distant aerosol mass to the lidar
is a measure of the aerosol mass distance from the lidar.  Lidar pulse rates for this particular
system were between 1 and 10 per second. Since the aircraft is also moving forward as the laser
is continuously pulsed, a two dimensional picture of aerosol burden in the area can be produced
during a single aircraft pass over an area. By flying multiple parallel passes over an area during
a several-hour period, a three dimensional picture of aerosol concentration in the  area can be
produced.

The  airborne lidar phase  of the study was supplemented by a number of ground-based
measurement systems including nephelometers  and stacked filter units (measuring 2-hour
intervals of fine and coarse aerosol concentrations) at three sites along the international border.
These measurement systems were included to yield ground-truth data with which to compare the
lidar aerosol density measurements. Winds aloft were also measured during aircraft flyovers with
two radiotheodilite stations in the region as well.  Continuous PMi0 data available from several
beta gauge monitors positioned on the rooftop of the Tillman Health Center in downtown El Paso
were also used for comparison with aerosol density measurements made with the airborne lidar.

The authors  prepared a series of contour plots of lidar backscatter intensity averaged in the
altitude interval ranging from about 6 to 40 m above the ground over the entire region by use of
interpolation routines that estimated lidar backscatter in the regions between those for which
overflights were completed. Figure 8 is a  typical example  showing data for the morning of
February 23. The actual flight tracks over the region are noted by plotted circles. The contour
lines resulting from the data interpolation routine are shown as solid lines. Contour units are
given as aerosol backscatter intensity-a parameter which is roughly correlated to aerosol density.
The highest aerosol levels are observed along both sides of the international border and in
                                          21

-------
particular near the downtown areas of both cities. High levels are also noted near the smelter
located close to the junction of the Texas, New Mexico and Mexico border. The highest aerosol
levels in the study were observed in the vicinity of a burning refuse dump located in the foothills
of the Juarez Mountains southwest of the El Paso city center.
       w

       o>
       E
       D)
       C
       o
       z
            "310000
                        351000
                                     366000         382000

                                      Easting, meters
                                                              396000
410000
  Figure 8 A graph showing airborne flight tracks (circles) on the morning of February 23 with
  near ground-level aerosol backscatter intensity, interpolated from the lidar data, shown as
  contour lines [from McElroy,  1990].
The  authors further discuss  the  results  of  several attempts  to relate  lidar backscatter
measurements  to measured ground-level aerosol concentrations.   First, the lidar backscatter
intensity measured in the vicinity  of the three ground-level sites, where aerosol monitoring
equipment was positioned, were compared in an effort to convert lidar backscatter intensity values
to aerosol density.  The authors report that although some parameters showed a reasonable

                                           22

-------
relationship to  lidar data,  in  general,  the  relationships between ground-based aerosol
concentration measurements and lidar backscatter measurements at the same locations were
inconsistent from site to site. The authors go on to discuss a number of factors which may mask
the observation of clear relationships between lidar backscatter intensity and ground-level aerosol
concentration.  First, the lidar data collection flight tracks were not flown directly over the
sampling sites where aerosol concentration measurements were being made,  thus, different
volumes of air may have been compared by the two methods. Second, ground reflection of the
lidar pulse interferes with the measurement of aerosol closest to the ground which is what the
ground-level samplers are measuring. Finally, the filter sampling interval at the ground locations
was 2 hours. This relatively short sampling period resulted in light mass loadings of material on
the filters, causing relatively high uncertainties in the mass concentration measurements derived
from the filters.  Furthermore, filter analysis was not carried out until several months had elapsed
such that artifact formation or volatile losses may have occurred on the filters in the intervening
storage interval. Thus, data uncertainties related to the measurement location, interval and the
storage time prior to analysis may have contributed to the difficulty in seeing a clearer relationship
between lidar backscatter and ground-based aerosol measurement.

In a review of data from CO and PM10 ground monitors, the authors note considerable temporal
correlation between the two parameters. Based on these observations, the authors suggest that
CO may be a reliable predictor of PM10. The authors go on to suggest that three lidar flights
occurring at about 8 AM, 2 PM and 8 PM would yield enough data to give a reasonable prediction
of 24-hour CO and PM10 levels encountered on any particular day.

As a part of this study, elemental analysis was carried out on a series of stacked filter units that
were sampled over a 2-hour interval on various days during the study.  The average mass
breakdown of the fine aerosol fraction (< 2.5 micrometer diameter) for the Chamizal site was as
follows: organic carbon, 40.3%; elemental carbon, 25.1%; sulfate species, 18.2%; crustal or soil
species, 12.8%; and other, 3.5%. The authors conclude that particulate sources in the El Paso-
Cd. Juarez area produce an extraordinary amount of organic particulate matter for transport and
dispersion in the region.  The authors do not report a similar species breakdown of the coarse
fraction filters.
Finally, to address the topic of aerosol flux across the border, the authors explored the use of a
diagnostic wind field model to construct a wind field of the region during selected intervals when
lidar data from aircraft overflights was also available. The two data sets (wind speed and direction
and lidar backscatter intensity, both as a function of position) for each lidar flight track were then
combined to arrive at an average aerosol flux value for a particular flight track across the region.
Results from these efforts were inconclusive. The authors note the complexity of this exploratory
effort and  observe that  additional work would be necessary to  more fully understand the
relationship between aerosol density, the wind field in this complex terrain and ultimately, the flux
of aerosol across the border.
4.2  Upper Air Winds and Visibility Study - University of Texas at El Paso, 1989

Researchers at UTEP were engaged by EPA Region 6 to support the airborne lidar and saturation
PMi0 studies carried out in El Paso-Cd. Juarez in late 1989 [McDonald, 1990 and Ballard, 1990].
Measurements were carried out concurrently with the EPA-EMSL airborne lidar measurements
and consisted of a tethered balloon system carrying a suite of meteorological sensors, a ground-
based meteorological monitoring system and a camera for visibility degradation assessment.

                                           23

-------
Measurements were taken with these systems at the Chamizal site, located southeast of the El
Paso downtown area immediately adjacent to the border, and at the UTEP campus located to the
northwest of downtown El Paso at a higher elevation. Atmospheric sounding and ground based
nephelometer measurements are reported for December 9 and 10, 1989.

Not surprisingly, the tethersonde data revealed temperature inversions in the early morning hours
extending up to at least 150 meters.  Data above 150 meters were unavailable since  FAA
regulations prohibit balloon flights above this altitude. The inversions were observed to dissipate
as the day progressed.

Of particular interest in this study was the deployment of a balloon-borne nephelometer at the
Chamizal site which allowed an indirect measurement of aerosol burden in the vertical dimension.
These data were collected in order to yield "ground-truth" data for comparison with airborne lidar
data collected at the same time. The report contains a selection of plots which show the measured
scattering coefficient and the balloon altitude as a function of time on selected sampling days.
In general, aerosol light scattering values are highest near the ground and drop off with increasing
altitude.  Not enough data are presented in the report, however,  to assess the extent to which
aerosol layers aloft were present in the 150 meter interval that the tethersonde system was flown.
The authors attribute some of the high aerosol scattering measured at the Chamizal site to local
activity caused by study participants. In general, aerosol scattering values 30-150 m above the
ground were a factor of two or less than those measured in the region zero to 20 m above the
ground.

In a companion report, researchers from UTEP describe a series of measurements made with a
video camera which was equipped with a video frame grabber  and software that enabled digital
processing of various images taken from the UTEP campus looking toward the south over the city
of Juarez. The objective of this effort was to calculate total light extinction and corresponding
visual range in a semi-continuous manner over selected days in the study period. The authors
describe in considerable detail the theoretical basis for how extinction values are determined from
the video images.   Path  averaged transmittance, path averaged extinction and visibility are
summarized In  a series of tables for a selection of video images taken on December 10, 1989.
With this technique, the authors note that the measurements are useful when compared against
each other, however, estimates of visibility parameters such as extinction may not necessarily
correspond to actual values present in the atmosphere when the measurements were taken.  The
authors note that in future studies,  the use of targets with well-defined optical characteristics
would enable more accurate measurements of visibility parameters. The authors make no attempt
to further interpret these data with regard to time dependance or meteorological effects on the
measured haze  parameters.

4.3  Saturation PM10 Study - EPA Region 6, 1989

An intensive short-term study of PM10 levels in the El Paso region was carried out by EPA Region
6 personnel in December  1989 [Kemp, 1990J. Objectives of this study were as follows:

       Integrate reference PM10 methods with other non-reference, short-term PM10 monitoring
       methods to characterize PM]0 levels along the US-Mexican border and answer the following
       questions: (1) What are PM10 concentrations in regions of the airshed that are not routinely
       monitored?  (2) Is the current reference PMi0 monitoring network in El Paso adequately
       characterizing the ambient air?

       Develop guidelines and study protocols for the operation of portable or screening type PM10
       monitoring in future short-term intensive studies of this nature.

                                          24

-------
       Establish working relationships and foundations for future joint air  quality studies
       between US and Mexican local and federal agencies.

The miniaturized PM10 monitor consists of a small battery-operated pump and filter that is
equipped with a size selective inlet such that only particles less than 10 micrometer diameter are
passed through the inlet and onto the filter. The filter is weighed prior to and following sampling
to measure the particulate mass collected on the filter during the passage of a known volume of
air through the filter.  Twenty-eight of these samplers were hung from utility poles in a gridded
pattern along the US-Mexico border with 24 samplers located in El Paso and four located in
Juarez.  Sampling intervals were either of 3.5 hours or 22 hours duration.  Hourly surface
meteorological data were also archived from two TACB monitoring sites (Ascarate Park and UTEP)
such that a wind back-trajectory calculation could be carried out for some of the aerosol data
collection periods.

Measured PMi0 concentrations from 8:30 PM until midnight on December 9, 1989 ranged from
a low of 24 ug/m3 in the vicinity of Sunland Park, NM near the Texas-New Mexico border to a high
of 745 ug/m3 in the western area of Juarez. PM10 levels in the eastern portion of the basin were
at about 90 ug/m3. The authors note that the high levels in western Juarez were a result of local
sources in the area.  During some sampling periods, problems were encountered with the nickel-
cadmium batteries used  to power the saturation PM10 monitors.  Internal battery resistance
increased to  high levels  at temperatures near freezing such that  sampler performance was
impaired. Despite these problems, data capture on selected days was good and provided a data
set suitable for subsequent analysis.

The 22-hour samples afforded enough sampling points on  selected days to  carry  out data
interpolation and contour mapping of estimated concentrations  across the entire airshed.  A
contour plot  showing estimated PM10 levels is shown in Figure 9, compiled from all 22-hour
measurements taken on December 11, 1989.  The highest levels are noted immediately south of
the junction of the New Mexico, Texas, Mexico border. In general, the contour results for most
days sampled show that PM10 levels are highest on the Juarez side of the border, however, as
noted above,  only four samplers were positioned on the Mexican side of the border.

Wind direction back-trajectories were calculated from two TACB meteorological ground stations
using 10-m tower data from several days to elucidate possible sources of PM10 in the region.  On
December 12, 1989, winds were generally flowing from the west and down valley near the UTEP
location where relatively high PM10 concentrations were observed. Sources along the western edge
of El Paso and Juarez are implicated in this analysis.  On the other hand, back trajectories for
Ascarate Park where only moderate PM10 levels were observed showed motion largely confined to
the eastern portion of the airshed. From these data, the authors concluded that the major PMi0
sources exist in the  western  portion of the basin.   The authors also  note that  significant
differences in PM,0 levels occur at different altitudes in the same general region of the city. PM10
levels in the range of 200 to 300 ug/m3 were observed in the valley bottom just southeast of the
pass while levels of about 60 pg/m3 were observed at the UTEP site, 1.6 km distant and 44 m
higher than the valley bottom sites.

The authors draw the following overall conclusions from this saturation PM,0 study:

        High PM10 levels were encountered regularly along both sides of the US-Mexico border in
        regions not normally covered by the local PM10 monitoring network.

       The existing PM10 monitoring network is judged to be of marginal use for characterization
        of PM10 over the  entire airshed.  Very high levels near the pass at the La Hacienda

                                          25

-------
    Restaurant are not adequately represented by any of the existing monitors in the PM10
    network.

    The Saturation PM10 Study was useful in establishing a working relationship with Mexican
    counterparts and will lead the way for additional cooperation between the US and Mexican
    agencies in the pursuit of regional air quality improvement.
         !—|  so  to  100

            100  to  150

            150  to  200

            200  to  250

            250  to  300
Figure 9  A contour plot showing PM-10 values interpolated from measurements on
December 11, 1989 during the Saturation PM-10 Study (from Kemp, 1990].

                                       26

-------
4.4 Cd. Juarez Gasoline Vapor Pressure Study — SEDESOL and others, 199O

In September 1990, personnel from SEDESOL, the City of Juarez. EPA and the City of El Paso
collected about 40 gasoline samples from a number of gas stations in Cd. Juarez for fuel volatility
measurements. Figure 10 is a histogram showing the distribution of vapor pressure readings from
all samples.
                 8-1
                       8.0 8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9 9.0 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.910.0
                                     Reid Vapor Pressure, psi
   Figure 10 Frequency histogram showing RVP measurements for 41 samples collected in Cd.
   Juarez [from Yarbrough, 1994].

The average reid vapor pressure (RVP) value for the Cd. Juarez samples was 9.02 psi with a
standard deviation of 0.29. This average was higher than an average RVP value of about 8.0 psi
currently encountered in El Paso. During the mid- to late- 1980's, RVP values in some Texas cities
remained quite high (over 10 psi). In light of this information, it was suspected that values in Cd.
Juarez would be similarly high. These survey results reveal a lower RVP value than anticipated.
Consequently, VOC emissions from Cd. Juarez vehicles are expected to be lower than estimates
made prior to this study.

4.5 Juarez Vehicle Tampering Survey *- Colorado State University, 1990

In conjunction with other elements of the region-wide study in  1989  and 1990, EPA Region 6
commissioned a study group from Colorado State University to conduct a tampering survey of the
vehicular fleet in Cd. Juarez [CSU, 1990). About 800 vehicles were surveyed between September
                                          27

-------
24 and October 4, 1990 at nine locations in Cd. Juarez.  Two of these sites were located at the
international bridges.  Data collected during the  survey Included tailpipe  measurements of
hydrocarbon and CO emissions, results from visual inspection of various  emissions  control
components and odometer readings of the surveyed vehicles.  A summary  of overall results.
showing the degree of vehicle tampering is given in Table 2.  Of the 514 light duty vehicles
surveyed, 318 or 62% showed signs of tampering. An additional 75 vehicles or 15% of the total
were placed in an "arguable" category where evidence  of tampering was observed although
inconclusive. Light and heavy duty trucks show similar high rates of tampering with rates in
excess of 67% for light duty trucks and 74% for heavy duty trucks observed in the survey.  In the
survey report, results are further broken  down into  such tampering categories as catalytic
converter, filler neck restrictor, air pump system, evaporative canister, exhaust gas reclrculatlon
system, heated air intake system and oxygen sensor. The highest three categories of tampering
were  observed  in  the  air  pump  system  (46%),  air  pump  belt  (44%)  and catalytic
converter/evaporative canister (both at 36%) categories. The authors make a comparison between
tampering rates factored Into the EPAMobile4 vehicular source emission model and observed Cd.
Juarez tampering rates and conclude that MOBILE4 significantly under predicts the degree of
tampering expected In a fleet composition such as encountered in Cd. Juarez. For example, at
the actual mileage of the Juarez vehicular fleet, MOBILE4 predicts a tampering rate of 37.9% as
compared to an observed rate of 67.2%. In light of these uncertainties, the authors point out that
MOBILE4 in turn would underestimate the true Juarez vehicular fleet exhaust emissions.
                                       Table 2
                                Juarez Vehicular Fleet
                         Tampering Survey Summary Statistics
Vehicle Type
Light Duty Vehicle
Total
Light Duty Truck
Total
Heavy Duty Vehicle
Total
Condition
Pass
Malfunction
Arguable
Tampered

Pass
Malfunction
Arguable
Tampered

Pass
Arguable
Tampered

Count
107
14
75
318
514
52
11
22
175
260
3
11
40
54
Rate
21%
3%
15%
62%

20%
4%
8%
67%

6%
20%
74%

                    from [CSU, 1990]
                                          28

-------
4.6 Cd. Juarez Emissions Inventory - Alliance Technologies and others, 1990

Beginning in 1990, EPA-Region 6 worked with SEDUE, SEDESOL's predecessor, to improve the
industrial emissions inventory for Cd. Juarez. A contract was placed with Alliance Technologies
Inc. to develop an emissions inventory questionnaire for distribution to Cd. Juarez industry. A
training session, hosted by Alliance Technologies in September 1990, was attended by over 200
persons who were instructed how to do the various calculations required to estimate industrial
pollutant sources.  Plans were also made for Alliance and SEDUE personnel to visit many Cd.
Juarez industrial sites in March 1991 to assist local plant engineers in the completion of the
questionnaire.

Unfortunately, the site visits planned for March 1991 were canceled by SEDUE for reasons which
were not communicated to US counterparts. Emissions data that were collected from industry
and submitted to SEDUE offices in Cd. Juarez were not shared with the EPA as originally
anticipated.  The exercise was moderately successful in bringing together government and
industry partners and increasing emissions data return to SEDUE. However, it also illustrates
the importance of having a clearly understood policy of cooperation with environmental officials
in the Mexican government.

4.7  Short-term Winter Participate Study ~ Texas Air  Control Board, Sandia National
Laboratories,  1990

A winter-season study staged at a number of air sampling  and meteorological sites spanning
about one-month in December 1990 was funded by EPA-Region 6 with the TACB and Sandia
National Laboratories taking lead roles in study design and data analysis. The primary focus of
this study involved the collection of size-segregated particle samples followed by determination of
the elemental and carbon species composition of the aerosol.  These data were used in an initial
assessment of the applicability of receptor modeling techniques for PM10 source apportionment.
Additional measurements of local meteorology were  carried out during periods of air stagnation.
Data from a number of continuous monitors were also collected in order to characterize temporal
features of the data during the winter season. A detailed discussion of this study along with
analysis results is found in Section 5 of this report.

4.8   PMJO Modeling Plan  Assessment  and  Recommendations  —  Systems Application
International, 1991

EPA-Region 6 commissioned a study by Systems Application International to review the various
PM10 modeling approaches that might be used in state implementation plan development for the
El Paso-Cd. Juarez airshed [Gray, 1991].  The overall objective of this study was to identify
candidate receptor and/or dispersion models that could be successfully used in the El Paso-Cd.
Juarez airshed to more fully understand the release and transport of PM10 in the local airshed.

The report presents a short summary of the multi-agency PM10 scoping study discussed at length
in a later section of this  report.  Following the summary, a number of air  quality models
potentially useful in the process of state implementation plan development are reviewed and
discussed. The authors break down the modeling approaches into the following three categories.

       Dispersion Models - Gaussian plume models, lagrangian trajectory models and eulerian
       grid models are all included in this category. All these models require emission inventories
       and detailed wind information.
                                          29

-------
       Receptor Models - Included in this category are chemical mass balance models, factor
       analysis and multiple linear regression techniques.  Receptor models, in contrast to
       dispersion  models,  do not require emission  inventories, but  instead  rely  on the
       characteristics of the collected data set at sampler locations to infer the type and strength
       of pollution sources  contributing to what is measured at the collection site.

       Wind Field Models - These models are used to produce gridded wind flow information for
       use by various dispersion models. These are further broken down into diagnostic and
       prognostic techniques.

