United
Environmental Protection
Aggney
Municipal Environmental
Laborvtory
Cincinnati OH 45268
EPA-600/2-79-050^
August 197&
Research and Development
Maximum
Utilization of
Water Resources in a
Planned Community
Bacterial
Characteristics of
Storm waters in
Developing Rural
Areas
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development. U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4 Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution-sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-79-050f
August 1979
MAXIMUM UTILIZATION OF WATER RESOURCES
IN A PLANNED COMMUNITY
Bacterial Characteristics of Stormwaters
In Developing Rural Areas
by
Ernst M. Davis
The University of Texas at Houston
School of Public Health
Houston, Texas 77025
Grant No. R802433
Project Officers:
!
Richard Field
Anthony N. Tafuri
Storm and Combined Sewer Section
Wastewater Research Division
Municipal Environmental Research Laboratory (Cincinnati)
Edison, New Jersey 08817
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection
Agency, nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
-------
FOREWORD
The Environmental Protection Agency was created because of increasing public
and government concern about the dangers of pollution to the health and
welfare of the American people. Noxious air, foul water, and spoiled land
are tragic testimony to the deterioration of our natural environment. The
complexity of that environment and the interplay between its components re-
quire a concentrated and integrated attack on the problem.
Research and development is that necessary first step in problem solution
and it involves defining the problem, measuring its impact, and searching
for solutions. The Municipal Environmental Research Laboratory develops
new and improved technology and systems for the prevention, treatment, and
management of wastewater and solid and hazardous waste pollutant discharges
from municipal and community sources, for the preservation and treatment
for public drinking water supplies and to minimize the adverse economic,
social, health, and aesthetic effects of pollution. This publication is
one of the products of that research, a most vital communications link
between the researcher and the user community.
This project focuses on methods of maximizing the use of water resources in
a planned urban environment, while minimizing their degradation. Particular
attention is being directed towards determining the biological, chemical,
hydrological and physical characteristics of stormwater runoff and its
corresponding role in the urban water cycle.
Francis T. Mayo
Director
Municipal Environmental Research
Laboratory
111
-------
PREFACE
The overall goal of this research was to evaluate the
water resource plan for The Woodlands, Texas, and to make
recommendations, as necessary, to maximize its effective
utilization through alterations in design and management.
Any recommended alterations were to be critically evaluated
as to their compatibility with the natural environment.
Collection and utilization of stormwater runoff for
recreational and aesthetic purposes was a major feature of
the water resources plan at The Woodlands. Control of down-
stream flooding was also of great importance and so storage
reservoirs, in the form of recreational lakes and wet wea-
ther ponds, were created by the developers. Water quality
was a concern if the impoundments were to be aesthetically
appealing and/or suitable for recreation. Therefore, a ma-
jor sampling and analytical program was designed to monitor
water quality and quantity at different locations in the
developing area. The Storm Water Management Model (SWMM)
provided the focal point for combining the water quality
and quantity data into a predictive tool for design and
management purposes.
SWMM was originally developed for highly urbanized
areas and, therefore, was calibrated for this project in an
urban watershed (Hunting Bayou). Subsequently, SWMM was
modified to model runoff and water quality from natural
drainage areas, such as The Woodlands. Because of the lag
in the construction schedule at The Woodlands, the dense
urban areas were not completed during the project period.
Consequently, Hunting Bayou and other urban watersheds were
sampled to provide a basis for predicting pollutant loads
at The Woodlands in the fully developed state.
Water analyses included many traditional physical,
chemical and biological parameters used in water quality
surveys. Pathogenic bacteria were also enumerated since
the role of traditional bacterial indicators in stormwater
were conducted to assess the eutrophication potential that
would exist in the stormwater impoundments. The source,
transport and fate of chlorinated hydrocarbons in storm-
water runoff was also investigated.
IV
-------
-Several of the large Woodlands impoundments will re-
ceive reclaimed wastewater as the major input during dry
weather. Besides their use as a source of irrigation water,
the lakes will be used for non-contact recreation—primarily
fishing and boating. Because the reclaimed wastewater must
be disinfected, there was a concern about disinfectant toxi-
city to the aquatic life in the lakes. Consequently, com-
parative fish toxicity tests were conducted with ozone and
chlorine, the two alternatives available at the water re-
clamation plant.
Porous pavement was considered by the developers as a
method for reducing excessive runoff due to urbanization
and an experimental parking lot was constructed. Hydraulic
data was collected and used to develop a model compatible
with SWMM, to predict the effects of using porous pavement
in development. Water quality changes due to infiltration
through the paving were also determined.
Pacteriological content of low flow and storm water
runoff was determined for the purpose of comparing densi-
ties of indicator bacteria with densities of bacterial patho-
gens. Those, in turn, were compared with data developed
from secondary treated chlorinated and unchlorinated muni-
cipal wastewater. This approach related the hygienic quali-
ty of the water when compared to new bacteriological water
quality standards for contact and non-contact recreation
purposes.
Hopefully, the results of this project will contribute
in a positive way to the development of techniques to uti-
lize our urban water resources in a manner more compatible
with our cherished natural environment.
-------
ABSTRACT
An investigation of low flow and storm water runoff
bacteria content from rural and urban areas was conducted
over a two and a half year period. Data were obtained for
total coliform, fecal coliform, fecal streptococci, Salmo-
nella sp., Pseudomonas sp., and Staphylococcus sp. for com-
parison to densities in known polluted sources such as se-
condary treated chlorinated municipal wastewater. The use-
fullness of the currently employed indicator groups of bac-
teria was evaluated with respect to the accompanying densi-
ties of pathogens. The hygienic quality of water when com-
pared to new bacteriological water quality standards for
contact and noncontact recreation was considered. Settling
of stormwater suspended solids was closely associated with
bacterial reductions in the water column. The most useful
indicators of pathogen content in stormwater runoff were
fecal coliforms. Total coliforms and fecal streptococci
were poor indicators of pathogenic bacteria densities.
Chlorine and ozone doses for disinfection of stormwater con-
taining high (^200 mg/1) suspended solids may exceed 8 mg/1
and 32 mg/1, respectively. Regrowth of total coliforms oc-
curs following disinfection. Indicator group densities in
urban stormwater runoff can easily exceed rural runoff den-
sities with continual increases occurring throughout a
storm event. Fecal coliform densities exceeded 2,000/lOOml
in 13 of 24 monitored hydrographs and exceeded 200/100 ml
in 22 of those hydrographs. Fecal coliforms and fecal
streptococci yielded the highest correlations with the phy-
sical factors, flow, suspended solids, and turbidity.
This report was submitted in partial fulfillment of U.S.
Environmental Protection Agency Grant No. R802433, by the
University of Texas at Houston, School of Public Health under
the (partial) sponsorship of the U.S. Environmental Protection
Agency. This report covers a period from November, 1973 to
June, 1976, and work was completed as of May, 1976.
VI
-------
CONTENTS
Foreword iii
Preface iv
Abstract vi
Figures viii
Tables x
Acknowledgments xi
1. Introduction 1
2. Conclusions 11
3. Recommendations 16
4. Materials and Procedures 18
5. Results and Discussion 24
References 79
-------
FIGURES
Number Page
1 Principal surface waters in primary test
site 3
2 Locations of soil sampling sites and lakes
A and B water quality sampling locations
(1/11/75 survey) 6
3 Total Coliform vs fecal coliform and fecal
streptococci relationships during low flow
periods. Station P-20 32
4 Total coliform responses to chlorine; Lake
A (n=6) 37
5 Fecal coliform responses to chlorine; Lake
A (n=6) 38
6 Fecal streptococci responses to chlorine;
Lake A (n=6) 39
7 Staphylococci responses to chlorine; Lake A
(n=6) 40
8 FC densities during 4/7-10/75 storm event;
the Woodlands Station P-10 52
9 Fecal coliform densities during 4/7-10/75
storm event; the Woodlands Station P-30 . . 53
10 FS densities during 4/7-10/75 storm event;
the Woodlands Station P-10 54
11 Fecal streptococci densities during 4/7-10/75
storm event; the Woodlands Station P-30 . . 55
12 Staphylococci densities during 4/7-10/75
storm event; the Woodlands Station P-10 . . 56
Vlll
-------
FIGURES (Continued)
Number Page
13 Staphylococci relationships during 4/7-10/75
storm event; the Woodlands Station P-30 ... 57
14 FC/FS ratios during 4/7-10/75 storm event;
the Woodlands Station P-10 58
15 FC/FS relationships during 4/7-10/75 storm