The authors present a discussion of the strengths and weaknesses of the various dispersion
models, both approved and unapproved by the EPA, for possible use in El Paso SIP development
activities. Highlights of these discussions are summarized below:

       RAM - Multiple Source Air Quality Algorithm, [EPA, 1986] ~ gaussian plume model -- EPA
       approved for PM10 — good source treatment — no  complex terrain ~ problems with
       stagnation conditions

       ISCST, ISCLT - Industrial Source Complex Short/Long Term, [EPA, 1986] -- gaussian
       plume ~ EPA guideline model for PM10 SIP development -- good point source treatment —
       problems with stagnation conditions ~ poor treatment of area sources

       PIC - Particle-in-Cell, [Marlia.  1990] - 3-D lagrangian trajectory model - not a EPA
       guideline model — area sources treated as ground level point sources ~ can determine
       contributions to overall PM10 from individual source categories — can accept complex
       terrain wind information

       WYNDvalley - Two layer Eulerian grid model, [Harrison, 1988] — not an EPA approved
       model — attempts to deal with dispersion under stagnation or near-stagnation conditions -
       requires trial and error approach for best results

       GRID.  [ODEQ,  1990] - Eulerian grid model -  not EPA guideline model  - can model
       different sources distinctly - includes met processor for hourly 3-D winds

       UAM - Urban Airshed Model, [EPA, 1990] — Eulerian grid model - EPA approved for urban
       ozone modeling - modifications required to run multiple particle sizes in the PMi0 category
       — takes complex terrain into account — computationally intensive and not well suited for
       annual modeling however alternative strategies have been suggested -- takes gridded wind
       field data from Diagnostic Wind Field Model

       RTM/UAM-V - Regional Transport Model, [Stewart, 1986] ~ Eulerian grid model - not an
       EPA guideline model ~ similar to UAM in many respects ~ forms the basis of recent UAM-
       V release ~ takes complex terrain into account — includes secondary sulfate aerosol
       formation mechanisms

A similar summary of a much shorter list of receptor models discussed by the authors is given
below:

       CMB - Chemical Mass Balance. [EPA, 1986] ~ primary receptor modeling  technique ~
       requires source profiles which are statistically fit to profiles measured at sampling sites—
                                          30

-------
       assumes no aerosol transformation from source to receptor — cannot distinguish between
       sources with similar chemical composition in their emissions

       SAFER - Source Apportionment by Factors with Explicit Restriction, [Henry, 1988] — uses
       factor analysis to extract information  on sources from data  ~ can predict source
       compositions and strength ~ source libraries not required

Finally, a summary of the various approaches to the development of wind field data required by
many of the dispersion models presented by the authors is given below:

       WEST - Winds Extrapolated from Stability and Terrain, [ODEQ, 1990] ~ a diagnostic wind
       field model — produces 3-D  (2 vertical layer) wind field each hour — extrapolates actual
       measurements from the grid — simple modeling approach

       DWM - Diagnostic Wind Model, [Douglas, 1988] — diagnostic model used for input to UAM
       ~ interpolates surface and upper air  measurements ~  calculates vertical velocity ~
       concerns about accuracy in vertical velocity calculations

       CSUMM - Colorado  State University Mesoscale Model,  [McNider, 1981] ~ a prognostic
       hydrostatic primitive equation model -- optimized for shallow slope terrain — exhibits
       questionable nocturnal flow features along highly sloped terrain

       HOTMAC - High Order Turbulence Model for Atmospheric Circulations, [Yamada, 1989] -
       - a  hydrostatic primitive equation model — predicts wind, temperature, humidity and
       atmospheric turbulence — can utilize nested grids (coarse for boundary conditions and fine
       for the area of interest)

       PSUMM4 - Pennsylvania State University Mesoscale Model, [Anthes. 1978] — hydrostatic
       terrain-following model ~ nested grids available ~ optimized for regional scales of 100 to
       1,000 km ~ costly to apply and run on a smaller grid size

       RAMS - Regional Atmospheric Modeling System — a non-hydrostatic prognostic model ~
       evolved from CSUMM — high horizontal resolution of small-scale processes possible — grid
       nesting available ~ output in spherical coordinates that require post-processing into the
       more conventional UTM coordinate system.

The authors draw several conclusions concerning the  application and use of dispersion and
receptor models to the problem at hand in the El Paso-Cd. Juarez locale. First, a recommendation
is made to  apply a dispersion model such as UAM to  PM10 dispersion in the area. The authors
note that the quality of the model output is only as good as the source information used for input.
The lack of credible emissions data from Mexico is also recognized as a potential problem in this
approach. The authors note that minor modifications would be necessary to apply the UAM model
to PM10. The model should be run in a so-called "inert" mode and similar emissions from Mexico
and US should be "tagged" separately; however, the authors do not give details here.  Secondly,
the authors recommend the use of the CMB receptor model to complement the dispersion model
efforts. Results from the CMB model may be useful in understanding sources in the region for
which emission inventory data are either missing or incomplete.
                                          31

-------
4.9 PM10 SIP Development -- Texas Natural Resource and Conservation Commission, 1991

The TNRCC in accordance with recent enactments of the US Clean Air Act have submitted a PM10
State Implementation Plan (SIP) that puts forth a detailed agency analysis of measured PM10
levels, inventoried PM10 sources  and dispersion model calculations intended to reveal PM10
pollution trends in the region along with the modeled effects of various PM10 control measures
[TNRCC, 1991J. An internal study of measured PM10 levels in El Paso during the years  1986 -
1990 revealed that exceedences of the PMJO standard occurred most frequently during  winter
months and that most occurred during winter meteorological stagnation events. The El Paso PM10
SIP also draws upon the results from special studies conducted in the area under the Annex V
program. Reference is made to the Winter 1990 Intensive PM10 study discussed in greater detail
in Section  5 of this report.  The  SIP report also describes the TNRCC's effort to develop
comprehensive emission inventories for El Paso County for the years 1990 and 1994. December
31,1994 is the CAA-mandated PM10 attainment date for El Paso County.

The TNRCC analysis was designed to predict local impact of US sources only, so no attempt was
made to account for emissions from the Mexican side of the border.  By employing Section 179(b)
of the CAA and showing that El Paso County would attain the PM]0 standard "but for" emissions
from Mexico, TNRCC would then  avoid a reclassification of El Paso's non-attainment status to
"serious" if El Paso failed to attain required air quality by December 31, 1994.

The US emission inventories were combined with meteorological data from the El Paso airport that
spanned the years 1985 through 1989.  An EPA regulatory model known as the Gaussian-Plume
Multiple Source Air Quality Algorithm (RAM) was used for predicting PM10 impacts from 1990 and
1994 emission inventories over a large receptor grid. The model assumes flat terrain which was
judged to be acceptable by TNRCC staff since major sources were understood to be at ground level
and located in the flat, valley bottom of the region. In general, model runs for one-year intervals
revealed annual average PM10 concentrations in the vicinity of 40 ug/m3 at so-called "hot spot"
receptor locations at which the highest predicted PM10 levels were encountered. Approximately
70% of this average annual PMJO value was attributed to sources in the inventory with the
remainder attributable to regional background aerosol. Similar model runs for 24-hour intervals
revealed maximum concentrations ranging from about 90 to 130 ug/m3.

The modeling analyses reveal no exceedences of US air quality standards by  1994 as a result of
US sources alone. Consequently, the TNRCC concluded that PM10 transport from Juarez sources
account for 24-hour PM10 levels above the US standard that are occasionally measured at the
various PM10 monitoring sites throughout El Paso.


4.1O  Cd. Juarez Industrial Emissions Study ~ EPA-SEDESOL, 1992-93

During September  1992, a cooperative study was conducted by EPA Region 6 along with
SEDESOL, EPA's Mexican counterpart, of air emissions from typical industrial source-types in Cd.
Juarez. This project was initiated following a specific request from SEDESOL at the June 1992
Binational Air Workgroup meeting. EPA Region 6 coordinated the involvement of personnel from
the EPA Office of Air Quality Planning and  Standards, Texas Air Control Board, California Air
Resources Board, California South Coast Air Quality Management District and the City of El Paso
in a series of one-week field visits with SEDESOL staff from Cd. Juarez, Mexico City and Tijuana.
Five facility reports were Jointly  written by US and Mexican  participants that estimated  air
emissions,  recommended easily  implemented controls and a Reasonably Available Control
Technology strategy for each industry visited. The data collected during this project provided the
first in-depth inventory of major sources in Cd. Juarez using US government personnel and gave

                                         32

-------
significant insights into similarities and differences between US and Mexican industrial plants.
The project also provided opportunities for technology  transfer between US and Mexican
counterparts.

The facility reports and raw data produced during this project are  considered proprietary
information, as specified in Article XVI of the 1983 La Paz Environmental agreement between
Mexico and the US.  Public release of the data or reports by the US is subject to the approval of
the Government of Mexico [Yarbrough, 1994].

In November 1992 and November 1993, TNRCC and EPARegion-6 personnel presented emissions
inventory training to over 100 SEDESOL engineers. The primary objective of the training exercise
was to transfer emissions gathering technology to SEDESOL staff so that they would be more
comfortable in planning and conducting future Juarez inventories. Although the training sessions
were judged to be successful in giving SEDESOL engineers the technical skills needed to carry out
emissions estimation calculations, no follow-on special Juarez industrial inventory was conducted
by SEDESOL as anticipated.

4.11 Vehicle Emissions Remote Sensing Study - University of Denver, 1993

In March of 1993 a consortium that included EPA Region 6 commissioned a research group at the
University of Denver to conduct a number of on-road measurements of vehicle emissions in both
El Paso and Cd. Juarez (Stedman, 1993A and Stedman, 1993B]. This university research group
has developed an infrared remote sensing  unit that is capable of measuring CO2, CO and
hydrocarbon tailpipe emissions from a single on-road vehicle in less than one second. An infrared
beam is aimed from a source across the road at tailpipe height, passed through a series of narrow
band filters and projected onto detectors that yield a measure of selected gases in the exhaust
plume. The system is capable of continuous sampling in a traffic stream such that the emissions
from a large number of cars can be recorded and compiled into a fleet profile for a particular area.
The license plate of the automobile is also recorded by a video system.  With appropriate labor
resources, these video data provide a means of determining vehicle year, model type and owner
for additional definition of the local vehicular fleet composition.

A prototype model of a remote sensing unit designed to concurrently measure NO emissions by
measuring its absorbance  in the UV portion of the spectrum was also tested on selected days
during the study interval.  A limited amount of data from this unit are reported as well.

Testing began on March 15 and was completed on March 25. The remote sensing apparatus was
sequentially set up at four locations in El Paso and seven locations in Cd. Juarez. Locations in
El Paso were the Yarbrough on-ramp to eastbound  I-10, the westbound on-ramp  to I-10  at
Sunland Plaza, the Altura on-ramp to southbound US 54 near Fort Bliss and the on-ramp from
both southbound and northbound US 54 onto eastbound I-10.  Locations in Cd. Juarez included
a right turn lane from northbound Lopez Mateos onto eastbound De La Razza, a left turn lane at
the San Lorenzo intersection, the northbound ramp to the Bridge of  the Americas (the only
cloverleaf junction in Cd. Juarez), a left turn lane from Thomas Fernandez onto Ave de la Industria
in one of the industrial sectors of the city,  a single lane near the Lucerna Hotel and a single
entrance lane into the parking lot of the Juarez municipal building.
                                          33

-------
A summary of the measurement results showing the mean and median values of CO  and
hydrocarbons (HC) as percent volume of total exhaust along with the number of cars monitored
at each site is given in Table 3.  Figures 11 and 12 provide additional detail of the tailpipe
emission distributions for CO and HC by apportioning the total number of cars sampled  into
deciles with an emission level  associated with each decile.  A number  of conclusions are
summarized from the report as follows:

       The average CO level compiled from about 16,000 vehicles at all sites in El Paso was 1.2%.
       Carbon monoxide levels in Juarez were nearly three-fold higher with a mean of about 3.0%
       as compiled from about 7,500 vehicles. The authors note that Juarez emission levels are
       clearly higher than typical US urban levels however, they are lower than levels measured
       in Mexico City.

       Hydrocarbon (HC) emission levels show similar trends as noted for CO, with the El Paso
       mean HC at 0.07% and the Cd. Juarez mean at 0.17%.

       Site-to-site differences for both CO and HC were minimal for the Cd. Juarez data set.
       Differences in CO and HC emissions were noted between the more affluent section of El
       Paso near Sunland Plaza and other parts of the city. The lower levels encountered near
       Sunland Plaza are attributed to a higher fraction of newer cars equipped with enhanced
       emission control equipment.

       The decile plots (Figures 11 and 12) for both CO and HC reveal differences in the
       emissions from the vehicle fleets of the two cities.  The highest CO emitting group in El
       Paso has a similar average %CO level as the three highest deciles in Cd. Juarez. Thus,
       there are more "broken" vehicles in Cd. Juarez, but the emission level of a high emitting
       vehicle is about the same in both cities.

       The lowest emitting vehicles in El Paso emit at a much lower (nearly three-fold) level than
       the lowest emitting category in Cd. Juarez. The authors attribute these differences to the
       presence of many more US vehicles possessing the newest "closed-loop" emission control
       technology.


The authors point out a two-fold utility in making such remote sensing measurements.  First,
individual high emitting vehicles can be detected and repaired or otherwise dealt with under
appropriate local agency programs.  Second,  a city-specific emission rate for CO, NO and
hydrocarbons  can be determined to reasonable accuracy since formulae are available for
converting measured average %CO values to grams of CO emitted per gallon of fuel consumed.
Fuel consumption figures can then be combined with the average emission rates to estimate total
emission levels of the El Paso and Cd. Juarez vehicular fleets for pollutant species such as CO.NO
or hydrocarbons that are  of interest in SIP development for PM10, CO or ozone control.
                                         34

-------
                              Table 3
               Summary CO and HC Tailpipe Emissions
          from the University of Denver Remote Sensing Study
Sampling Site
Yarbrough
Yarbrough
Sunland
Aftura
Eastbound 1-10
El Paso Overall
Lopez Mateos
San Lorenzo
Bridge Ramp
Cloverleaf
Industrial
Hotel
Municipal Bldg
Juarez Overall
Date
3/15/93
3/16/93
3/17/93
3/18/93
3/19/93

3/22/93
3/22/93
3/23/93
3/23/93
3/24/93
3/24/93
3/25/93

Vehicle
Count
5273
5811
4035
2220
6534
15986
391
1802
601
673
1303
1239
1631
7640
Mean
CO
(%, v/v)
1.55
1.42
0.84
1.59
1.25
1.22
3.48
2.90
2.81
2.81
2.72
3.21
3.04
2.96
Median
CO
(%, v/v)
0.63
0.47
0.26
0.53
0.37
0.37
2.73
2.20
1.72
1.96
1.87
2.35
2.41
2.18
Mean
HC
(%, v/v)
0.08
0.07
0.08
0.07
0.07
0.07
0.19
0.18
0.15
0.13
0.17
—
0.18
0.17
Median
HC
(%, v/v)
0.06
0.05
0.05
0.05
0.04
0.04
0.11
0.10
0.08
0.08
0.08
—
0.09
0.09
from [Stedman, 1993A]
                                35

-------
            9

            8

            7-
          o
          c
          o
            3-

            2-

            1-

            0
                       T     I    T     1
                                                   I     i     r
                    1234-56789   10
                                     Decile number
                                  I Juarez  •§ 0 Paso
Figure 11 Decile emissions in %CO for vehicles measured in El Paso and Cd. Juarez Jfrom
Stedman,  1993B].
           0.9

           0.8

           0.7

         00.6


         
-------
4.12 MOBILES Revisions — Energy and Environmental Analysis Inc., 1993

In order to obtain better vehicular emission estimates for Cd. Juarez, the Emission Control
Strategies Branch of the EPA commissioned Energy and Environmental Analysis Inc. to revise the
EPA MOBILE5 Vehicular Emissions Code so that it could be applied to the Cd. Juarez fleet [EEA,
1993].  The current MOBILESa model has hard-coded, non-changeable inputs for vehicle control
technologies as they were implemented in vehicles sold in the US (with the exception of the State
of California) over the past few decades. As such, the code is applicable to all US states with the
exception of California since different control technologies have been implemented there. By the
same argument, the code is also not suitable for use in Cd. Juarez. Modifications were made to
the code to allow greater input flexibility and thus applicability to Cd. Juarez. The modified code
is designated MOBILESc and is suitable for use in any US or foreign location, provided that the
appropriate input data for the vehicular fleet of interest are available.


4.13 Cd. Juarez Vehicle Fleet Characterization — Texas Transportation Institute,  1993

In the interest of better definition  of Cd.  Juarez vehicular pollutant sources,  the  TNRCC
commissioned the Texas Transportation Institute to conduct a multi-tasked study in  the city
during the summer of 1993 [TT1, 1994]. A total of six specific  tasks related to the refinement of
the Cd. Juarez vehicle fleet characteristics, designed to yield better MOBILES input, are described
below.  A summary of the more important results  is also given for each task.


       Vehicle Speed Study -  Various surveys were carried out to estimate travel time and
       average vehicle speeds throughout the city. Average speeds ranged from as low as 14.5
       mph during midday periods on undivided arterials in  the central business district to a
       high of 33.1 mph at midday on rural divided arterials. Table 4 summarizes overall average
       speeds by roadway class and time of day.

                                       Table 4
                Average Speed (MPH) by Roadway Class and Time of Day
Roadway Type
Divided Arterial
Undivided Arterial
Other Arterial
Collector
AM
27.0
20.2
22.8
17.0
Off-peak
25.4
18.9
21.6
17.0
PM
24.6
18.6
22.2
16.4
Daily
24.8
18.6
21.8
16.4
             from [TTI, 1994]
       Vehicle Miles Traveled (VMT) Mix Study - Vehicle count surveys were carried out to
       specify the fraction of total vehicle miles traveled in the urban area by each of the eight
       vehicle categories used in the MOBILES model.  The overall VMT mix by vehicle
       classification is given in Table 5. An estimate of traffic activity showing the percent of total
       24-hour activity by hour of the day is given In Figure 13.
                                          37

-------
                               Table 5
                      Overall Cd. Juarez VMT Mix
Vehicle Type
Light Duty Gas Vehic.
Light Duty Gas Truck"!
Light Duty Gas Truck2
Heavy Duty Gas Vehic.
Light Duty Diesel Vehic.
Light Duty Diesel Truck
Heavy Duty Diesel Vehic.
Motorcycles
Percent of Total
61.1
24.2
6.6
5.6
0.0
0.0
1.9
0.6
                   from fin, 1994]
      8%
      4%--
      0%
         06   08  1    1    14   16   1    2   22   24   02   04
            07   09   11   13   15   17   19   21   23   01    03   05
                              Time of Day
Figure 12 Estimate of daily Cd. Juarez traffic activity by hour of the day
[from TTI, 1993].
                                 38

-------
Operating Mode Study - Home interviews on vehicle usage, normally used to determine
vehicle mode usage, were not carried out in this study. Instead, estimates of the fraction
of vehicles operating in the area under various engine temperature conditions (e.g. percent
cold vehicle starts) were derived from data collected in other US cities.  These were
combined with average trip duration estimates for Cd. Juarez to estimate overall engine
mode usage. Not surprisingly, overall results pertaining to mode usage did not differ
significantly from other US cities.

Vehicle Registration - An extensive vehicle sampling survey was conducted to determine
the percentage of total vehicles in the city that were Mexico-registered. US-registered or
non-registered.  A compilation of all survey results revealed that about 62% of all vehicles
had Mexico registry, 15% US registry with the remaining 23% having no registry. Survey
results are summarized by age of vehicle in Table 6.

International Bridge Study - Vehicle sampling surveys were conducted at the Paso del
Norte Bridge (Sante Fe Street), Bridge of the Americas (Free Bridge) and Zaragosa Bridge
to determine typical vehicle wait times for US-bound vehicles throughout the day. The
authors observe that the number of vehicles passing over the various bridges in a 24-hour
period ranged from 10,000 to more than 31,000 in July 1993 when  the survey was
conducted. The highest traffic volume was noted on the Bridge of the Americas. Typical
wait times for crossing from Mexico to the US at this bridge were a function of the day of
the week and the time of day and ranged  from less than 10 minutes to in excess of 30
minutes with the longest times occurring  on weekdays between 1200 and 1500 hours.
About half of the vehicles passing though were of Mexican registry with the remainder of
US registry.

                                 Table 6
                 Registration Information by Vehicle Age
Age of Vehicle,
(Total No.)
Pre-1979
(99,961)
1979-1985
(156,193)
1986-1993
(53,311)
Overall
(309,195)
Percent of Total Vehicles In Category
Mexico
Registered
61
59
54
59
US
Registered
9
16
33
17
Non-
Registered
30
24
13
24
             from [TTI, 1994]
                                   39

-------
       VMT Estimation - A number of traffic count surveys were carried out and coupled with
       estimates of average vehicle trip lengths to produce an estimate of total VMT in the city.
       The results of this survey are summarized and compared with similar El Paso data in
       Table 7. The results indicate that Cd. Juarez has about 34% of the EL Paso VMT. while
       it has about 16% fewer cars than El Paso. Thus, average ear usage is less in Cd. Juarez
       than in El Paso.
                                       Table 7
                    Cd. Juarez and £1 Paso Vehicle Use Comparisons
Parameter
Total VMT
Total Vehicles
Total Population
Total VMT per Capita
Vehicles per Capita
Cd. Juarez
1993
3,394,000
309,000
1,200,000
(est.)
2.8
0.26
El Paso
1990
9,900,000
364,000
591,600
16.7
0.62
JAZ/ELP
Ratio
0.34
0.85
2.02
0.17
0.42
             from [TTI. 1994]
4.14  Cd. Juarez Brickmaker Study ~ El Paso Natural Gas Co./FEMAP, 1993

The El Paso Natural Gas Company, in collaboration with the Mexican Federation of Private Health
and Community Development Associations (FEMAP), sponsored a study of the  brickmaking
industry in Cd. Juarez [Johnson,  1994]. The primary focus of this study was to strengthen the
ability of the traditional Mexican brickmaker to earn a living from his trade in the face of mounting
pressure from Mexican pollution control agencies such as SEDESOL to curtail emissions from
primitive brick kilns.  The authors estimate that approximately 400 brickmaking operations are
located  in  Cd.  Juarez providing employment for about 2,500 citizens.   Traditionally,  the
brickmaker scavenges for fuel to fire the kiln.  Fuels include such materials as sawdust (either
clean or contaminated with plastics or glues), used motor oil and rubber tires. Use of these fuels
results in appreciable pollution emissions with the brickmaking industry ranked the fourth largest
pollution source in Cd. Juarez by some. The authors report accomplishments in the following
areas as a result of this on-going project:

      Process and Economics Improvement - The newly-formed coalition has set guidelines
      for the use of clean burning fuels, standardized brick dimensions, and brick price
      stabilization.

      Testing Burner Efficiency - Studies were conducted to evaluate the efficiency of various
      propane burners used in kiln firing operations. Emissions rates of pollutant gases such
      as SO2, CO and NOX were measured to optimize burner design and fuel usage.
                                         40

-------
       Pollutant Emissions from Various Fuel Types - Emission testing was carried out on the
       various fuels traditionally used to fire the brick kilns.  The authors report measurable
       levels of such pollutants as CO, NO,  NO2 and SO2 for such fuels as clean sawdust, used
       motor oil, old propane burner, contaminated sawdust and new propane burner.  The
       contaminated sawdust category showed the highest emission rates for all pollutants.
       Particularly high levels of CO (2,700 - 29,000 ppm) and SO2 (20 - 225 ppm) were measured
       for contaminated sawdust.  The authors do not report pollutant emissions in terms of
       emission factors (mass of pollutant per mass of fuel consumed) however. The emission
       measurements did not include an analysis of the elemental composition of the particulate
       emissions. The authors noted that bulk motor oil was analyzed for elemental content and
       observed  that  the oil contained  significant  quantities  of lead,  arsenic,  chromium,
       cadmium, copper, selenium and mercury.

       Improved Kiln Design - Various Innovative kiln designs were tested at a local training
       center set up in Cd. Juarez as a part of the project. Kiln designs were evaluated with
       regard to heating uniformity, brick processing time and waste heat recovery. The authors
       note that in many cases, brick production  and quality were improved by implementing
       these changes.

4.15 Pollutant Emissions from Residential Heaters — University of Utah, 1993

A consortium of southwestern US universities, known as the Southwest Center for Environmental
Research and Policy (SCERP)  has also  secured EPA-Headquarters funding to study air pollution
issues along the US-Mexico border. In recent years, the University of Utah, one of the member
institutions, has conducted studies to assess the air emissions from domestic heating systems
commonly used in Cd. Juarez.  The  overall objective of the project was to develop  low-cost
technology to reduce toxic hydrocarbon emissions from domestic heating and other Incineration
systems fired with various fuels [Lighty, 1993]. Studies were conducted jointly with the Technical
Institute of Juarez to measure emissions from a typical residential heater using three fuel types,
namely, Mexican-made pallets, US-made  pallets and scrap particle board covered with plastic
laminate. Emission measurements included on-line determinations of CO, THC, CO2, O2 and NO
along with speciation of volatile hydrocarbons using a fast-response GC-MS. Particulate matter
samples were also collected using EPA Method 5H. A summary of gas emission factors for the
various fuels tested is given in Table 8.

                                       Table 8
                     Gas and Total  Hydrocarbon Emissions Factors
                            for Selected Waste Wood Fuels
Fuel Type
US Pallet
Mexico Pallet
Particle Board
AP-42 Residential Wood
(EPA, 1985]
CO, g/kg
60 ±3
37 ±9
95 ±4
61.1
THC, g/kg
4.7 ± 0.3
3.5±1.3
3.1 ±0.6
95.1
NO, g/kg
0.8 ±0.1
0.7 ±0.1
3.7 ±0.1
0.9
             from [Lighty, 1993]
                                         41

-------
The morphology of collected smoke particles was examined using scanning electron microscopy.
The investigators report the presence of chain agglomerate structures indicative of soot particles.
Further analysis was done in both the gas phase and the collected particulate material using an
ion trap GC-MS  system to  speciate volatile  and semi-volatile hydrocarbons.   Four major
compounds identified In the samples were furan, benzene, toluene and furaldehyde. The presence
of polynuclear aromatic compounds was also reported. The Investigators note that analysis for
chlorinated hydrocarbon species was not carried out in this set of experiments, however they
observe that the laminated particle board contained significant amounts of chlorine which would
likely produce chlorinated organic species during combustion.

4.16  Upper Air Wind and Temperature Data Collection - University of Texas at El Paso,
1993

The Electrical Engineering Department at UTEP was funded as one of the member universities of
the Southwest Center for Environmental Research and Policy to install and operate a three-axis
915 MHz wind profiler and acoustic sounding system for the collection of winds and temperature
aloft data at the UTEP campus.  The system was installed and operational by October 1993.
Temperature data are available from about 100 to 600 meters above ground level. Wind data are
available from 100 meters to 2,000 meters above ground.  Data are processed and archived as
hourly, dally, monthly and seasonal averages and variances. The system was used for a winds
aloft data cross comparison in a recent lidar technology demonstration project briefly discussed
in the final section of this report.


4.17  Oxygenated Fuel Use in El Paso-Cd. Juarez - SEDESOL and others, 1993

El Paso is classified as a moderate CO non-attainment area.  While the US Clean Air Act
Amendments of 1990 do not require El Paso to Institute a winter oxygenated fuels program, the
City of El Paso nevertheless chose to Implement one, beginning In December 1991. The initial
oxygenate blend was 2.4%, and was raised in winter 1992-93 to 2.7%. The city made the decision
to initiate such a program as a result of the apparent CO reductions experienced by other
Southwestern US cities, such as Albuquerque, with oxy-fuels programs. At this point In time no
systematic studies have been carried out to examine the effects of oxygenated fuels on winter
season CO levels in El Paso. However, average CO levels at least qualitatively reveal a decline over
the past several years.

The  US EPA  and the  Government of  Mexico have discussed the possible introduction of
oxygenated fuels in Mexico as well.  A pilot program was carried out from January 1, 1993 to
February 15, 1993 during which Mexican officials allowed oxygenates to be included in unleaded
gasoline supplied by US refineries to Cd. Juarez markets.  The effects of this introductory program
have not been systematically evaluated at this point In time.


4.18  Other Activities - Paso del Norte Air Quality Task Force, 1993

A coalition of government, industry and non-government organization representatives came
together in  1993  to form the Paso del Norte Air  Quality Task Force.   The group, with
representatives from EPA, SEDESOL, Cities of El Paso and Cd. Juarez, the States of Texas, New
Mexico, Chihuahua as well as non-government organizations and the public originally came
together as an advisory committee to the TACB, TNRCC's predecessor.  The organization was
elevated to a task force in September 1993 when the TACB was combined with other Texas state
agencies to form the TNRCC.

                                         42

-------
At present, the group, working in concert with the Environmental Defense Fund, is focusing on
short-term air pollution emission reduction strategies that can be easily implemented.  Current
efforts at  air emissions reduction include:  developing strategies for the reduction of VOC
emissions from small auto body shops; defining actions to reduce excessive idling vehicle queues
at the international bridges; and, assisting in the training of mechanics who are charged with the
repair of vehicles that fail the inspection and maintenance program in Cd. Juarez. The task force
has also contributed to the Cd. Juarez Brickmakers study summarized  earlier in this report.

A longer-term goal of the Task Force is the development of an International  Air Quality
Management District (IAQMD). The formation of this international air shed has been proposed
by the Task Force as an annex to the 1983 US-Mexico Border Environmental Agreement.  The
IAQMD would consist of a governing board of directors that would encourage common approaches
to enforcement, monitoring, regulations and  public outreach among the various air pollution
control agencies in the airshed.  Economic incentive programs would also be proposed to aid in
air pollution reduction in the airshed. For example, one such economic incentive might involve
the designation of international air pollution credits trading, much like those implemented as a
part of  the recent US  Clean Air Act legislation.  This scheme would enable a US  industry
interested in plant expansion on the US side of the border to finance air pollution controls on the
Mexican side of the border, and in the process, "bank" air pollution credits  for use as an offset for
pollution increase on the US side that may arise as a result of plant expansion.  Provided that the
Mexican plant pollution reduction is greater than the US pollution added (by some ratio  to be
agreed upon), the credit trading system offers some attractive alternatives for control of Cd. Juarez
industrial pollution sources through financing by US investments while allowing US industry to
expand  in an otherwise highly restrictive non-attainment environment in  El Paso.

The EPA and the US Department of State are  currently reviewing the proposal for the formation
of a formal IAQMD, forwarded by the State of Texas on behalf of the Task Force in September
1993.

4.19  Comparison of Vehicle Emissions Inspection and Maintenance Programs in Cd. Juarez
and El Paso ~ Paso del Norte Task Force, 1994

A comparison of the El Paso and Cd. Juarez vehicle emissions inspection and maintenance
program was carried out in early 1994 by a member of the Paso del Norte Air Quality Task Force
[Rincon, 1994].  In his introductory comments, the author notes the significant population growth
of the Pas del Norte trade area, as shown in Figure 14. The author's estimates of total population
in the region by the year 2000 is on the order of 2.6 million people. Motor vehicle registrations
show similar dramatic increases over the past 20 years, particularly in  Cd. Juarez.

Other observations made by the author with respect to the I/M programs in both El Paso and Cd.
Juarez are summarized in the following paragraphs.

       Tailpipe emission standards  for CO and hydrocarbons in El Paso and Cd. Juarez are
       summarized in Figure 15. In general,  the requirements are more stringent in El Paso for
       both  CO and hydrocarbons.  For  1981 and newer vehicles. El  Paso's standards are
       markedly more stringent.

       In 1993, about 160,000 vehicles were inspected in Cd. Juarez, a  number equivalent to
        about 50% of the  total number of registered vehicles  in Cd. Juarez.  The Municipal
       Ecological Committee reports a failure rate of 9.5% of all vehicles tested.
                                          43

-------
             3.0i
                                                                El Paso
                                                                Cd. Juarez
                                                                ES3
                                                                LasCruces
                                                                Total
                       1970
1980       1990
     Year
2000
Figure 14  Population trends in El Paso and Cd. Juarez (from Rincon, 1994].

     As of April 15, 1994,  about 96,000 light duty trucks and passenger vehicles were
     inspected in Cd. Juarez. A failure rate of about 48% is reported, which stands in marked
     contrast to 1993 inspection data.

     Inspection data for El Paso as tabulated by the TNRCC for 1993 is shown graphically in
     Figure 16. The inspection failure rate is shown by vehicle year. The overall 1993 failure
     rate for all vehicles was about 15.3%.

     The author notes that current Mexican import laws tax imported US vehicles less than 5
     years old at 100% of their import value. These restrictions result in an older vehicle fleet
     with greater tailpipe emissions in Cd. Juarez as compared to El Paso.

     Many of the inspection and maintenance centers in Cd. Juarez do not have fully trained
     mechanics  doing I/M testing and repairs.  The Comite Municipal de Ecologica has
     implemented introductory technical seminars for I/M inspectors to assess thei1- level of
     knowledge and provide instruction on analyzer operation (See Section 4.20).

     Educational programs  are needed in Cd. Juarez  in order to raise awareness of the
     population to the requirements and benefits of an I/M program.
                                        44

-------
              @
              .1 5'
              .2

              I 4
              Q.
              O
              O
                   Q—B—a '••\o
                                  V
                                     -B	B	S	B-
A
    IS	S	B	S	B-

    1   I   I	1   I—
3>

                                        -»—<• m "••-•*- «• tt —«t - •• *	•  • -r- m ~~m— m
                                                                          800
-700  E
     a.
     &

     ®
     o
 500  '8
     E
     a>
 400  -g-
                                                                          -300

                   75 76 77 78 79 80 81  82  83  84 85  86  87  88  89  90  91  92
                                         Vehicle Year
—
us
CO -
•+— JAZ
CO ->
*- us
HC
~e-
JAZ
HC
  Figure 15 Vehicle CO and hydrocarbon emission standards established in El Paso and Cd.
  Juarez I/M programs [from Rincon. 19941.
The author also lists a number of recommendations for improvement of the Cd. Juarez I/M as
summarized below.

       Test all vehicles in the Cd. Juarez fleet under an expanded Cd. Juarez I/M program.

       Develop an inspector/mechanic "hot line" for prompt resolution of technical questions.

       Implement data management techniques to track I/M program performance.

       Develop educational programs that are targeted to the general populace in order to raise
       their collective environmental consciousness.

       Prioritize both traffic flow control measures and road pavement projects as a means of
       achieving pollution reduction associated with vehicle use in Cd. Juarez.
                                           45

-------
             3Oi
                    75 76 77 78 79 80 81
82 83 84 85 86 87 88 89 90 91 92 93 94
  Model Year
  Figure 16 Vehicle failure rate by vehicle year encountered during 1993 in the El Paso I/M
  program [data from Rincon, 1994].
4.20 Technology Transfer Session with I/M Technicians - Colorado State University, 1994

Colorado State University, under contract to the EPA Office of Mobile Sources, performed two
hands-on technology transfer sessions with Juarez I/M technicians in 1994 [Friedt, 1994). The
sessions were designed to better equip Cd. Juarez commercial mechanics to diagnose and repair
typical vehicle  emissions control system problems. The program was specifically developed to aid
Juarez Comite Municipal de Ecologia in enhancing the positive impacts of the city's I/M program.


The project was initiated with a  questionnaire administered to  14 automotive technology
instructors in Cd. Juarez. The questionnaire was designed to assess the training needs of the Cd.
Juarez automotive technicians and to assist in the development of the training course content.
Survey respondents indicated a lack of training aids such as exhaust gas analyzers for instruction
sessions on engine diagnosis and tune-up.  Survey respondents also ranked academic needs for
a course on mechanics training in emissions system repair as follows: (1) more effective teaching
techniques; (2) emissions control systems training; (3) chemistry of pollutant  formation; (4) fuel
injection systems; and (5) electronic ignition systems.

The project team put drew a number of conclusions and recommendations from the survey results
as briefly summarized below.
                                          46

-------
In general, the technician group In Cd. Juarez requires additional training In the areas of
teaching methodology, causes of excessive vehicle emissions, emissions control system
function and diagnosis/repair techniques.

The learning styles and reading skills of potential course participants vary widely and
must be taken into account when designing course material for this technician population.

The instructional course should be a two-phase program with an approximate 6-month
interval between phase  one and two. This Interval would allow an assessment of the
degree of success reached with the Introductory phase of the course, allowing further
adaptation of the second phase of the course to student needs.

A communication network should be established between Mexican and US border
community colleges with Instructional capabilities In these areas to provide continued up-
to-date training approaches and skills for automotive repair Instructors and technicians.
                                    47

-------
48

-------
5.0 Short-term Winter Season PM10 Study

5.1 Study Purpose and Scope

A winter-season study staged at a number of air sampling and meteorological sites spanning
about one-month  In December 1990 was funded by EPA-Region 6 with the TACB and Sandia
National Laboratories taking lead roles in study design and data analysis. The primary focus of
this study involved the collection of size-segregated particle samples followed by determination of
the elemental and carbon species content of the aerosol.  These data were used in an initial
assessment of the applicability of receptor modeling techniques for PM10 source apportionment.
Additional measurements of local meteorology were carried out during periods of air stagnation.

5.2 Measurement Techniques

Multiple samplers were deployed in this study in an effort to more completely characterize the
pollutant  composition  and spatial distribution during winter stagnation events.   Funding
limitations prevented complete pollutant characterization during the entire study interval, however
the aggregate of various measurements,  although not continuous, did allow  a reasonable
characterization of winter season air quality. Table 9 lists the measured parameters along with
instruments and lab analysis, where applicable.

                                       Table 9
                          Air Pollution Parameters Measured
                           In the 1990 Winter Season Study
Measured Parameter
PM,0 Concentration (24-hr avg)
Aerosol Concentration (12-hr avg)
Fine, Coarse & PM,0
Elemental Concentrations (12-hr avg)
Fine, Coarse & PM10
Elemental and Organic Carbon Cone.
(12-hr avg) Fine, Coarse and PM)0
PMto Concentration (1-hr avg)
Aerosol Light Scattering Coefficient
Inorganic Gases and Aerosol
Gas Criteria Pollutants (CO, NOX, O3, SO2)
Semi-volatile organics
Wind Speed & Direction
Temperature
Winds aloft, atmospheric mixing height
& turbulence
Winds, absolute humidity, temperature
aloft
Winds, humidity, temperature & aerosol
light scattering coefficient aloft
Instrument
Hi-vol Sampler
Dichotomous Sampler w/
Teflon Filter
Dichotomous Sampler w/
Teflon Filter
Dichotomous Sampler w/
Quartz Filter
Beta-gauge sampler
Nephelometer
Annular Denuder
Continuous Monitors
Hi-vol sampler with PDF
Cartridge
Anemometer
Thermocouple
Acoustic sounder
Rawinsonde
Tethersonde/Nephelometer
Lab Analysis
Gravimetric analysis
Gravimetric analysis
X-ray fluorescence
2-stage combustion
analysis
None
None
Ion chromatography
None
Gas Chromatography/Mass
Spectroscopy
None
None
None
None
None
                                          49

-------
5.3 Site Descriptions

5.3.1 Site selection process

A site survey trip was made to the El Paso-Cd. Juarez area In the spring of 1989 to inspect
existing and potential new sites for meteorological and air monitoring purposes. Discussions were
held with EPA, TACB, El Paso City-County Health District, ASARCO Corp., SEDUE and other
contractor representatives.  Sites were selected based on such criteria as topography, pollution
sources,  available facilities  (shelter and power), security and personnel access. Final detailed
selections were made by the EPA contractor designated to install field instrumentation (this
contractor later withdrew from the proj ect before completing instrumentation installation). Sandia
acquired towers, meteorological equipment and data acquisition systems for seven sites—five in
El Paso and two  Cd. Juarez.

5.3.2 Air sampling enhancement at existing sites

Carrying out a preliminary PM10 assessment using receptor modeling approaches required that
sites be  equipped with  sampling instrumentation that would provide a suitable data set for
subsequent analysis. This study, insofar as possible, made use of existing sites operated by either
the City of El  Paso or the TACB in order to minimize setup and instrument acquisition costs.
Attempts were also made to use existing SEDUE sites in Mexico; however, a survey of these
Mexican  sites revealed  that many existing PM10 or meteorological  sites did not satisfy the
minimum criteria established for such issues as available space, security and proximity to local
obstructions.  For those sites selected on both sides of the border, augmentation consisted of
installation of dichotomous samplers which provide for the separate collection of both fine particle
mass (FPM)1 and coarse particle mass (CPM) aerosol fractions. At selected sites, nephelometers
were also installed to give a real-time measure of the aerosol light scattering coefficient, and by
inference, aerosol mass concentration. These continuous measurements of aerosol concentration
provided a data set from which temporal and spatial variations in pollutant concentrations could
be examined.  A summary of site names, locations and associated instrumentation is given in
Table 10. A topographical map of the area showing site locations is shown in Figure 17. A brief
description of  each of the sites follows.

5.3.3 General site descriptions

El Paso Airport - Meteorological

  The National Weather Service wind set at the airport is mounted on a 10m mast north of the
FAA-NWS Building, which  is about 12 km northeast  of downtown El Paso.  Routine hourly
weather observations are taken round the clock and are regularly disseminated through weather
communication channels.  Exposure (about 40 m above the valley floor) is representative of the
flat plateau northeast of town that eventually merges with the Tularosa Basin to the east of the
Franklins.  ELP  is also an upper-air observation station with twice daily radiosondes taken at
noon and midnight universal time (0500 and 1700 mst).  Rawinsonde balloons are released
between 45 minutes to an hour before the nominal observation times.  A typical balloon ascent
to 10 mb, or 31 km msl (above mean sea level) takes nearly two hours to complete at a balloon
ascent rate of 5 ms"1.
   1 The Fine Particle Mass (FPM) fraction encompasses those particles less than 2.5 \un aerodynamic
diameter. The Coarse Particle Mass (CPM) encompasses those particles larger than 2.5 jim and less than
10 ^m aerodynamic diameter. Taken together they represent PM,0.