event; the Woodlands Station P-30 59
16 Scalar approach to FC/FS patterns in
different land use areas 70
17 Fecal streptococcus vs fecal coliform
regression lines for five land use areas . . 71
18 Indicators and pathogen relationships; low-
flow and storm events. Station P-10, the
Woodlands 73
19 Indicators and pathogen relationships; low-
flow and storm events. Station P-30, the
Woodlands 74
20 Low-flow vs storm rainfall occurrences at
Stations P-10 and P-30 76
21 Comparison of Ipw-flow, storm, sewage, soils,
and stream sediment indicator and Pseudomonas
densities 77
22 Woodlands seasonal low-flow variations .... 78
IX
-------
TABLES
Number Page
1 Descriptions of Low-Flow Water Quality
Stations Established in November/ 1973;
The Woodlands Test Site 4
2 Schemes used for Isolation, Identification,
and Enumeration of the Genus Salmonella ... 21
3 Physical, Chemical, and Bacterial Ranges in
Lakes A and B. Values are from Surveys of
1/11/75, 10/19/75, 12/15/75, 1/24/75, and
2/15/76 34
4 Summary of Indicator and Pathogen Base Levels,
Peak Levels and Times during Storm Events . . 43
5 Summary of Physical Factor Peak Values and
Times during Storm Events 49
6 Fecal Coliform Geometric Means during
Storm Events 61
7 Summary Table of Modeling Equations 63
8 Means, Slopes, Elevations, and Correlation
Coefficients for FS on FC Regression Lines
for Land use Areas Shown in Figure 17 .... 71
x
-------
ACKNOWLEDGMENTS
This research investigation was supported, in part, by
The U.S. Environmental Protection Agency, Storm and Combined
Sewer Section, Municipal Environmental Research Laboratory, Edison, New
Jersey and the University of Texas at Houston, School of Public Health,
Houston, Texas.
The principal investigator wishes to acknowledge the
dedication and patience of his research participants, no-
tably those who contributed directly to the research sup-
ported conclusions developed from this investigation; J. D.
Moore and D. M. Casserly-
The cooperation and helpful suggestions given by Rice
University faculty and staff members were most appreciated
throughout the project. Data which were obtained from Rice
University and developed by staff of that Institution and
which were required for completion of this investigation
included those for turbidity, suspended solids, storm hy-
drograph, and rainfall data. The cooperation of the Wood-
lands Development Corporation is acknowledged.
XI
-------
SECTION 1
INTRODUCTION
Meaningful decisions regarding effective management of
water resources must be derived following the use and correct
interpretation of water quality data. This is especially
true of management of rainfall runoff, otherwise known as
stormwater. An essential part of this investigation was to
develop data which could be used for comparison purposes in
assessing the quality of stormwaters. Of equal importance
was the task of determining the effect of urbanization on
stream water quality during storm events. The need for
producing those results is emphasized by the meager amount
of data existing today which are directed to the specific
tasks stated above. Most of the emphasis on bacterial pol-
lution has been placed in municipal waste water bacterial
densities. Therefore it is justifiably important to compare
municipal data with stormwater bacteriology from rural areas.
The Woodlands, Texas served as the principal test site
for low flow and stormwater quality investigations, and is
located approximately thirty miles north of Houston, Texas.
A more detailed site description can be found in the water
quality report to the U.S. Environmental Protection Agency
by Rice University, Environmental Science and Engineering
Department; Dr. W. G. Characklis, Principal Investigator
(Grant No. R802433).
The uniqueness of the primary test site was its rural,
forested, and semitropical setting. Pines and oak abound
in the area with thick forest floor cover. The loblolly
pine (Pinus taeda), shortleaf pine (Pinus echinata) and
various oak species (Quercus spp.) constitute the majority
of stands. The soils are of a red-yellow podzollic nature
which are characteristic of a mild climate, abundant rain-
fall, and the mixed conifer-deciduous forest cover. While
it is generally believed that soils of that type are highly
leached, the test site forest floor always had a 0.5 to 4
inch leaf cover. Such decaying vegetation contributes con-
siderable amounts of organic materials to the area and the
underlying kaolinite clays which are found in those soils.