                                          50

-------
                                    Table 10
           Winter Season Study Site Locations and Monitoring Equipment
Site Name - Abbreviation
Sunland Park, NM - SPK
Lindbergh. TX - LND
Vilas School. TX - VLA
CAMS-12(UTEP)-UTP
Northeast - NET
Ivanhoe - IVN
Riverside - RVR
Moon City - MCY
Border Patrol - BRP
ELP Airport - ELP
Fort Bliss - FBL
Sun Metro - SUN
Tillman - TIL
CAMS-6 - CAM
Chamizal - CHM
Ascarate Park - ASC
Advance Transformer - ADV
Tecnologico • TEC
Site
Coordinates
UTM
352552 East
3518673 North
349919 East
3525389 North
357941 East
3514810 North
357450 East
3515600 North
366777 East
3530156 North
374670 East
3517550 North
369926 East
351 1603 North
378400 East
3504700 North
359380 East
3513390 North
368590 East
3519480 North
365080 East
3520950 North
357900 East
3514470 North
359620 East
3514400 North
359270 East
3514980 North
362120 East
3515200 North
36690 East
3513790 North
362030 East
3506820 North
367910 East
3509770 North
Site
Elevation
m, msl
1141
1195
1100
1170
1197
1210
1120
1117
1133
1195
1178
1138
1135
1138
1128
1126
1167
1123
Site
Equipment
PMIO Hivol
PM)0 Hivol
PM,0 Hivol
PM.o Hivol, CO
PM10 Hivol
PM,, Hivol, CO, Met
PM10 Hivol
Met
Met
Met
Met
Dichot, Nephelometer. Met
PM10, CO, Met
PMIO Hivol, Dichots, Denuder,
Nephelometer,
CO, NOX, PUF Hivol, Met
PM10 Hivol, PM,0 Beta Gauge,
Dichots, Nephelometer, CO,
Met, Met Sounding
CO
PM10 Hivol. Dichot, CO. Met
PM10 Hivol. Dichot, CO. Met
Note:  The five specially-equipped sites are shown in bold type.
                                       51

-------
                                              NET
   3530000-/
         LMD
   3525000-
   3520000-
S

-------
Sun Metro - Meteorological and Air Sampling

A 10 m tower with wind and temperature sensors and data logger was installed in a fenced storage
yard in  the extreme northwest comer of the bus yard on the west side of town.  Here, the
Interstate 10 freeway, and two railroad lines squeeze together to turn from their westerly heading
toward the northwest to follow the river course (and hence, border) out of Texas, into New Mexico.
The exposure represents the opening of the canyon containing the Rio Grande as it enters the
broad valley to the southeast. The site is about 1.5 km west of downtown El Paso and about 2
km east of an active quarry operation located in the northwest outskirts of Cd Juarez across the
river.  Air monitoring  equipment consisting of two dichotomous  samplers and an integrating
nephelometer were also installed  at this site.  The dichotomous samplers were positioned on a
platform about 1.5 m above ground level since a shelter was not available at  this site.  The
nephelometer was installed in a small unheated shelter adjacent to the dichotomous samplers.
The Sun Metro site was selected approximately one week prior to  the onset of the air sampling
study and came about as a result  of an inability to find a suitable site in the southern downtown
district of Cd. Juarez.  Instrumentation intended for this third met and air sampling site in Cd.
Juarez was then moved to the Sun Metro site on short notice.

Border Patrol Building at del Norte Bridge - Meteorological

The US Immigration and Naturalization Service (INS) graciously allowed us to mount wind and
temperature sensors on their 15 m communication tower located between the two  one-way
bridges, about two blocks north of the river, halfway between the downtowns of both cities (one
km north and south). The exposure here is now flat, having smoothed out from the slopes at Sun
Metro, just 2 km distant to the northwest.

Fort Bliss - Meteorological

The US Army allowed us to install a 10 m tower in a yard on their post at Fort Bliss. This site is
about 4 km northwest of the airport site and about 9 km north-northeast of downtown El Paso.
The exposure is flat,  just at the edge of the  plateau before  it ascends toward the Franklin
Mountains' ridge line 6 km to the west.

Ivanhoe Fire Station -  Meteorological and Routine Air Sampling

The El Paso City-County Health District borrowed two crank-up towers from El Paso Natural Gas
Company and arranged for one to be installed at the Ivanhoe Fire Station in the far northeast part
of the El Paso. The site is located  about 6 km east-southeast of the airport site and about 16 km
east-northeast of downtown at the southwest  comer of Ivanhoe and Lee Trevino Drives.  The
elevation, at  1210m msl, is the  highest of the installed stations. The exposure is flat in all
directions with only residential and small commercial buildings in the vicinity. In addition to the
meteorological equipment, instruments for CO. O3. and PM10 were  routinely operated by El Paso
City-County personnel at this site.

Moon City Clinic -  Meteorological and Routine Air Sampling

This El  Paso site was  at the City-County Health Clinic in the far  southeast valley some 21 km
southeast of downtown El Paso. The site is in valley bottom land, about 4.5 km northeast of the
Rio Grande at an elevation of 1117 m msl (the lowest  elevation site in the study) near the
intersection of North Loop Road  and Old Hueco Tanks Road. The other borrowed tower was
installed next to the building, through a tree, which should not seriously  compromise the wind
measurements at the low wind speeds which are of most Interest in this project.

                                          53

-------
Institute Tecnologico de Monterrey - Meteorological and Air Sampling

This Cd. Juarez site is on the campus of the Institute Tecnologico y de Estudios Superiores de
Monterrey (Campus Juarez), near the corner of Avenida A. J. Bermudez and Avenida Tomas
Fernandez, about 0.5 km due south of the Hipodromo-Galgodromo (Juarez Race Track). It is in
eastern Cd. Juarez, about 10 km southeast of downtown El Paso. The exposure is flat bottom
land with many one story maquiladores buildings in the vicinity. In addition to the 10 m weather
tower, PM10 and dichotomous samplers were positioned on site along with a trailer containing CO
and O3 continuous monitors.

Advance Transformer Co. - Meteorological and Air Sampling

This Cd. Juarez site is about  6.5 km south-southeast of the downtown section near the
intersection of Calle El Cid and Calle Beckett just off Calle Magneto. The site was located on the
north side of the Advance Transformer Co. maquiladora at an altitude of 1167 m msl, about half
way between the Rio Bravo and Sierra Juarez  crest to the southwest.   The  exposure is
characterized by many one story buildings on gently sloping ground whose fall line aims toward
the east-northeast. The area is surrounded by maquilas and small tile/brick kirns which at times
were observed to emit copious quantities of smoke. Continuous monitors for ozone and carbon
monoxide as well as a hivol PM10 sampler were also located at this site. For this short term study,
the site was also augmented with two dichotomous samplers with quartz and teflon filter media.


Chamizal National Memorial Park - Meteorological and Air Sampling

This site, on the north side of the Rio Grande just 3 km east-northeast of downtown El Paso at
an altitude of 1128 m msl, is on park land formerly known as Cordova Island.  The exposure here
is superior in that no structures of consequence are within at least 1 km in any direction, and the
terrain is quite flat near the river bottom. The El Paso City-County Health District maintains an
air monitoring site here that includes a meteorological tower,  a hivol PM10 sampler, a beta-gauge
PM10 sampler and continuous monitors for ozone and carbon monoxide.

Personnel from the Atmospheric Sciences Laboratory  set up a 10 m tower at Chamizal Memorial
which included UVW propeller anemometers at 10 and at 2 m. thermometers at  10 m,  2 m,
surface and soil (-0.05 m) as well as a radiometer (PSP).  Data were taken on a Campbell data
logger in 15 minute intervals. A Remtech acoustic tri-axial doppler acoustic sounder (or SODAR -
 sound detection and ranging) was placed about 25 m away from their tower to give wind profiles
up to about 1000 m agl. Some problem was experienced with this unit, however, because of the
presence of the nearby Border Highway which  contributed substantial background noise
interference and limited the sounder's range. The data from the ASL system were later supplied
to Sandia in hardcopy and diskette form.

Sandia personnel set up a portable radio theodolite unit (AIRInc., "Intellisonde") to take soundings
during selected periods of the two-week measurement interval. Normally, sounding balloons were
released at 0800 and 1200 mst to better define atmospheric stability and upper air flow in the
period between the 5 AM and 5 PM NWS soundings, as well as to cover the additional vertical
profile from valley bottom to airport plateau top,  37 m higher.  Data from this system  were
telemetered, processed and recorded every 5 seconds to about 10,000 m msl.

A helium-filled tethered balloon known as a tethersonde (AIR Inc.) equipped with a meteorological
sensor package and a  Sandia-designed integrating nephelometer  was  also operated at the
Chamizal site. The sensor package was winched up to a maximum profile height of 250 m agl as

                                         54

-------
allowed by FAA regulations. Flights started at sunrise and continued until about 1400 mst
(except the first day when they went until sunset).  About 90 minutes were required to complete
a calibration and round trip profile.  Measurements of pressure, temperature, humidity, wind
direction,  wind speed and aerosol light scattering coefficient were telemetered to the ground
receiver and recorded to disk file at 5 s intervals.

The Chamizal site was also augmented with two dichotomous samplers with Teflon and quartz
filter media and an integrating nephelometer equipped with a heated inlet.  Data acquisition at
this site was accomplished using an existing City-County system and a separate data logger for
the nephelometer.

CAMS6 - El Paso

The Continuous Air Monitoring Station 6 (CAMS6) is one station in a network of Texas Air Control
Board air sampling sites located across the state of Texas and was the most extensively equipped
site in the study. This site is located on the northeast edge of the downtown El Paso district near
the intersection of Campbell Street and Interstate 10.  Instrumentation on the roof of the shelter
included   a meteorological tower and associated sensors, PM10 samplers,  two automated
dichotomous samplers with quartz and teflon filter media, an additional dichotomous sampler for
microscopy study, annular denuder samplers and a high volume sampler equipped with a PUF
cartridge for semi-volatile organic sampling. Samplers inside the shelter included continuous
monitors for sulfur dioxide, carbon monoxide, oxides of nitrogen, ozone and a nephelometer.

Air quality and wind measurements were also available at two other TACB continuous monitoring
sites—one at Ascarate Park (CAMS-30), 8 km east of downtown, and the other at UTEP (University
of Texas  - El Paso,  CAMS-12).  Wind data were also obtained from the ASARCO Company's
Executive Center 30 m tower, 5 km northwest of downtown; and from the New Mexico Department
of Environment's Sunland Park station (6ZG), 8 km northwest of downtown, northwest of the
Sierra de Cristo Rey, in the south end of the Mesilla Valley.

5.4 Study Results

5.4.1 Meteorological Measurement Results

5.4.1.1 Description of the overall meteorology during the study interval

The synoptic weather maps during the first week of December showed a large surface high
pressure system centered over the Southwestern US (Great Basin and southward along the
Continental Divide) with only weak easterly surface winds over southern New Mexico and far west
Texas.  The flow aloft was dominated by a broad ridge of high pressure over the Continental Divide
area which kept many short wave disturbances well to the north.  Skies were clear,  and winds
were light, which is the "ideal" situation for air stagnation conditions to begin and persist. This
stable  meteorological situation prevailed until December 11, when a low pressure trough aloft
approached the region from the west causing the ridge to collapse and the winds to increase,
thereby ventilating the valley.  High winds also made flying tethered balloons  hazardous as a
result of power lines adjacent to the Chamizal site. For the next two weeks many short wave
disturbances traversed the area causing unsettled weather until just after Christmas. However,
as noted above, the meteorology on December 8-10 produced intense stagnation conditions which
caused enhanced pollution concentration.
                                          55

-------
5.4.1.2 Typical network winds during the stagnation period

As was expected, the behavior of the near surface winds strongly depended upon location. At Sun
Metro the wind was nearly always from the northwest as air drained out of the Mesilla Valley
through the pass.  Figure 18 shows the direction and speed curve for the 7th through 11th.
Notice the direction hovering around 310° except during the afternoons of the 7th and 8th when
a more south and southwesterly direction prevailed until 1800 hours each evening when drainage
flow resumed.  The dichotomous sampler at Sun Metro showed its highest concentration during
the evening of the 8th after the direction had been averaging southwesterly for over 6 hours and
the speed decreased from three to two to one ms'1. A notation made by a. technician in a log book
that evening while at that site to service some equipment reads "...very murky!"

At Tecnologico, away from immediate terrain influences, the direction variability was extreme. See
Figure 19 for a plot of direction and speed from this station. Speeds were nearly always less than
2 ms"1 until 0600 on the 11th as they started to increase toward 6 ms"1 by 1600 mst. At Advance
Transformer there was slightly more order as can be seen from Figure 20. Here the wind speed,
again less than 2 m s"1 until the 11th, tended to come from the southeast in the morning and
swing around  from  the west  and  northwest by evening and  night.  This picture fit with the
expectation of upslope flow when the sun shines on the eastern slope of Sierra Juarez in the
morning, followed by down slope drainage  as the sun moves to the west side of the range by
evening.  In general, variable wind directions are to be expected during periods of light winds.

For Chamizal, a pair of plots showing temperature/humidity and wind direction/ wind speed are
shown for the 24-hour period from midnight to midnight on the 8th in Figure 21. The speed here
rarely exceeded 1 m s"1, but the sensor was a UVW triaxial propeller set whose direction cosine
response may be questionable, especially at low speeds. After about 0015 mst the direction, with
an indicated speed of less than 0.5 m s"1, veered slowly from east through south through west
through north to northeast by 0300 where it remained, with a speed of slightly greater than one
1 m s"1 until 0630. From about 0700 until 1730 flow was variable from the south, backing around
to northeast by 1745. Figure 22 shows sodar winds at heights of 100 and 150 m above ground.
These data show the generally northerly (between northwest and northeast) flow until about 0600.


5.4.1.3 Sounding data from Chamizal tether- and rawin-sondes

Figure 23 shows temperature sounding data for midmoming of the 8th from both the radiosonde
to 6000 m msl, and the tethersonde to 250 m agl (above ground level) respectively. The latter plot
essentially fits  inside the bottom rectangular grid box of the radiosonde plot. The agreement is
quite satisfactory. The data for the 8th are being emphasized here because, as discussed later,
that is the day selected to test the Diagnostic Wind Model.  In general, the radiosonde profile
shows marked stability (inversion) to about 822 m agl with less stability to 2500 m agl (3630 m
                                         56

-------
                EL PASO - SUN METRO  (7-11 DEC 1990)
        &
        cc
        Q 20--
        Q
        UJ
        O
        CO
           00:00  12:00 00:00 12:00 00:00  12:00 00:00 12:00 00:00 12:00 00:00
                                   TIME



Figure 18 Wind speed and direction for December 7-11 at the Sun Metro site.
             CD JUAREZ - TECNOLOGICO (7-11 DEC 1990)
                 	•*•
00:00
                12:00 00:00 12:00 00:00 12:00  00:00 12:00 00:00 12:00 00:00

                                   TIME
Figure 19 Wind speed and direction for December 7-11 at the Tecnologico site.


                                    57

-------
          CD JUAREZ - ADV TRANSFORMER (7-11 DEC 1990)
00:00 12:00 00:00 12:00
                            00:00 12:00  00:00 12:00 00:00 12:00 00
                                 TIME
00
Figure 20 Wind speed and direction for December 7-11 at the Advance Transformer site.
                                  58

-------
                   EL PASO - CHAMIZAL, 1990 DEC 8
          20
       UJ
-5-
00:00   04:00
                           	-V-4-/-	-,	20
                                                       10
                        08:00   12:00    16:00
                             TIME, MST
                                             100
 20:00   00:00
                                                   (A)
                  EL PASO - CHAMIZAL (1990 DEC 8)
           00:00   04:00
               08:00   12:00   16:00
                   TIME, MST
20:00   00:00
                                                             (B)
Figure 21 Temperature/humidity (A) and wind speed/wind direction (B) for December 8 at
the Chamizal site.
                                   59

-------
                  CHAMIZAL SODAR,  DEC 08,  100m
          400-
           00:00  04:00   08:00   12:00   16:00   20:00   
-------
               Chamizal Radiosonde,  0926mst 8 Dec 90
         6000

         5500-

         5000-
       CD
         1000
            -15
-10
-5        0
Temperature, C
                                                               (A)
             CHAMIZAL TN-SONDE, #3D 0920M 08 DEC 90
       - 150"
                            2       2.5      3
                             TEMPERATURE, C
                                                               (B)
Figure 23 Chamizal temperature sounding data from the morning of December 8 from
radiosonde (A) and tethersonde (B).
                                  61

-------
msl).  The profile above this is closer to neutral at a lapse rate of about 8.0 "Clkm)"1 (the neutral
adiabatic process lapse rate is 9.8). Much detail can be seen in the stratification which is typical
of stable profiles.

Figures 24 through 31 show scattering coefficient profiles from the balloon-borne nephelometer
for both ascending (U) and descending (D) traverses of the tethersonde taken throughout the day.
Typical clean air values are less than a 0.01 km"1 whereas late in the day values approached 0.06
km"1.  (The apparent height displacement of the elevated peak between the two profiles of Figure
24 is judged to be a response-time effect  as the instrument was reeled up or down during the
runs). A complete listing of all radiosonde and TN- (tethered nephelometer) sonde flights made
during the intensive measurement period  is listed in Table 11.
                                          62

-------
              CHAMIZAL TN-SONDE, #1U 0708M 08 DEC 90
       LU
        - 150-
             00  0.02  0.04  0.06  0.08  0.10  0.12  0.14  0.16  0.18  0.20
                           SCATTERING COEFF, kmM
              CHAMIZAL TN-SONDE, #1D 0732M 08 DEC 90
            '6   0.02   0.04  0.06  0.08   O1  QA2  0.14  0.16  0.18  0.2
                           SCATTERING COEFF., kmM
Figure 24 Scattering coefficient profiles during balloon-nephelometer ascent (upper) and
descent (lower) during 0708-0732 hours on December 8 at the Chamizal site.
                                   63

-------
             CHAMIZAL TN-SONDE, #2U 0753M 08 DEC 90
            0   0.02  0.04  0.06 0.08   0.1  0.12  0.14  0.16 0.18   0.2
                           SCATTERING COEFF., kmA1
             CHAMIZAL TN-SONDE, #2D 0824M 08 DEC 90
          300
            0   0.02  0.04  0.06  0.08   0.1  0.12  0.14  0.16  0.18   0.2
                        SCATTERING COEFFICIENT, kmM
Figure 25 Scattering coefficient profiles during 0753-0824 hours at the Chamizal site.

                                  64

-------
              CHAMIZALTN-SONDE, #3U 0850M 08 DEC 90
          300
          250"	<—i
        CD
        LU
        - 150-
             0   0.02  0.04 0.06  0.08   0.1  0.12  0.14  0.16  0.18  0.2
                         SCATTERING COEFFICIENT, kmM
              CHAMIZALTN-SONDE, #3D 0920M 08 DEC 90
          300-
          250-
          200-
        - 150-
       LLI
       X
           50-
          100-	
                                                   ...	
             0  0.02  0.04  0.06  0.08  0.1   0.12  0.14  0.16 0.18   0.2
                         SCATTERING COEFFICIENT, kmM
Figure 26 Scattering coefficient profiles during 0805-0920 hours on December 8 at the
Chamizal site.
                                   65

-------
              CHAMIZAL TN-SONDE, #4U 1045M 08 DEC 90
          300-
          250-
          200-
        - 150-
       lit
       X
           50-
          100-	
                        ..	....
^	
             0   0.02  0.04  0.06  0.08  0.1   0.12  0.14  0.16  0.18  0.2
                         SCATTERING COEFFICIENT, kmM
              CHAMIZAL TN-SONDE, #4D 111OM 08 DEC 90
          300-
          250-
          200-
        - 150-
        LII
           50-
          100-
                 0.02  0.04  0.06  0.08  0.1   0.12  0.14  0.16  0.18  0.2
                         SCATTERING COEFFICIENT. kmM
Figure 27  Scattering coefficient profiles during 1045-1110 hours on December 8 at the
Chamizal site.
                                   66

-------
             CHAMIZALTN-SONDE, #5U 1235M 08 DEC 90
         300
         250-
         200- •
       UJ
       x
          50-



           0
          100-	
^	
                                     L ......J
                                     L ......J
                                               	
                0.02  0.04  0.06  0.08   0.1  0.12  0.14  0.16  0.18  0.2

                        SCATTERING COEFFICIENT, kmM
             CHAMIZALTN-SONDE, #5D 1305M 08 DEC 90
         300
          250-
          200-
       -  150-
       g
       01
       x
          100-
          50-
            0   0.02  0.04  0.06  0.08   0.1  0.12  0.14  0.16  0.18  0.2

                        SCATTERING COEFFICIENT, kmM
Figure 28 Scattering coefficient profiles during 1235-1305 hours on December 8 at the

Chamizal site.
                                  67

-------
              CHAMIZALTN-SONDE,#6U, 1408M, NEPHEL
       UJ
        _- 150-
            0.1
0.14
 0.18       0.22
SCATT. COEF., kmA-1
0.26
          300-
          250H
       <:  200H
       8
          100-
          50-
            0.1
                  CHAMIZALTN-SONDE, #6D, 1440M
0.14      0.18       0.22
        SCATT. COEFF., kmM
                    0.26
          0.3
Figure 29 Scattering coefficient profiles during 1408-1440 hours on December 8 at the
Chamizal site.
                                 68

-------
         300-
         250-
       - 150"
       UJ
                  CHAMIZALTN-SONDE,#7U, 1615M
                    0.04
  0.08       0.12
SCATT. COEFF., kmM
0.16
0.2
           I.25
                  CHAMIZALTN-SONDE, #7D, 1642M
       0.35
 SCATT. COEF.. .kmA-1
          0.45
Figure 30 Scattering coefficient profiles during 1615-1642 hours on December 8 at the
Chamizal site.
                                 69

-------
                  CHAMIZALTN-SONDE, #8U, 1718M
JW

-------
             Table 11
Tethersonde and Radiosonde Soundings
at the Chamlzal Site on Dec 8-10, 1990
Date
Dec 08
"
"
"
"
"
M
M
"
"
"
"
"
M
N
"
H
"
Dec 09
"
H
"
M
"
N
H
II
"
H
"
M
"
Dec 10
"
*
"
"
"
"
n
"
"
"
*
Ascent/Descent
No.
1U
10
2U
2D
3D
3D
4U
4D
5U
5D
6U
6D
7U
7D
8U
8D
Radiosonde
Radiosonde
1U
1D
2U
2D
3U
3D
4U
4D
5U
5D
6U
6D
Radiosonde
Radiosonde
1U
1D
2U
2D
3U
3D
4U
4D
5U
5D
Radiosonde
Radiosonde
Start Time
706
733
753
824
850
920
1045
1110
1235
1305
1408
1440
1615
1642
1718
1726
926
1259
701
728
803
830
855
920
958
1024
1018
1138
1210
1240
812
1200
705
730
811
850
920
951
1057
1120
1159
1223
809
1159
Stop Time
727
749
820
842
914
936
1104
1130
1257
1330
1430
1455
1634
1707
1723
1732
1050
1400
725
748
828
851
916
935
1019
1045
1133
1159
1235
1254
844
1246
727
749
843
843
945
1023
1117
1145
1223
1247
1001
1326
Max. Altitude
m
265

261

268

262

264

264

263

49

13500
9477
255

254

254

265

258

255

5358
7789
264

263

263

273

261

21680
15277
                 71

-------
Various minor problems crept into the operations to prevent 100 percent data recovery (sonde
battery  failure,  computer  software  hangups,  keyboard  malfunction in cold, subfreezing
environment), but a rough qualitative  estimate is  that better than 95 % data retrieval was
achieved.

The elevated scattering layer at about 200 m msl in Figures 24 through 26 is somewhat of a
puzzlement as to possible origin.  It is too high to have come off the northeast heights of El Paso
as drainage. Actually, the layer at 70 m agl may have come from this source since, as mentioned
above, there were several hours of light northerly wind in the morning before this profile was
measured. The 200 m layer is 64 m below the actual ASARCO stack top, and plume stabilization
is probably about 10 m higher than stack top. However, in the 6.5 km straight line distance from
the stack to Chamizal it is possible for some settlement to have occurred, or some downward flow
motion could have occurred in response to the diverging easterly wind flow caused  by terrain
features below. A look at the wind direction for 1330  m msl from the radiosonde released at 0926
shows it to be from 030°.

The variability in aerosol scattering coefficient from  profile to profile is striking. Profile 7D is of
questionable validity as the tethersonde battery failed shortly after starting descent. Prominent
features to be noted are: disappearance of the 200 m agl layer by 0850: cleansing of the  entire
profile by 0920 as a result of increasing mixed layer  (supported by following profiles); buildup of
values after 1300 (note scale shifts); and very large values near ground level by sunset. (Balloon
flying above 50 m after sunset was disallowed, hence the truncated #8 profile).  High aerosol
scattering values (0.45 km"1) were also seen on the nephelometer located in the adjacent shelter.

5.4.2 Diagnostic Wind Field Model Runs

5.4.2.1  Diagnostic Wind Model Description

EPA-6 requested that Sandia look at the Diagnostic Wind Model (DWM), a component part of the
larger Urban Airshed Model (UAM) used for oxidant modeling as  part of the SIP generation
process.  This submodel, prepared by SAI over many years, is a diagnostic,  as opposed to a
prognostic, formulation which calculates wind flow over complex terrain using a network of
surface  and upper air wind  observations.  Prognostic models  generally contain a  more
comprehensive suite of physics formulations which enable a realistic forecast of flow over complex
terrain.  The latter generally requires fewer input  observations than diagnostic approaches.
However, the complexities involved in a prognostic model usually mean an enhanced cost of
development and operation  over a diagnostic approach.

For application to the El Paso situation a 1:100,000  topographic map was studied to best locate
a computational domain that would include all areas of interest, observation points and influential
terrain features.  A 30 x 30 km square was marked out which best appeared to fill these criteria
and that fell on even metric grid blocks of the UTM (Universal Transverse Mercater) grid system
for the area. The grid domain is shown in Figure 1. A 40 x 40 grid of 750 m interval was chosen
with 14 layers in the vertical as being an acceptable resolution for faithfully representing the
terrain which did not overwhelm the computer memory and computational resources. Vertical
grid spacing varied from 25 m near the ground to 200 m at the top (1500 m agl). A digital tape
file of terrain data for the El Paso region and some associated processing software was received
from EPA contractor Lockheed Engineering & Sciences Co., Las Vegas, NV, at the request of
Region 6.
                                          72

-------
The model is formulated in terrain-following coordinates which allows computation of wind vectors
at constant heights above ground using conservation of mass (only) principles.  Input data
required are: grid description, gridded terrain heights and type, domain-mean winds, domain-scale
stability (dT/dz), and actual surface and upper air wind observations.  The model generates
gridded component (westerly, U, and southerly, V) winds in a two step procedure. A domain mean
wind is adjusted for kinematic effects of terrain forcing, thermodynamically generated slope flows
and blocking effects producing a spatially-varying, mass-consistent, gridded field for each specified
vertical layer. The second step  incorporates the addition of actual wind observations to the step
one domain averaged field. A specified radius of influence for observations is set by the user, such
that the domain mean is used outside the influence radius where observations are lacking.
Details of the procedures used in the code are given in published code documentation [EPA, 1990].


Processing  of surface and upper air data are handled by separate preprocessing routines which
require information about number, identifier and location of stations; starting and ending time;
hourly direction and speed for surface data; and observation time and height for upper air data.
Values are  interpolated spatially and temporally to provide inputs for each model level for each
hour of simulation. The resulting processed data are then written to separate files which are read
incrementally by the wind model. Another input file is prepared which contains the many control
options and parameters needed (over 40 are listed in the manual).  It  is clear that the best
combination for these parameters would take many experimental runs to optimize.

Output from the DWM is a series of tables of component wind values for each level and grid point.
These tabulated values were input to a post-processor package on a PC (Surfer®, Golden Software,
Boulder, CO) which produced graphic vectors at each grid point showing direction and speed class
for various pre-selected hours.  Results for  one of the selected stagnation days are shown later.


5.4.2.2 Selection of DWM test  day

Of the three days comprising the intensive measurement period (8th, 9th and 10th), the 8th was
selected as the most interesting single day for DWM performance evaluation. This judgement was
based on completeness of data records and elevated nature of ground monitoring data of both
particulate material and gas monitors.

5.4.2.3 DWM Model results

The required input files for the  DWM were prepared from the wind data then available (excluded
were data from Sunland Park, Executive Center, CAMS- 6, -12 and -30 that are now available and
will be included at a later date). Reformatting of the various data records into that required was
done by hand since three different formats were involved. The three yet unused groups mentioned
above are even different still. The DWM code had been installed on a VAX 8700, and later moved
to a VAX 3800 (Micro VAX) for reasons of economy. A typical run took approximately 45 minutes
on the 786, and roughly twice that on the Micro; however, per run charges were avoided on this
smaller computer.


A rigorous evaluation of the model's performance is beyond the scope of this report, however some
general observations about its  performance can be made. The output from the code is a series
of tables displaying U and V wind components at each grid cell for each height layer. These files
were downloaded to a desktop  computer, reformatted and run through the SURFER® program

                                          73

-------
mentioned above to produce graphical output.  Figures 32-37 show vector wind categories for
selected heights above terrain and hours of the day for each grid cell.  In addition, elevation
contours are shown at 30 m intervals.  In general, the vectors behave  In a logically intuitive
fashion: winds tend to flow around the major obstructions, speeds increase with height, increase
toward mid-afternoon, flow up slopes when heated, and down when cooled. Only the few right
angle turns violate the minimum amazement principle.  On the other hand, they occur generally
in very light wind areas.  No doubt some of this can be reduced by adjustment of some of the
many input parameters such as influence distance for observed winds, or the convergence criteria
for the divergence minimization procedure, etc. Another Interesting feature shown Is the very light
wind regime over the main downtown area of the dual metro  area centers. This may just be a
manifestation of the pooling effect of collecting, cool drainage flow.  A  complete collection of
computed results for the selected test day Is given in Appendix A.
                                          74

-------
  3524 r
  3518
c.
o
  3512
  3506
  3500
                                                       El Paso - Cd Juarez
                                                           1990 Dec 8
                                                            0400 MST
                                                            12.5m AGL
                                                                         ontour interval = 30m
                                                                              wind speed key

                                                                                » 0—0.9 m/s
                                                                                » 1-1.9 m/s
                                                                                t 2-2.9 m/s
      350
356
                                362
                             UTM km East
368
                                                           374
   Figure 32 Early morning (0400 hours) ground-level (13m agl) wind fields predicted by the
   Diagnostic Wind Field model on December 8.
                                              75

-------
3524
3518
3512
3506
3500 V
    350
El Paso —  Cd Juarez
    1990  Dec 8
     0400 MST
    1000.0m  AGL
                                                                       :ontour Interval - 30m
    wind speed  key

      »0—0.9 m/s
      H-1.9 m/3
      12-2.9 m/s
      13-3.9 m/9
      14-4.9 m/s
            m/s
                                                         374
                           UTM
 Figure 33 Early morning (0400 hours) upper-level (1000 m agl) wind field predicted by the
 Diagnostic Wind Field model for December 8.
                                             76

-------
  3524 \-
  3518
o
z
  3512
  3506
  3500
                                             0 Paso — Cd Juarez
                                                 1990 Dec 8
                                                  1200 MST
                                                  12.5m AGL
                                                                         (contour interval = 30m
                                                                              wind speed key

                                                                                « 0—0.9 m/s
                                                                                t 1-1.9 m/s
                                                                                » 2-2.9 m/s
      350
                   356
   362
UTM km East
                                              368
                                                            374
  Figure 34   Midday (1200 hours) ground-level  (13 m agl) wind field  predicted by the
  Diagnostic Wind Field model for December 8.
                                              77

-------
  3524
  3518
 o
z
  3512
  3506
  3500
El Paso — Cd Juarez
    1990 Dec 6
     1200 MST
    1000.0m AGL
                                                                          :ontour interval = 30m
                                                                              wind  speed  key

                                                                                 • 0-0.9 m/s
                                                                                 0-1.9 m/s
                                                                                 t 2-2.9 m/s
      350
                   356
                                 362
                              UTM km  East
                                              368
                                                            374
Figure 35   Midday (1200 hours)  upper-level (1000  m agl) wind field predicted by the
Diagnostic Wind Field model for December 8.
                                            78

-------
   3524 r
   3518
 .c
 o
   3512
   3506
   3500
                                                        El Paso — Cd Juarez
                                                            1990 Dec 8
                                                             1600 MST
                                                            12.5m AGL
                                                      contour interval  = 30m

                                                           wind speed key

                                                             > 0-0.9 m/s
                                                             '1-1.9 m/s
                                                             t 2-2.9 m/s
       350
356
   362
UTM km East
                                                368
                                                             374
Figure 36  Afternoon (1600 hours)  ground-level (13 m agl) wind field predicted by the
Diagnostic Wind Field model for December 8.
                                             79

-------
   3524
   3518
   3512
   3506
   3500
 El Paso  — Cd Juarez
     1990 Dec 8
      1600 MST
     1000.0m AGL
:ontour interval = 30m

     wind speed key
       • 0-0.9 m/s
       » 1-1.9 m/3
       f 2-2.9 m/s
       t 3-3.9 m/3
      350
                    356
                                 362
                              UTM km East
                                               368
                                                            374
Figure 37  Afternoon (1600  hours) upper-level (1000 m agl) wind field predicted by the
Diagnostic Wind Field model  for December 8.
                                            80

-------
5.4.3 Summary of FPM, CPM and PM10 Measurements

5.4.3.1  PM10 composition and correlation analysis

Aerosol mass concentration measurements are summarized for all sites in Table 12. The highest
aerosol concentrations in all categories were observed in Cd. Juarez at the Advance Transformer
site. A close examination of the data reveals that the highest measurements in both the US and
Mexico were encountered in the evening hours. Based on pooled measurements from all sites, the
mass fraction of PMi0 that was attributed to FPM was about 0.33 for El Paso sites and 0.43 for
the two Juarez sites.

Graphs showing 12-hour average FPM and CPM in a stacked bar format are given for the five sites
equipped with dichotomous samplers in Figures 38 through 42.  In general, the CPM  levels
constitute the majority of the total mass and, as noted earlier, the highest levels are observed in
the evening periods. Data from all five sites reveal the same patterns of pollutant concentration
over the two week study interval.  Particulate mass concentration levels were observed to rise to
a maximum during a three-day stagnation period that occurred from December 7 to 10. Levels
dropped considerably as the surface winds increased on the morning of December 11. Another
increase was observed on the  evening of December 15 followed by several days of high surface
winds which again decreased aerosol mass concentrations in the area. Another brief increase of
pollutants was  observed on December  19 and 20 just prior to study completion.  The highest
levels were observed at the Advance Transformer site followed by the Sun  Metro site.  Sites
showing intermediate particulate mass concentration levels were the Tecnologico site in Juarez
and the CAMS-6 site in El Paso.  Somewhat surprisingly, the Chamizal site showed the lowest
overall levels despite its location on the valley bottom astride the US-Mexico border.

                                       Table 12
                 Summary of 12-hr FPM, CPM and PM10 Measurements
Parameter
FPM All Sites
FPM El Paso
FPM Cd. Juarez
CPM All Sites
CPM El Paso
CPM Cd. Juarez
PM10 All Sites
PM10 El Paso
PM10 Cd. Juarez
N
183
111
72
183
111
72
185
111
74
Mean
ugrrr3
32
23
45
52
45
63
83
68
106
Minimum
ugm'3
0
0
0
0
0
0
0
0
0
Maximum
Mgm*
229
155
229
392
318
392
621
473
621
              Note: FPM = Particles <2.5 micrometer; CPM = Particles >2.5 and
              <10 micrometer; PMIO = FPM + CPM
                                          81

-------
           700-
                  3  4  5  6  7  B  9 10 11 12  13 14 15 16 17 18  19 20 21
                                Day of December 1990
Figure 38 A stacked barplot showing FPM and CPM for all sampling periods at the Advance
Transformer site.
                  3  456  7  8  9 10 11 12  13 14 15 16 17 18  19 20 21
                                Day of December 1990
Figure 39  A stacked barplot showing FPM and CPM for all sampling periods at the CAMS6
site [note scale change).

                                        82

-------
           16O
                  3  4 5  6  7  8 9  10 11 12 13 14 15 16 17 18 19 20 21
                                Day of December 1990
Figure 4O  A stacked bar plot showing FPM and CPM for all sampling periods at the
Chamizal site.
                  3  4  5  6  7  8  9 10 11 12  13 14  15 16  17 18  19 20 21
                                Day of December 1990
Figure 41  A stacked barplot showing FPM and CPM for all sampling periods at the Sun
Metro site [note scale change].

                                        83

-------
                 250-1
                        345
'8' "9  10 11 12 13 14 15  16 17  18 19 20 21
 Day of December 1990
  Figure 42   A stacked barplot showing FPM and CPM for all sampling  periods at the
  Tecnologico site.

The results of a correlation analysis among FPM, CPM and PM10 are given in Figures 43-45 for 12-
hour average data pooled from all five of the special study sites. In general, the plots show that
both FPM and CPM correlate well with PM,0. This correlation is in large part influenced by the
so-called meteorological factor.  This factor is related to the atmospheric mixing volume which,
in turn, is influenced by the vertical mixing height and local wind speeds encountered during any
particular time period.  When atmospheric vertical mixing is limited and wind speeds are low, all
pollutants, regardless  of their source will show concentration increases in the mixed layer.
Conversely, when the mixed layer is deep and wind speeds are high, the same pollutants sources
will mix into a larger volume resulting in lower ambient concentrations.

The correlation analysis of FPM with PM10 from all sites produced a correlation coefficient of 0.91
and a multiplier of about 2.4. That is to say, the PM10 mass concentration was a factor of about
2.4 that of FPM measured at the same site during the same 12-hour interval.  The correlation of
CPM with PMi0 produced an even  higher correlation coefficient of 0.97 with a CPM  to PM10
multiplier of about 1.4.  Correlation analysis reveals that at least during the winter season, FPM
is a reasonably good  predictor of PM10 levels in the region.  Furthermore,  these preliminary
analyses results suggest that combustion sources, a broad source category for FPM, account for
30 to 40% of the measured PM10.  The remaining 60-70% of the PM10 mass originates from CPM
which is derived primarily from crustal sources. CPM, to a large extent, originates from particles
produced by mechanical processes such as wind blown soil or re-suspended  soil particles from
vehicular activity.
                                           84

-------
      700
     -100
                         FPMvsPMIO ALL SITES
                     PM10 = 9.1976 + 2.3675 *   FPM
                        Correlation: r = .90945
                        ?	    II
20
60     100    140
      FPM, ug/mA3
                                              180
                                  220
          260
                                      Regression
                                      95% confid.
Figure 43  A scatterplot of FPM and PM-10 for all sampling sites and periods.

                        CPMvs  PM10 ALL SITES
                     PM10 = 8.0115+ 1.4496*   CPM
                         Correlation: r = .97116
   eo
   <
   o>
   3
     -100
         -50
50
150       250
CPM, ug/mA3
350
                                    450
                                                   Regression
                                                   95% confid.
Figure 44 A scatterplot of CPM and PM-10 for all sampling sites and periods.

                                     85

-------
                             FPMvsCPM  ALL SITES
                         CPM = 9.1976 + 1.3675 *   FPM
                             Correlation: r = .78409
        450
                   20
60     100    140     180     220
       FPM, ug/mA3
260
Regression
95% confid.
  Figure 45 A scatterplot of FPM and CPM for all sampling sites and periods.
5.4.3.2 Additional TNRCC Observations on Aerosol Mass Measurements

TNRCC analysts found that night period concentrations were typically higher and had wider
ranges than day concentrations [Dattner, 1993J.  Two stagnation episodes marked by elevated
pollution levels were observed during the study interval. One of brief duration occurred on Dec
4 followed by a second extended period from December 7 to 10.  Not surprisingly, meteorological
conditions measured during these stagnation periods revealed low wind speeds and conditions
conducive to the formation of nighttime temperature inversions.  Synoptic scale weather maps,
corresponding to the same December 7-10 time frame, revealed a high pressure ridge to the north
of the El Paso-Cd. Juarez area, clear skies and low wind speeds aloft. The highest PMIO levels
during the study interval occurred on the evening of December 10  when vertical mixing in the
atmosphere was suppressed by a temperature inversion.

The authors of the TNRCC report discuss the presence of outliers in plots of fine and coarse mass
concentrations from various measurement sites and suggest that further investigation into the
observed outliers may yield information about sources and sites.

The results of a correlation analysis between sites is also discussed in the  TNRCC report.  The
observation is made that the  lowest correlation  is noted for CPM inter-site comparisons.  A
suggestion Is offered that CPM is more localized and less likely to be transported and mixed over
the entire airshed by virtue of its relatively short atmospheric residence time. FPM, on the other
hand, shows the highest correlation presumably a result of its longer residence time in the mixed
layer.
                                         86

-------
The  12-hour dichotomous sampler measurements were combined Into 24-hour averages to
estimate the number of PMi0 NAAQS exceedences likely to have occurred during the study. The
results are summarized in Table 13. showing exceedences at all sites except Chamizal.  The
Advance Transformer site tops the list with eight exceedences.

                                      Table  13
              Estimated Number of Exceedences of the US PM10 Standard
            Based on a Combination of 12-hour Dichotomous Sampler Results
Site
ADV - Advanced Transformer
SUN - Sun Metro
CAM - CAMS6
TEC - Tecnologico
CHA - Chamizal Park
No. of Violations
8
5
1
1
0
Min, Max (pg/m3)
179,453
161,319
196
156

       from [Dattner, 1993]

5.4.3.3  Spatial Distribution of PM10

TNRCC analysts merged data from sites where 24-hour PM10 samples were collected with the five
special sites equipped with dichotomous samplers and  12-hour sampling intervals in order to
conduct a spatial analysis. Data from a total of fifteen sites were used with a data interpolation
and plotting package to produce contour plots for 24-hour sampling periods on December 7,  13
and 19. Figures 46 - 48 illustrate the results of this particular analysis. A weighted average was
used to estimate a midnight to midnight 24-hour average from the dichot samplers which were
started and stopped at SAM and 5PM. The contour plots, in general reveal a gradient of increasing
PM10 as one moves in a southerly direction toward the Cd. Juarez sites. The contour plots look
similar to those produced during the 1989 saturation PM10 study discussed in an earlier section
of this report (Figure 9). The Advance Transformer site appears to be a so-called "hot spot" with
measured levels of aerosol mass usually well in excess of levels observed at the other sites.
TNRCC researchers speculate that local brick kilns produced thick black smoke which may have
significantly impacted the samplers at this site. A moderate particulate concentrations gradient
from high to low is also observed in the  eastern portion of the basin between Cd. Juarez and
eastern El Paso.
                                          87

-------
                   ':<.>K':-x-?V!***f/v*""*:*'- f-£*.*+'^if•&•*••••'•<•

Figure 46 A contour plot of PM-10 levels Interpolated from measurements at 15 sites on
December 7 [from Dattner,  1993].
                                          88

-------
        December 13
Figure 47 A contour plot showing PM-10 levels Interpolated from measurements at 15
sites on December 13 [from Dattner, 1993].
                                 89

-------

       December 19
Figure 48 A contour plot showing PM-10 levels Interpolated from measurements at 15
sites on December 19 [from Dattner, 1993).
                                 90

-------
5.4.3.4  Conclusions from Particulate Mass Concentration Analysis

The  following conclusions are drawn from  both TNRCC and  SNL  analyses of the mass
concentration data collected during the short-term study.

       Considerable variation was observed in the PM10, CPM and FPM measurements made at
       the various monitoring sites in the study. Concentration levels were clearly the highest
       at the Advance Transformer site in southwestern Cd. Juarez and may have been heavily
       influenced  by local combustion  sources such as brick kilns.   The lowest mass
       measurements were observed at the Chamizal site which was located more or less at the
       geographic center of the valley.

       By combining the 12-hour mass concentration measurements, exceedences of the US PM10
       standard level were estimated to have occurred at all sites except the Chamizal site during
       the study interval.

       Night period measurements of partlculate mass were typically higher than day period
       measurements at all sites, presumably a result of diminished vertical mixing of the
       atmosphere during the evening and early morning hours.  Considerable altitude effects
       were observed as well, with low-lying sites showing much higher levels than adjacent sites
       several hundred feet higher.

       On average, the FPM was 30 to 40% of the PM10.  Both FPM and CPM correlated with PM10
       reasonably well, suggesting that either could be a reliable predictor of PM10.

       Spatial analysis of the particle concentration data suggest a concentration gradient highest
       in Cd. Juarez and lowest in El Paso; however, the relatively small number of sampling
       sites in Cd. Juarez limits the significance of this observation.

5.4.4 Elemental Species Measurements

A receptor modeling approach for air pollution source identification hinges upon a detailed
understanding of the composition of the particulate sample.  A determination of the elemental
composition of the sample is one of several aerosol speciation techniques employed in this study.
The conventional method for the determination of aerosol elemental composition is through the
use of X-ray fluorescence analysis. In this method, the particle-loaded filter is placed in an X-ray
spectrometer and bombarded with X-rays. This produces characteristic electronic transitions in
the sample elements.  The type and intensity of electronic transition is then directly related to a
specific element along with its mass loading on the filter. The methodology is non-destructive and
relatively cost-effective  since  a single  analysis  lasting  several minutes gives  quantitative
measurements for more than 30 elements.

Funding limitations required the selection of a subset of all filters to be analyzed by X-ray
fluorescence. Fifty fine filters and 22 coarse filters were selected for analysis. Filters from heavily
polluted periods were selected over those from periods with lower particulate mass concentrations.
A summary of fine and coarse elemental concentrations from two sites. Advance Transformer and
CAMS6 are shown in Tables 14 and 15. The reader is referred to the TNRCC report for a complete
description of the elemental analysis [Dattner,  1993]. Significant levels are noted for the crustal
elements like calcium, silicon and iron.  Chlorine and sulfur are also observed at high levels.
Lower, but nonetheless, measurable levels are observed for lead, bromine, arsenic and cadmium.
                                           91

-------
               Table 14
Summary FPM Elemental Analysis Data for
 a Cd. Juarez and El Paso Monitoring Site
Element
FPM
Al
Si
S
Cl
K
Ca
Ti
V
Cr
Mn
Fe
Cu
Zn
As
Se
Br
Sr
Pb
Cd
Sn
Sb
Ba
Advance Transformer
Average
ng/m3
102,100
172
1,101
1,514
4,219
608
2,414
23
10
3
20
316
81
160
51
11
100
8
470
10
10
40
4
Std. Dev.
ng/m3
67,480
57
546
792
3,506
399
1,459
14
8
3
17
170
55
87
44
12
53
4
232
5
8
37
7
CAMS6
Average
ng/m3
39,770
71
611
1,266
941
249
862
14
10
7
19
285
114
180
86
11
33
5
236
8
5
15
16
Std. Dev.
ng/m3
31,434
61
378
627
1,348
198
705
9
6
4
7
145
64
105
67
7
31
2
125
4
5
12
8
                  92

-------
               Table IS
Summary CPM Elemental Analysis Data for
 a Cd. Juarez and El Paso Monitoring Site
Element
CPM
Al
Si
S
Cl
K
Ca
Ti
V
Cr
Mn
Fe
Cu
Zn
As
Se
Br
Rb
Sr
Pb
Zr
Cd
Sn
Sb
Ba
Advance Transformer
Average
ng/m3
28,820
498
2,339
154
130
342
3,594
49
4
4
15
607
12
40
2
1
6
2
22
28
2
2
0
0
16
Std. Dev.
ng/m3
22,169
501
1,781
115
171
263
3,106
40
4
5
11
423
6
69
2
0
5
1
15
21
2
1
1
1
17
CAMS6
Average
ng/m3
54,950
923
4,166
279
117
607
7,007
99
9
9
57
1,521
137
71
18
3
5
4
27
59
4
4
2
4
46
Std. Dev.
ng/m3
21 ,849
292
1,524
41
42
228
3,393
47
2
1
4
404
61
8
1
1
3
1
11
2
0
1
2
1
7
                  93

-------
5.4.5 Aerosol and Semi-volatile Carbon Measurements

5.4.5.1  Aerosol Carbon

TNRCC  analysts extended considerable effort in the analysis of the large collection of fine and
coarse fraction quartz filter pairs collected at the five specially equipped study sites. A total of 150
filter pairs were analyzed by conventional staged thermal combustion methods.  Using this
method, the aerosol carbon is further broken down into such categories as organic and elemental
carbon which can provide additional information on source contributions.  A number of general
conclusions from the TNRCC aerosol carbon analysis are summarized below.

    The mass fractions of FPM  and  CPM that were attributable to particulate carbon are
    summarized in Table  16 for the five study sites. About 50 to 60% of the FPM is attributable
    to volatile and non-volatile carbon aerosol. The aerosol carbon mass fraction for the CPM was
    on the order of 15 to 20% and 10 to 20% of PM10.

                                       Table 16
               Summary Aerosol Carbon Mass Fractions by Sampling Site
Site
ADV
TEC
CAM
CHM
SUN
FPM Carbon Mass
Fraction
0.4810.16
0.40 ±0.1 2
0.52 ±0.13
0.50 ±0.13
0.47 ±0.16
CPM Carbon Mass
Fraction
0.16 + 0.19
0.31 ±0.16
0.13 ±0.06
0.10 ±0.05
0.09 ± 0.04
PM10 Carbon Mass
Fraction
0.27 ±0.14
0.36 ±0.1 2
0.28 ± 0.09
0.23 ± 0.08
0.20 ± 0.07
No.
Samples
29
25
27
21
34
             from [Dattner, 1993]

    As normally encountered in urban aerosol, volatile aerosol carbon concentrations were higher
    than the non-volatile concentrations at all sites, usually by a factor of two or greater.
    Differences between day and night concentrations were similar to those observed for overalf
    mass concentrations—levels were highest during the evening periods.

    The TNRCC conducted a "math spectral analysis" on the aerosol carbon data and reported
    that 12 distinct spectra were derived from the data set.  No implications in terms of source
    attribution are reported however since the analysis methodology and its physical significance
    are still under development.

5.4.5.2  Semi-volatile Organic Aerosol

A separate sampler at CAMS6 was deployed to collect the semi-volatile component of the aerosol
so that the concentration of individual semi-volatile organic species could be determined. Analysis
by GC-MS gave a measure of total polycyclic aromatic hydrocarbons (PAH) as well as selected
species, such as benzo(a)pyrene, pyrene and others, some of which are either known or suspect
human carcinogens. A series of 12-hour samples were collected over the same time frame as the
coarse and fine particulate samples.  A graph showing total PAH as well as selected PAH species
                                          94

-------
is shown in Figure 49 showing concentration increases that correlate to increased stagnation
around December 7. Not surprisingly, increases in total PAH and individual species correlate with
the stagnation periods encountered during the study.  Increases during  these periods were
observed to be as  high as 10-fold above the background levels encountered during "clean air"
conditions.
                  700
                                          8      10      12
                                         Day of December 1990
                     14
         16
18
                       Total PAH
                       Naphthalene
Benz(a) Anthracene
Phenanthrene
Acenaphthene
  Figure 49  Total polycyclic aromatic hydrocarbon and selected species concentrations
  measured at the CAMS6 site during a portion of the study interval [from Dattner, 1993].

5.4.6  Other Measurements

5.4.6.1 Annular Denuder Measurements

TNRCC made a limited number of measurements with annular denuders2 at the CAMS6 site
during the  study  interval.  The measurements are deemed  relevant since they allow for the
speciation of inorganic acid gases and aerosols, such as sulfur dioxide, aerosol sulfate, nitric acid
and aerosol nitrate.  Measurements of these particular species allow inferences to be drawn
concerning  the extent to which secondary aerosol is a factor in the winter season.  Secondary
aerosol is particulate matter that is formed by reaction of various gaseous pollutants, such as
sulfur dioxide or nitric acid, with other atmospheric constituents. Knowing these species' relative
concentration levels can also give an indication as to the relative "age" of the aerosol in the region;
   2 Annular denuder sampling technology allows differentiation and quantitation of related gas and particle
species such as sulfur dioxide and aerosol sulfate from the same sampled volume of air.

                                           95

-------
information which may be further used to infer whether the gaseous sources such as sulfur
dioxide are local or whether they are regional and transported into the local area by wind flow.

Results from TNRCC analysis of particulate and gaseous sulfur and nitrogen species reveal that
the dominant sources in this category are local.  A ratio of gaseous to particulate sulfur of about
0.9 is reported which indicates a relatively "new" source of sulfur gas. Gaseous sulfur sources
transported into the region from distant sources would show a much lower ratio since much of
the gas would be converted to the particulate sulfur phase during the transport process. This
finding is corroborated with other sulfur aerosol measurements suggesting local  sources and is
consistent with the presence of the local ASARCO copper smelting operation from which gaseous
sulfur emissions are known to occur.

A lower gas to particle ratio  of about 0.7 was observed  for nitrogen species. TNRCC analysts
suggest that this ratio may reflect faster gas to particle reaction rates for the nitrogen species
when compared to the sulfur species or the fact that higher levels of particulate nitrogen are
released from primary sources.  The TNRCC analysts also point out that the gaseous nitrogen
emissions are primarily from ground level vehicular sources as compared to  gaseous sulfur
emissions that are released  from elevated smelter stacks.  Ground level emissions are thus
released into the mixed layer which results in longer residence time as compared to gaseous sulfur
emissions aloft which are only periodically mixed into the inversion layer during the late afternoon
when the mixed layer approaches the stack height.

Measurements of sulfur dioxide by annular denuder over an approximate two-week interval reveal
higher daytime levels than nighttime levels.  The largest daytime 12-hour average observed was
about 190 ug/m3.  The higher daytime values are taken to confirm relatively shallow mixed layer
that exists during the evening hours.  During these times the sulfur gas emissions occur above
the mixed layer and are not effectively transported down to ground level.

Smaller differences are observed between nighttime and daytime gaseous nitrogen species such
as nitric acid and nitrous acid. This observation suggests that sources of gaseous nitrogen species
are primarily vehicular in nature and by virtue of their ground location are always in the mixed
layer. Thus, day-night differences are minimized for the nitrogen species.

5.4.6.2 Electron microscopy analysis

The TNRCC collected fine and coarse aerosol fractions on a specialized  dlchotomous sampler
designed to optimize particle loadings on the filter for microscopy analysis. The intent of this effort
was to combine receptor modeling approaches with  information on the elemental composition of
discrete particles via electron microscopy in much  the same way as done in the earlier Energy
Technologies study discussed earlier in this report. Samples were collected at the CAMS6 station
and sent to the EPA-AREAL laboratory for analysis.  The filters were determined to be overloaded
with particles such that the analysis could not be completed.

5.4.7 Temporal PM10, Nephelometer and CO data analysis

The TNRCC conducted an analysis of continuously  monitored PMi0, CO and nephelometer data
collected at the Chamizal site during the December 7-10 stagnation event which  revealed some
interesting patterns and relations among these pollutants. A plot of these three parameters is
given in Figure 50 showing a high degree of correlation among the three parameters. A number
of conclusions drawn by TNRCC analysts from this analysis are summarized below.
                                          96

-------
  Figure 50  A graph of aerosol light scattering. CO and PM-10 measurements from the
  continuous monitors at the Chamizal monitoring site [from Dattner, 19931.
   The high degree of correlation over reasonably short time scales (1-hour) suggests a common
   source type for these three measured parameters.

   The three parameters  show peaks during the morning and evening  commute periods
   suggesting that vehicular activity may be the predominate source of these parameters.

   Carbon monoxide is commonly associated with combustion sources and with vehicular sources
   in particular. Biomass burning also produces CO however the degree to which this source is
   a major contributor in the El Paso-Cd. Juarez region is unclear from these data.

   Late evening CO peaks are not as well correlated with nephelometer and PMj0 peaks suggesting
   the presence of biomass combustion sources not in the vehicular category.

   The high degree of PM10 correlation with CO suggests the presence of two possible PM10 sources
   related to vehicular sources, namely: tailpipe emissions and entrained road dust associated
   with vehicular travel. One or both of these source categories are responsible for the observed
   PM10 levels.

A TNRCC comparison of nephelometer data from the Chamizal and the CAMS-6 sites over the
same measurement interval reveals an appreciable degree of correlation, however, not as high as
that observed between the three parameters at the Chamizal site discussed previously.  This
comparison is taken to suggest that pollutant measurements  at a particular site reflect, to a
significant extent, the presence of sources local to that particular site. In general, the results of
these analyses  illustrate the effectiveness of using time-resolved measurements to help clarify
source contributions in the area.
                                          97

-------
5.4.8 Receptor Modeling Analysis

5.4.8.1 Approach and Factor Analysis Method Description

One particular aspect of the short term winter-season study involved a preliminary evaluation of
receptor modeling techniques as applied to the data set produced during this two week intensive
study interval.  Receptor modeling approaches are viewed as one of the ways that predictions
concerning current and future pollutant levels can be made by the State of Texas in their
development of a PM10 State Implementation Plan for the El Paso-Cd. Juarez region. The other
traditional approach used in SIP analyses for various regulated pollutants is the combination of
emission inventories with air dispersion modeling. This approach relies heavily on the preparation
of a pollutant emission database which incorporates source estimates for all significant sources
in a particular region. These include area, point, vehicular or mobile sources and other categories
as well.  These emissions data are combined with  hourly measurements of meteorological
parameters made at selected locations in the area of interest, including such measurements as
wind direction, wind speed, atmospheric  mixing height  and atmospheric turbulence.  The
meteorological parameters are used to predict the transport and diffusion of pollutants from the
numerous sources and ultimately, to estimate pollutant concentrations at selected receptor sites
of interest. The discussion of the SAI report in Section 3 gives a brief review of some of the
dispersion models available for these types of analyses.

Receptor modeling techniques, on the other hand, differ from dispersion modeling techniques in
that a detailed pollutant data set is collected at a specific receptor site or sites in discrete sampling
Intervals. Analysis of the collected aerosol provides detailed composition of the sample including
the elemental and carbon species composition. The elemental composition of the sample is then
used to draw inferences about various local and regional sources and their relative degree of
contribution to measured mass categories such as FPM or PM10.  For example, the extent to which
sulfur containing aerosols are present In a sample may be directly related to the amount of sulfur
gas emitted from a local copper smelter; or, the extent to which silicon particles are observed in
a sample may be used  to estimate the extent to which soil sources such as wind-lofted dust or
dust re-suspended by vehicles from unpaved roadways are present In the sample.  One of the
most powerful receptor models available for source strength determination is the EPA Chemical
Mass Balance or CMB model. This model takes the elemental signatures of multiple sources and
by classical least squares fitting techniques determines the extent of each source contribution to
a measured sample. This approach has merit;  however, a limitation to successful application of
this model is a requirement for measurements of the elemental composition of the major sources
of Interest In the region—a requirement which can result in costly collection  and analysis of
numerous source samples. Alternatively, literature values of various source categories can be
used. In many cases, however, these may be of Insufficient detail or specificity for a particular
Industrial process or area.

A related receptor modeling technique employs the use of factor analysis techniques which allow
classification of the numerous variables in a data set from a receptor site or sites into groupings
of variables that may be logically related to actual source categories.  In general factor analysis
and its companion approach principal component analysis are statistical approaches used to: (1)
simplify a data set composed of many variables into a fewer number of variables or factors
(principal components analysis); or, (2) to discover underlying structure in a data set composed
of many variables (factor analysis).  In either case, the methods attempt to extract so-called factors
from the data set in such  a way that the factors account for as much of the variability in the
original data set as possible. Factor analysis has often been applied to the data sets obtained
from air sampling as a receptor modeling tool. Partlculate matter that has been analyzed for
elemental composition yields a data set that has many variables well suited for this approach.

                                          98

-------
Since many sources emit distinctive or unique elements, a factor analysis treatment of the data
can often help elucidate sources in a particular region.

Factor analysis is a desirable approach since it allows the data set to reveal source categories with
a minimum of analyst bias or intervention.  Correlations among variables are mathematically
determined with limited analyst intervention.  The CMB approach on the other hand requires the
analyst  to select appropriate source categories for least squares fitting routines and source
strength determination.  It is often desirable to use factor analysis as a first step in exploring
potential sources, with a follow-on analysis using the CMB model after significant sources have
been identified in the data set.

5.4.8.2  Factor Analysis Results

Factor analysis was carried out on the combined data set and subsets that were assembled from
all five special study sites by Sandia National Labs researchers using factor analysis routines in
the Statistica® (Statsoft® Inc. Tulsa, OK)  software package.   A detailed  description of the
underlying theory and approaches to factor analysis is  beyond the scope of this report, however
many references are available that describe this approach in more detail [Hopke, 1985; Stevens,
1986; Harman, 1967].

Factor Analysis on the FPM Data Set

Factor analysis was carried out on a raw and normalized FPM elements data set which included
the suite of elements analyzed by XRF along with the measured fine particle mass concentration
for each filter.  Earlier work has shown  that normalization of all variables in the data set to a
parameter such as FPM helps to reduce co-linearity of all variables caused by the meteorological
factor.   Variable co-linearity can weaken the ability of factor analysis to detect the underlying
structure in the data set since all variables tend to extract under a single meteorological factor.
In this case all 12-hour average fine fraction element concentrations were divided by the 12-hour
average FPM, and coarse fraction  element concentrations by the 12-hour average  CPM.  The
degree to which the extracted factors explain the variability encountered in the FPM data set is
given in Table 17. The results reveal only a moderately successful model with about 58 percent
of the variance explained by the five factors that were extracted.

                                       Table 17
                         Goodness of Fit Results for a 5-Factor
                        Solution to the Normalized FPM  Data Set
Factor
1
2
3
4
5
Total
Variance
%
24.5
16.8
7.2
5.3
4.0
Cumulative
Variance
%
24.5
41.4
48.5
53.8
57.8
                                           99

-------
The loadings of the elemental variables on the five factors after a varimax3 rotation are given In
Table 18. For clarity, those elements with loadings in excess of 0.5 are shaded. Several source
patterns can be identified in the data set.  The first factor shows high loadings for Cu, Zn, As, and
Pb. Clearly, this factor is indicative of a smelter source since all of these elements are emitted
during the copper ore smelting process.  Sulfur is also moderately loaded on this factor with a
loading of 0.47. It is interesting to note in this analysis that Pb is loaded onto the smelter factor
and not some other factor related to vehicular emissions.  A second factor shows high loadings
for Si, K, Ca, Ti, and Fe and is suggestive of crustal sources in the region.  A third factor shows
intermediate loadings for S, Ni and Sn. A fourth factor shows loadings for Br, and Cl and V, Cr,
Mn and Ge are loaded on a fifth factor. While the first two factors can be physically related to
sources within the region, the same cannot be conclusively said for the last three factors.  Day
and night subsets of the  data were similarly analyzed by factor  analysis with results showing
similar  characteristics to the combined day and night data set.
   3 Varimax rotation is one of the matrix manipulation techniques used within the factor analysis routine
to improve the resolution of the variables into groups of factors.

                                           100

-------
                   Table 18
Normalized Fine Particle Factor Analysis
Results
Element
SI
S
CL
K
CA
Tl
V
CR
MN
FE
Nl
CU
ZN
GA
GE
AS
SE
BR
RB
SR
PB
ZR
MO
AG
CD
SN
SB
1
BA
Expl.Var
Prp.Totl
FACTOR 1
0.02
0.47
-0.08
-0.02
0.01
-0.03
0.14
0.00
0.32
0.18
0.06
: 0,89
;- 0,31
0.03
0.35
0,82
0£l
0.08
0.31
-0.04
! $3$ ,
-0.11
0.50
s - w
! s 0,69
0.31
0.46
0.33
0.03
4.86
0.17
FACTOR 2
o.sa
-0.10
-0.03
0,71
.. ,0,9$
0.7$
0.10
0.32
Use
ft.QO
0.00
0.15
0.12
-0.05
-0.01
-0.03
0.01
-0.12
0.58
I 0,69
-0.07
0.40
0.15
0.01
0.03
0.05
-0.12
0.02
0.31
5.33
0.18
FACTOR 3
-0.01
O.S*
-0.20
0.14
-0.03
-0.07
-0.04
0.06
0.29
0.11
0.70
0.18
0.20
0.02
0.09
0.24
-0.13
0.33
-0.17
0.04
0.22
0.23
0.37
-0.18
0.29
0.72
0.39
0,60
03S
2.94
0.10
FACTOR 4
0.02
0.22
-0,63
-0.16
-0.03
0.00
0.06
0.11
-0.11
0.10
-0.14
0.02
0.07
0.20
0.15
0.08
-0.01
-OS
-0.08
0.05
-0.14
0.25
0.32
0.00
0.09
-0.10
-0.35
0.22
0.15
1.53
0.05
FACTOR 5
0.07
0.14
-0.07
-0.06
0.03
0.36
,0,54
< - 0,67
! 0-50
0.14
0.26
0.23
0.01
0.02
OJ5
0.22
0.31
-0.04
0.41
0.01
-0.08
0.23
0.12
0.16
-0.01
-0.01
-0.10
-0.25
-0.07
2.09
0.07
                     101

-------
Factor Analysis on the CPM Data Set

The combined coarse particle data set was analyzed in a similar manner as for the FPM data set.
The overall variance in the data set accounted for by the five factors extracted in this solution is
given in Table 19. The fit of the factor model to the data set is slightly better for the CPM data set
when compared to the FPM data set. The CPM model accounts for about 68% of the variability
in the CPM data set as compared to 58% for the FPM data set.

                                       Table 19
                         Goodness of Fit Results for a 5-Factor
                       Solution of the Normalized CPM Data Set
Factor
1
2
3
4
5
Total Variance
%
25.0
14.8
10.6
10.4
6.6
Cumulative
Variance
%
25.0
39.7
50.4
60.8
67.4
The loadings of the various elements on the 5 factors extracted from this data set are given in
Table 20.  The first factor shows high loadings for Ti, V, Cr, Ni, Ga, Zr. and Mo. Sources related
to this profile are not immediately obvious.  A second factor shows high loadings for Cu, Zn and
As and is  probably related to smelter emissions. The elements of Br, S and Cl are loaded on a
third factor and may be related to vehicular sources as evidenced by the presence of Br in this
elemental  profile.  Elements of Si, K, Ca, Ti and Fe are loaded on a fourth crustal factor, while a
fifth factor shows loadings for Sb, I and Ba and Cd for which a physical source is not obvious.
Sulfur is loaded at a lower (0.34) level in the CPM smelter factor than its loading of 0.51 in the
FPM smelter factor.  This  may be a result of the fact that sulfur is emitted from the process
primarily in the gaseous form followed by the formation of particulate sulfate species in the FPM
size category by atmospheric oxidation.
                                         102

-------
                    Table 20
Normalized Coarse Particle Factor Analysis
Results
Element
SI
S
CL
K
CA
Tl
V
CR
MN
FE
Nl
CU
ZN
GA
GE
AS
SE
BR
RB
SR
PB
ZR
MO
AG
CD
SN
SB
1
BA
Expl.Var
Prp.Totl
FACTOR 1
0.13
0.14
-0.06
0.18
-0.02
* ' ,$.71
, " &$4
\ 0,75
' , 0.69
0.16
' 0.86 ..
0.07
0.06
^0.81
! &5
-------
5.4.8.3  Source Strength Estimation by Tracer Elements

Use of the CMB7 receptor model on the data set collected during this short term study was beyond
the scope of analysis in this study. Estimates of the degree to which selected sources contributed
to the overall FPM and CPM were obtained, however, by using a number of unique tracer elements
for several sources known to exist in the El Paso-Cd. Juarez region. The degree to which soil
contributes to measured FPM and CPM was estimated by taking literature-derived values for the
Si mass fraction of soil in both the fine  and coarse size fractions [Watson, 1979]. Mass fractions
of Si in FPM and  CPM were 259 and 321 milligrams per gram of soil, respectively, resulting in
multipliers of 3.86 and 3.12 to scale from Si concentrations to crustally-derived FPM and CPM
respectively.  The following calculation was carried out for each 12-hour sampling period:
                                    '"" •    FPM    iww                          (1)
where %FPMS01I is an estimate of the percent mass fraction of total FPM that is derived from the
crustal source, SiFPM is the Si concentration measured in the 12-hour dichot filter sample, M is
the multiplier used to scale Si to crustal FPM, and FPM is the measured fine particle mass
concentration in the 12-hour sample.  In this calculation we assume that airborne Si is unique
to the crustal source.  In other words, all Si detected on the filter has crustal origins.  Similar
calculations were carried out for the coarse fraction measurements using a different multiplier and
CPM instead of FPM.  The results for both fine and coarse fractions are plotted as frequency
histograms in Figures 51 and 52 for all 12-hour periods. For the fine fraction, estimated crustal
source strength ranges from about 4 to 18 percent with most occurrences falling in the range of
4 to 8 percent. The average value was 6 ± 3 percent.  Estimates for soil contribution to the coarse
aerosol fraction range from a low of about 5 percent to a high of 60 percent. The median value
for the coarse size fraction falls at about 30 percent.  These estimates reveal that which is already
known to be true for most urban  air masses—crustal sources such as windblown dust or
entrained dust from vehicular activity constitute a significant fraction of the CPM size category.


Similar calculations were carried out using As as a unique tracer for smelter sources in the region.
Literature values for the elemental concentrations  measured in smelter plumes were used to
estimate the mass fraction of As in both fine and coarse sizes at about 4 percent [Small, 1981].
This value corresponds to a multiplier of 25 for use in the calculation for smelter source strength
analogous to that shown in equation 1  for the crustal source. Figures 53 and 54 give frequency
distributions for the smelter source strength in both the fine and coarse size fraction for all 12-
hour sampling periods.  The smelter contribution to FPM is the highest of the two sizes with a
median value of about 2 percent. Smelter contribution to the coarse fraction is an approximately
eight-fold less with a median value of about 0.25 percent.


Source strength estimates for vehicular and biomass burning were carried out using Pb and K as
elemental tracers. Lead, although not uniquely identified with vehicular sources, was nonetheless
used to  determine an  upper bound estimate for vehicular source contribution to both fine  and
coarse aerosol.  For this calculation we assumed that all airborne Pb originated from vehicles
                                          104

-------
burning leaded fuel. We further determined through conversations with local air quality officials
that about 95% of all fuel consumed in El Paso is unleaded gasoline and that about 80% of fuel
consumed in Cd. Juarez is unleaded [Reynoso, 1994]. We used the total VMT results for El Paso
and Cd. Juarez, posted in Table 7, to calculate a weighted average of fuel type consumption for
both cities at 91% unleaded and 9% leaded fuel. From the literature, we used 0.2 as an estimate
of the ratio of fine particle Pb to total FPM for vehicular sources burning leaded fuel [Pierson, 1976
and Watson, 19791. Thus increasing the measured Pb value by a factor of 5 provides a measure
of vehicular FPM, assuming that all vehicles are burning leaded fuel.  Since only 9% of the fuel
consumed is leaded fuel in the region, we incorporate another multiplier of 11.1 to account for
those vehicles burning unleaded fuel, if we conservatively assume that vehicles burning unleaded
fuel emit the same amount of fine particle  aerosol as those burning leaded fuel.  The following
expression then yields a worst case estimate of FPM from vehicles:
                                      _       *5*11.1 ,»,ftA
                               vehlcle --    - * 1 0 0
A similar expression was used for estimates of vehicular source contribution to the coarse particle
fraction. In this case a multiplier of 6.76 was used instead of 5 to account for lower Pb emissions
in the coarse particle fraction [Watson, 1979]. Figures 55 and 56 give estimates of the percent
contribution of the vehicular source category to overall FPM and CPM for all sites and sampling
periods, using the assumptions noted above.  We are careful to point out that these estimates
represent an upper bound since other sources of aerosol Pb exist within the airshed. The overall
average contribution of the vehicular source to FPM was 31 ± 18%, and 10 ± 6% to CPM. The
range of contribution for both size categories is quite variable ranging from 10 to 55 percent for
the fine fraction and 3 to 20 percent for the coarse fraction.

A similar calculation was carried out using K as a tracer for biomass burning. We corrected the
total K measured on the filter for that contribution from crustal sources by using the ratio of soil
K to soil Si. We used values of 0.040 for the fine fraction and 0.069 for the coarse fraction as
reported in the literature [Watson, 1979]. Crustal K can be estimated taking the product of Si
concentration and the  K^/St,,,,, ratio if one assumes that  all Si is crustally derived.  The
expression used to determine FPM derived from vehicular sources for a particular sampling period
follows:
                                                         *100                    (3)
where FPMb,omass is the fine aerosol concentration attributable to biomass burning in a particular
12-hour sample, K is the total potassium concentration measured on the filter with K^,, and Si^u
the soil composition of the two elements as noted above. The multiplier in this case is 100, which
approximates the range of literature values for the mass fraction of K in fine and coarse biomass
smoke [Hopke, 1985]. Figures 57 and  58 show estimates of the percent contribution of the
biomass source category to total FPM and CPM respectively. The distribution is quite narrow for
the fine fraction with well over half of the values falling between 50 and 60%. The mean
contribution of this source category to the fine fraction aerosol was 53 ± 14%. A lower mean
contribution of 31 ± 8% was noted for the coarse aerosol size fraction.
                                          105

-------
                   01  2  3  4  5  6  7  8  9  10  11  12  13  14  15
                               CONTRIBUTION TO FPM, %
Figure 51   Estimates of crustal source contribution to FPM  for all sites and sampling
periods.
          12-1
       ffi
                 0 5 101520253035404550556065707580859095 100105110115120
                             CONTRIBUTION TO CPM, %
Figure 52  Estimates  of crustal source contribution to CPM for all sites and sampling
periods.
                                         106

-------
                    0   2   4   6   8   10   12  14  16   18  20  22   24
                               CONTRIBUTION TO FPM, %
Figure 53 Estimates of smelter source contribution to FPM for all sites and all sampling
periods.
        ffi
        Q-

        O
                  0.00 025 0.50  0.75 1.00 125  1.50 1.75 2.00  2.25 2.50 2.75  3.00
                              CONTRIBUTION TO CPM, %
Figure 54  Estimates of smelter source contribution to CPM for all sites and sampling
periods.
                                         107

-------
        UJ
        a:
                   0   5   10  15  20  25   30   35   -40   45   50   55
                              CONTRIBUTION TO FPM, %
Figure 55  Estimates of vehicular source contribution to FPM for all sites and sampling
periods.
           121
                  0 1  23  456  7  8 9 10 11 12 13 14 15 16 17 18 19 20
                             CONTRIBUTION TO CPM, %
Figure 56 Estimates of the vehicular source contribution to CPM for all sampling sites and
periods.
                                        108

-------
                   0   10  20  30  40  50  60  70  80  90  100  110 120
                             CONTRIBUTION TO FPM, %
Figure 57  Estimates of biomass combustion source contribution to FPM for all sampling
sites and periods.
                  04   8   12  16  20  24  28  32  36  40  44  48 52
                             CONTRIBUTION TO CPM, %
Figure 58  Estimates of biomass combustion source contribution to CPM for all sampling
sites and periods.

                                        109

-------
Overall summary of source contributions to FPM, CPM and PM10
A summary of average source strength estimates for both the fine and coarse aerosol fractions
using tracer elements is given in Table 21. The overall composition of PM10 during this winter
season study  was 39% FPM and 61%CPM as given earlier in Table  12.  Using  this PM10
composition, a weigh ted average of the source strength estimates given in Table 13 was calculated
for the PM10 size category and is shown in Figure 59.   Our estimates reveal that biomass
combustion and crustal sources together account for nearly 80 percent of PM10 during the winter
season.  The vehicular category is at about 20% and about one-fourth of the biomass and crustal
categories combined. The smelter category is a minor contributor at a level of about 2%.  It is
arguable that the crustal category is to a large extent influenced by the vehicular category since
the use of vehicles on both paved  and unpaved roadways will result in the entrainment of soil
particles into the air. Furthermore, it is readily apparent that many miles of unpaved roadways
exist in the residential sections of Cd. Juarez.

                                       Table 21
                         Summary Source Strength Estimates
                         for Pine and Coarse Aerosol Fractions
Source Category
Vehicular
Biomass Combustion
Smelter
Crustal
Fine Particle
Fraction (%)
31 ±18
53 ±14
4 + 5
6±3
Coarse Particle
Fraction (%)
10 + 6
31+8
<1 + 1
57 + 22
               Crustal (38.2%)—,
                    Smelter (2.2%)
                                                   Vehicular (18.7%)
Biomass (40.8%)
  Figure 59 Estimates of source category contribution to total PM-10 during the winter season
  El Paso-Cd. Juarez.
                                         110

-------
These results reveal several important features of winter season air quality in the El Paso del Norte
airshed. First, our estimates reveal that biomass combustion sources are a significant contributor
to PM10.  The use of wood products for residential heating and brick kiln fuel appears to  be a
significant source of PM10 during the winter season. The extent to which wood combustion is used
for residential home heating in El Paso has not been reported in any of the reports reviewed as
a part of this effort. Wood or other biomass combustion in Cd.  Juarez is recognized both in the
brick making industry and for residential home heating since many Cd. Juarez residences lack
natural gas or electric central heating systems. Further investigations Into the extent of biomass
combustion would be warranted to confirm the estimates derived here.  Secondly, the crustal
source is a major contributor to the  PM10 category and is probably very closely tied to vehicle use
in Cd. Juarez. In the studies reviewed in this work, no comparison is made between El Paso and
Cd. Juarez with respect to unpaved  roadways. However, in general unpaved roadways are much
more prevalent in Cd. Juarez than in El Paso. The results from this study would indicate that
programs to curtail biomass burning and to pave roadways would do much to control the PMj0
levels encountered during the winter season in this particular airshed.

Further refinement of source strength by receptor modeling is  best carried out with use of the
CMB model. As noted earlier, however, effective use of this model is not without its own set of
problems. The analyst Is faced with the selection of appropriate source profiles for use in the
statistical fitting routines by either the judicious  use of library source  profiles or through the
collection of source profiles in the El Paso-Cd. Juarez area at considerable effort and expense.
Short term studies such as the one  discussed here prove useful by providing initial estimates of
source category strength from which follow-on studies can be thoughtfully planned and executed.
                                          Ill

-------
                    6.O Summary Conclusions and a Forward Look

The most important findings from the studies reviewed in Sections 3 and 4 of this report are
summarized along with the major findings from the short term winter PM10 study presented in
Section 5.  A brief review of some recent technological developments relevant to air quality issues
in the El Paso del Norte region is also given.

6.1 Summary Findings

   Industrial stationary sources do not contribute significantly to airborne partlculate
   matter in the El Paso del Norte region -  Studies in the 1980's focusing on total suspended
   particulate matter and more recent studies dealing with PMJO both reveal that large industrial
   stationary sources such as copper smelters in the region are minor contributors to winter
   season airborne particulate matter.

   Winter season PM10 levels are highest in the Cd. Juarez-El Paso downtown areas and in
   general show a concentration gradient increasing toward Mexico - A number of studies,
   among them the EPA-EMSL airborne lidar study, EPA-6 saturation PM10 study, and the short
   term winter PM10 study, all reveal higher PM10 levels as one moves toward Cd. Juarez. Further
   indirect evidence for this concentration gradient comes by way of the TNRCC PM10 SIP analysis
   which reveals that no exceedences of PMj0 are predicted in El Paso if only El Paso particulate
   matter sources are taken into account. In actual studies, however, PM10 concentration levels
   in excess of air quality standards are encountered on both sides of the border.

   Emissions from the average vehicle in Cd. Juarez are about three-fold  higher than from
   the average vehicle in El Paso - The University of Denver remote sensing studies for tailpipe
   CO and hydrocarbons show that while the amount of pollutants emitted from a high-polluting
   car is the same in both Cd. Juarez and El Paso, there are more of these high-polluting cars
   operating in Cd. Juarez than in El Paso. Other studies undertaken to characterize the age of
   the vehicular fleet in Cd. Juarez reveal that the average age of the fleet is older in Cd. Juarez
   than in El Paso.  This observation is consistent with the measured higher  tailpipe emissions
   in Cd. Juarez.

   Vehicle miles traveled in El Paso are about three-fold higher than in Cd. Juarez - Results
   from a series of studies by the Texas Transportation Institute on vehicle usage in Cd. Juarez
   reveal about 3.4 million VMT in Cd. Juarez as compared to 9.9 million VMT in El Paso.  Per
   capita mileage in El Paso is about six-fold higher than in Cd. Juarez revealing very different
   vehicle usage patterns in the two cities. Should Cd. Juarez residents adopt US driving habits,
   vehicular emissions could significantly increase in the Paso del Norte region.

   Winter stagnation events and complex terrain significantly limit pollutant dilution within
   the region - Meteorological data taken during the PM10 short term study confirm the presence
   of very shallow vertical mixing heights in the evening and early morning hours during winter
   season. Wind flow at various meteorological monitoring locations throughout the El Paso del
   Norte region reveal local terrain influence during winter stagnation periods.  Wind fields
   predicted by the Diagnostic Wind Field Model show general agreement with observed winds.
   A rigorous comparison was not carried out as a part of this study however.  In some test cases
   abnormal discontinuities in  the  predicted wind fields were  observed suggesting  that
   optimization of the model may be required to obtain representative results.
                                         113

-------
   Winter season PM10 in the Paso del Norte region reaches levels in excess of the National
   Ambient Air Quality Standard - Analyses completed by the TNRCC as a part of the short term
   winter PM10 study revealed numerous exceedences of the 24-hour PM10 standard at both Cd.
   Juarez and El Paso sites during the winter season study period.  In some cases, PM10 levels
   three-fold higher than the air quality standard were observed.

   Aerosol carbon is a major constituent of fine, coarse and PM10 aerosol in the region -
   Carbon species  measurements  of collected particulate  matter  reveal  aerosol  carbon
   composition much like that observed in other urban areas. High levels of both elemental and
   volatile carbon are observed and originate from vehicular, biomass combustion and other local
   combustion sources.  Limited measurements of individual toxic aerosol species such as
   benzo(a)pyrene show that, in general, these vary proportionately with overall aerosol carbon
   levels. The relatively high levels of elemental or soot carbon have important implications for
   visibility in the region as well since elemental carbon is an important contributor to visibility
   reduction.

   Crustal and biomass combustion sources together constitute nearly 80 percent of the
   winter season total PM10 measured in the Paso del Norte region - Preliminary estimates of
   source category strength using tracer elements reveal that these two sources are the major
   contributors to winter season PM10. Average contribution of the crustal source to PM10 is about
   40%.  The crustal source is understood to be closely linked to vehicular sources which are
   estimated to contribute no more than 20% to total PM10. Vehicle traffic on many unpaved
   roadways, prevalent in Cd. Juarez residential areas, results in the suspension of both fine and
   coarse fraction crustal material, thus linking the vehicular and crustal source categories. The
   use of soil-corrected potassium as a tracer for biomass combustion reveals about 40% average
   contribution of this source to PMIO as well.


6.2 New Technologies

Several recent studies have employed new technologies to further explore air quality issues in the
El Paso del Norte region. The EPA is interested in employing remote sensing  and other new
techniques to  the El-Paso-Cd.  Juarez air pollution problem.  Thorough  understanding of
instrument performance and data validation of particular techniques are needed prior to wide-
spread use of such techniques for compliance monitoring however. These new technologies are
briefly described below.


6.2.1 Lidar Systems

Los Alamos National Laboratories Mini-Lidar

   Los Alamos National Laboratory has on two occasions field tested a new "mini-lidar" system
   to map aerosol spatial distribution and to estimate wind fields in the El Paso-Cd. Juarez
   region.  This is a relatively small, transportable, ground-based lidar system  that measures
   elastic back scatter from airborne particles along the laser beam path. Like any lidar system,
   the travel time of the light pulse out and it reflection back to the detector is used as a measure
   of the distance that the aerosol is from the lidar. The intensity of return is used to gauge the
   amount of aerosol as a function of range from the lidar. The lidar is fixed to a gimble mount
   along with receiving optics so that the beam can be scanned over a wide path thus allowing
   aerosol density mapping.
                                         114

-------
   Further applications are being explored with this system that involve the estimation of wind
   fields by tracking aerosol features in the lidar return signal over time. At the present time this
   wind field determination not straightforward, being computationally intensive and requiring
   considerable analyst intervention in the data processing steps.  The research goal however, is
   to develop a system that can map out wind fields, giving both wind direction and speed over
   time In a gridded format using the  scanning lidar system. The wind field data are desirable
   for use in pollutant dispersion models that are more accurate than many of the gaussian
   dispersion models currently in use  by the regulatory community.

Sandia National Laboratories UV-DIAL

   Sandia has developed a state-of-the-art ultraviolet differential absorption lidar (UV DIAL) for
   the remote sensing of chemical species in air.  Preliminary testing of a prototype system has
   recently been completed. DIAL is a variation of conventional absorption spectroscopy that has
   been adapted to provide the range resolution required by many remote sensing applications.
   In DIAL two laser pulses are transmitted sequentially—one tuned "on" and the other tuned "off"
   an absorption band of the chemical species of interest.  The back scatter of the light pulses
   from aerosol and gas molecules in the air provides a recording of signal intensity versus time
   for both pulses. By ratioing the two signals, a range-resolved measure of concentration of the
   selected molecular species can be obtained. By scanning the pointing direction of the optical
   system with a moveable mirror system,  a full  three-dimensional spatial map  of species
   concentration  can be recorded in a short time.

   The Sandia UV DIAL system is well suited for many different applications in environmental
   monitoring.  The system Is housed in a semi-trailer that can be parked in a single location.
   The concentration of selected species such as ozone or nitrogen oxides can be mapped out over
   a large (in some cases several kilometer) radius by rotating and tilting the system's scanning
   mirror. The broad tunability of the laser system provides an unusual degree of flexibility for
   a UV DIAL system, making it straightforward to tailor sensitivity, range, spectral resolution and
   temporal resolution for urban pollution applications. For example, understanding the spatial
   distribution of ozone (both horizontally and vertically) throughout the day may yield important
   information about localized pollutant sources contributing to the ozone problem.

Sandia National Laboratories UV-Fluorescence Lidar

   Sandia researchers have also developed and recently tested a UV-Fluorescent lidar system for
   the remote detection of selected organic and inorganic pollutants. The system uses a laser
   pulse to cause electronic transitions in molecules along the beam path.  Electromagnetic
   radiation is emitted as the molecules decay back to their ground state which is, in turn,
   detected with an  optical configuration at the  lidar.   Frequency-agile lasers enable the
   wavelength of the outgoing light pulse to be varied which yields chemical specificity for the
   system, allowing in many cases individual pollutant species such as benzene or toluene to be
   detected.  The fluorescence signal,  detected over a wide frequency band, is processed using
   advanced multivariate algorithms to detect both the amount and range of pollutants from the
   lidar.  Laser beam direction is accomplished with a scanning mirror  such that pollutant
   mapping can be accomplished.  Detectability is expected to be in the mid to high parts per
   billion for many of the organic  species commonly encountered in urban air.  While tills
   particular device may not be suitable for general ambient air monitoring because of relatively
   low sensitivity, it may prove useful for the remote detection of stack emissions  or fugitive
   emissions from stationary sources.
                                           115

-------
6.2.2 Model Development

Various regulatory and research level pollutant chemistry and transport models were briefly
described in the review of the SAI report in Section 4.7.  In this section we highlight several new
approaches to urban air pollution modeling that offer promise for improved pollutant forecasting
and air pollution control  program  evaluation.   A number of prognostic models are under
continuing development and improvement. These models accept a minimum of meteorological
data and from them calculate wind fields over a wide area much like the Diagnostic Wind Field
Module of the Urban Airshed Model  discussed earlier.  A complementary effort is underway to
carefully model the emissions from the wide range of vehicles and vehicle usage patterns that
occur on time scales as short as minutes in a typical urban area. Such parameters as vehicle
count over the roadway system of a city, the operating mode of the vehicle (cold-start, accelerating,
decelerating, idling at a traffic light, etc.) are used by the model to estimate pollutant emissions
over an entire urban area.  Coupling this information with wind field data produced by a
prognostic model makes possible the prediction of air quality in an urban area over the duration
of a commute period in a wide range of weather and traffic flow patterns. Los Alamos National
Laboratories researchers are developing such a system called TRANSIMS. The system is currently
in its early  stages of development and requires significant computational effort  and further
understanding of source emission characteristics. Improvements in computer distributed and
parallel processing techniques are expected to speed the computational process to the point where
urban simulation routines such as TRANSIMS may become a useful tool for city and regional
planners at some point in the future.

6.2.3 Complementary Technologies

Many of the new technologies  under development  offer important advances in ambient air
monitoring issues currently faced by such urban areas as Cd. Juarez-El Paso.  It is unlikely that
any of the technologies noted above will completely replace conventional point monitoring methods
currently in use. These technologies undoubtedly will prove useful however in special studies over
relatively short time frames such as reviewed earlier in this report. A combination of these new
innovations with conventional monitoring and modeling techniques will yield better understanding
and predictive capabilities for air quality trends in international urban areas such as El Paso-Cd.
Juarez.
                                         116

-------
                                   7.O References
Ballard, H. N., M. Izquierdo, C. Tate, G. Romero, M. Paganlnl, C. Becerra, J. De La Reza, 1990.
"Study of Upper Air Winds in El Paso During an Air Stagnation Event with Aerosol Measurements
as Related to Visibility"; Contract 89/90-064; Electrical Engineering Department. University of
Texas at El Paso, El Paso TX.

Bath, C. R, 1983. 'Vehicles and Air Pollution in El Paso-Cd. Juarez", C. R Bath (ed), Center for
Inter-American and Border Studies, University of Texas at El Paso, El Paso, TX 79968.

CSU, 1990. "1990 Juarez, Mexico Tampering Survey Statistics". Colorado State University. Fort
Collins,  CO.

Douglas. S. G., and R C. Kessler, 1988.  "User's Guide to the Diagnostic Wind Model. Version
1.0"; Systems Applications International, San Rafael, CA.

EEA, 1993. "Aversion of the MOBILES Emission Factors Model with User-Revisable Input Data
Files"; Energy and Environmental Analysis, Inc., 1655 North Fort Myer Dr Suite 600. Arlington.
VA 22209.

Energy Technology Consultants, 1983A  "Identification of the Sources of Total  Suspended
Particulates and Particulate Lead in the El Paso Area by Quanitative Microscopic Analysis, Volume
I"; Murrysville. PA.

Energy Technology Consultants, 1983B.  "Identification of the Sources of Total  Suspended
Particulates and Particulate Lead in the El Paso Area by Quanitative Microscopic Analysis, Volume
II"; Murrysville, PA.

Environmental Research and Technology  Inc., 1983.  "Assessment of Primary Inhalable and
Thoracic Particulate Matter Emissions in the Houston-Galveston and El Paso Areas of Texas":
Report No. P-B467-070A, Atlanta. GA.

EPA, 1985. "Compilation of Air Pollutant Emission Factors", Document No. AP-42, Fourth Edition.
US EPA Office of Air Quality and Standards,  Research Triangle Park, NC.

EPA, 1986.  "Guideline on Air Quality Models (Revised)"; Report No. EPA-450/2-78-027R- US
Environmental Protection Agency, Washington D.C.

EPA, 1990.  "Users Guide for the Urban  Airshed Model"; Report No.  EPA-450-4-90-007. US
Environmental Protection Agency, Washington D.C.

Friedt, P. M., B. Wolff,  L.  Bohren, 1994.  "Mexico Automotive Technician Training Needs
Assessment Summary"; National Center for Vehicle Emissions Control and Safety, Colorado State
University, Fort Collins, CO 80523.

Gray. H.  A,  C. A.  Emery, M.  P Ligocki,  1991.  "Modeling Program  for the  PM10 State
Implementation Plan Development for the El Paso/Ciudad Juarez Airshed"; Report No. SYSAPP-
91/109, Systems Applications International,  San Rafael, CA.

Harman, H. H., 1967. Modem Factor Analysis. Chicago, University of Chicago Press.
                                         117

-------
Hemy, R and B. M. Kim, 1988.  "A Factor Analysis Receptor Model with  Explicit Physical
Constraints"; in Receptor Models in Air Resource Management: J. G. Watson, ed. Air and Waste
Management Association, pp 214-225.

Hopke, P. K., 1985. Receptor Modeling In Environmental Chemistry, New York, John Wiley and
Sons Inc.

Johnson, A. T., J. Soto Jr., J. B. Ward, 1994.  "Successful Modernization of an Ancient Industry:
The Brickmakers of Cd.  Juarez, Mexico", presented at New Mexico  Conference on the
Environment, April 25. 1994, New Mexico Environment Department, Sante Fe, NM.

Kemp, M.; "El Paso/Juarez Saturation PM,0 Study"; Report No. 906-R-92-001; US EPA Region 6,
Dallas, TX.

Lighty, J. S. and D. W. Pershing, 1993.  "Control of Pollutant Emissions from Waste Burning";
Department of Chemical Engineering, University of Utah, Salt Lake City, UT.

Marlia. C..B. M.KimandR Wu, 1990. "Annual PM10 Dispersion Model Application to the South
Coast Air Basin", Draft Technical Report V-D, Draft Air Quality Management Plan, 1991 Revision,
South Coast Air Quality Management District.

McDonald, C., M. Paganini. G. Romero, C. Becerra, 1990.  "Study of Upper Air Winds in El Paso
During an Air Stagnation Event (Visibility and Extinction Measurements)"; Contract 89/90 - 064;
Electrical Engineering Department, University of Texas at El Paso, El Paso TX.

McElroy, J., 1990. "1989 El Paso-Juarez Particulate Pollutant Transport Study", Environmental
Monitoring Systems Laboratory, US Environmental Protection Agency, Las Vegas NV.

McNider. R T.  and R A. Pilke,  1984. "Numerical simulation of slope and mountain flows", J..
Climate and Appl. Meteorol. 23:1441-1453.

Oregon  Department  of Environmental  Quality,  1990.   "Modeling Documentation,  State
Implementation Plan for the Medford-Ashland Oregon Particulate Matter Non-attainment Area";
Salem, OR

Pierson, W. R and W. W. Brachaczek, 1976.  "Particulate Matter Associated with Vehicles on the
Road," SAE Automotive Engineering Congress and Exhibition, Detroit,  Michigan.

Radian Corporation, 1983A, "A Study of Ambient Air Quality and Emission Source Contributions
to Ambient Air Pollution for El Paso County, Volume I, Analyses and Conclusions, Final Report";
Report No. DCN 83-144-772-02, Austin, TX.

Radian Corporation, 1983B. "A Study of Ambient Air Quality and Emission Source Contributions
to Ambient Air Pollution for El  Paso County, Volume II, Evaluation of the Quality of Available
Monitoring Data and Appendices"; Report No. DCN 83-144-771-05, Austin, TX.

Reynoso, J. 1994. Telephone Conversation, October 1994.

Rincon, C. A.. 1994. 'Vehicle Emissions Inspection and Maintenance Program in Ciudad Juarez.
Chihuahua. Mexico and El Paso, Texas"; submitted to Peter Emerson. Environmental Defense
Fund by the El Paso del Norte Task Force.
                                        118

-------
Small. M., M. S. Gennanl, A. M. Small, W. H. Zoller, and J. L, Moyers. 1981. "Airborne Plume
Study of Emissions for the Processing of Copper Ores In Southeastern Arizona";  Environmental
Science and Technology 15:293-298.

Stedman, D. H., 1993A. Interim Report on Vehicle Measurements In El Paso, Texas and Juarez
Mexico", University of Denver, Denver. CO, 80208-0179.

Stevens, J., 1986. Applied Multivariate Statistics for the Social Sciences. Hillsdale, NJ, Erlbaum.

Stedman. D. H. and J. H. Lesko,  1993B.  "Second Interim Report on the [El Paso-Cd. Juarez]
Remote Sensing Study", University of Denver, Denver, CO 80208.

TNRCC, 1991. "Revisions to the State Implementation Plan for Inhalable Particulate Matter",
Texas Natural Resource Conservation Commission, Austin, TX 78753.

TTI,  1994.  "Ciudad Juarez MOBILES Data Collection", Texas Transportation Institute, College
Station. TX

Watson, J. G., 1979. "Chemical Element Balance Receptor Model Methodology for Assessing the
Source of Fine and Total Suspended Particulate Matter In Portland, Oregon"; Ph.D. Thesis, Oregon
Graduate Center, Beaverton, OR.

Yamada, T., C. Y. J. Kao and S. Bunker,  1989.  "Airflow and air quality simulations over the
western mountainous region with a four-dimensional data assimilation technique", Atmos. Envir.
23:539-554.

Yarbrough, J, 1994. Telephone conversation, August 18, 1994.
                                         119

-------
     Appendix A
Wind Fields Calculated
       by the
Diagnostic Wind Model

 for December 8, 199O

-------
                        Diagnostic Wind Model Graphical Output

The graphical output for all periods and altitude layers run in this particular investigation are
contained in this appendix. The specific times and altitudes are as follows:

       Date:  08-Dec-90

       Hours (Mountain Standard Time):  0000, 0400. 0800, 1200, 1600, 2000

       Altitude Layers (meters above ground level):  13, 38, 75, 350, 1000

Wind speed and direction are given as a vector at each model grid point. The direction of the
vector arrow indicates the direction toward which the wind is blowing. The length of the arrow
is a measure of the wind speed. A wind speed key to the right of each vector plot gives wind speed
ranges in units of meters per second. The local topography is also shown with a contour plot on
each of the vector plots with contour altitude markings in meters above mean sea level.

-------

-------

-------
iiii

-------
iilii

-------

-------
                             *^y  *  » ,  , ,  x  x x x  x x y* -y.  x Vx
                             ky »***»>-xxxx  ^^<  x y x  x/x x  X *  »  •
^ •  *jrf/ •  *  *  *  *  *  *  *  »^>i'fc^'»N»XXS
                                                 oi>| rim

-------
UJ>| Win

-------
iu>) Win

-------

-------

-------
nin

-------

-------
tu>| run

-------
% jX*-^>«->«  »»»»»»«»•
«T  »  »^y  »»«•»•»••

-------
liflii
 O> O> OS Ol
d ^   '

-------OCR error (C:\Conversion\JobRoot\00000CKF\tiff\200149CH.tif): Saving image to "C:\Conversion\JobRoot\00000CKF\tiff\200149CH.T$F.T$F" failed.

-------
 D  no


     o
T>   0>
O  Q
 i   e
 O

D.
        o   E
    _  o   in
    03  ^   i<
                                              o
                                              n
                                               o
                                              ^J
                                               c
                                                             _   _
                                                             o>  o>


                                                             ?  7
                                                       c

                                                      '*

-------

-------

-------
 D
 3
-3


•D

O
10  (_  O
 o  en  <
 o  2   _
Q        E
    o  o


i  §  I
                                                  1
                                                  0)
^     ^^^

 0)     O)   O  O)

 JJ     o   ^  e>i
 O.     ill
 Ifl      III
       O   »-  (N
"U      *   *-  *•
 c

'i

-------

-------

-------

-------

-------

-------

-------

-------

-------
                 o
  ro
\  o o
  o> o
O O5 
-------

-------
                                   TECHNICAL REPORT DATA
                           (Please read fnitrucrions on the reverie before completing}
 1. REPORT NO.
 . EPA-906-R-95-001
3. RECIPIENT'S ACCESSION NO.
 4. TITLE AND SUBTITLE
  Winter Season Air Pollution in El Paso-Ciudad Juarez
6. REPORT DATE
    March 1995
                                                          6. PERFORMING ORGANIZATION CODE
 7. AUTMORIS)
 - Wayne Einfeld
  Hugh W. Church
8. PERFORMING ORGANIZATION REPORT NO
   SAND95-0273
9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Sandia National Laboratories  -    ~. •- -
  Environmental- Characterization Department
  Albuquerque, NM  87185  •
10. PROGRAM ELEMENT NO.
11. CONTRACT /GRANT NO."
 DW89933419^01
 12. SPONSORING AGENCY NAME AND ADDRESS
  U.S.  EPA - Region 6
  1445  Ross Avenue
  Dallas, TX  75202
 3. TYPE OF REPORT AND PERIOD COVER
 Interagency Agreement 1989-
 . SPONSORING AGENCY CODE
 IS. SUPPLEMENTARY NOTES

 Project Officer;  James w. Yarbrbugh
 16. ABSTRACT
      This report  discusses  results from the  Winter 1990  particulate
      matter  intensive  field study sponsored  by EPA in  El Paso-Ciudad
      Juarez  and adjoining parts of New Mexico.  The report also
      summarizes all other known,  substantive,  quantitative studies of
      winter  season air pollution  in this international airshed.
                              KEY WORDS AND DOCUMENT ANALYSIS
                 DESCRIPTORS
                                            b. IDENTIFIERS/OPEN ENDED TERMS
              COSATI Fieid/Cr
    Particulate matter
    Carbon monoxide
    Air pollution
    Meteorology
    U.S.-Mexico air pollution
 Unlimi ted
                                            19. SECURITY CLASS (TTtu Rtfortf
                                             Unclassified
                                           20. SECURITY CLASS iThapage/
                                             Unclassified
EPA f»,m 2220-1 (R... 4.77}   PUCVIOUI coi TION 11 O«»OU*TK
            21. NO. OF PAGES
                161
           23. PRICE

-------