-------
This test area has channel slopes of about 0.2% and land
slopes of 0-2%, which yield low runoff coefficients (10-20%
of reported precipitation values) (1).
In the earliest stage of this investigation (November,
1973) an intensive effort was made to determine the water
quality of each of the principal streams in the primary
test site. To that purpose field reconnaissance indicated
that sixteen sampling locations would be required. Their
locations are indicated in Figure 1 and described in Table 1.
The majority of those locations were used for low-flow water
quality sampling.
Storm event monitoring occurred principally at four
locations: Stations P-10, P-30, the influent to Lake B
(Swale No. 8), and the discharge point from Lake A. Station
P-10 represented the stream water quality above development
areas whereas P-30 waters indicated the changes with distance
downstream with increased development.
The test site was well suited for testing insofar as the
purpose of this investigation was concerned. Bear Branch,
Panther Branch, and the two impoundments (Lakes A and B) were
calculated to have drainage areas of the following areal
magnitudes:
Area Drained to that Point
Location Acres (Hectares)
P-10 16,050 (6,497.9)
P-30 21,600 (8,744.9)
B-09 9,651 (3,907.3)
Lake B 337 ( 136 )
Lake A 483 ( 195 )
During periods of low flow, less than one cubic foot per
second (<0.028 cu. m/sec) has been measured at Station P-04
and approximately 5 cfs (0.14 cu. m/sec) at P-30.
Lakes A and B were constructed to provide aesthetic value
to the adjacent conference center and for the purpose of
providing irrigation water for the golf course. Figure 2 rep-
resents their locations relative to Panther Branch sampling
Station P-30 and shows the locations within the Lakes of
sampling sites for the January 11, 1975 field survey.
Locations which were chosen for bacteriological analyses of
soils near the Lakes are also included in Figure 2. The
following data characterize the lakes.
-------
p-02
B-01
Wood J ands
Boundary
LEGEND
B-01 Sampling station
p_02 Sampling station
41
t
S-09
Figure 1. Principal surface waters in primary test site and
original low sample station locations.
-------
TABLE 1. DESCRIPTIONS OF LOW-FLOW WATER QUALITY STATIONS
ESTABLISHED IN NOVEMBER 1973; THE WOODLANDS TEST
SITE.
Station Location
N-03 Southwest of Egypt, Texas on Egypt-
Decker Prairie Road in Nickaburr
Creek.
B-01 Bear Branch, vicinity FM 1488; 1.7
miles west of Egypt, Texas.
B-03 Bear Branch, in vicinity of FM 2978
(Egypt-Decker Prairie Road) southwest
of Egypt/ Texas.
B-05 Bear Branch, one mile below Station
B-03; on Louis Lane, an unpaved road
off FM 1488.
B-09 Bear Branch, 50 feet upstream of con-
fluence with Panther Branch.
P-02 Panther Branch, NE of Egypt, Texas on
Old Conroe-Magnolia Road.
P-04 Panther Branch, vicinity FM 1488.
P-08 Panther Branch, SE of Egypt, Texas on
unmarked gravel road off FM 1488 in
W.J. Jones State Forest.
P-09 Panther Branch, 100 feet upstream of
confluence with Bear Branch.
P-10 Panther Branch, 500 feet downstream of
confluence with Bear Branch. U.S.G.S.
Station No. 08068400.
P-20 Panther Branch, 1 mile upstream from
Station P-30.
P-25 Panther Branch, 1/4 mile upstream from
Station P-30. North of Sawdust Road
Gaging Station.
-------
TABLE 1. DESCRIPTIONS OF LOW-FLOW WATER QUALITY STATIONS
(CONtf'D.) ESTABLISHED IN NOVEMBER 1973; THE WOODLANDS TEST
SITE.
Station Location
P-30 Panther Branch, U.S.G.S. Gaging Station
at Sawdust Road. U.S.G.S. Station No.
08068450.
P-40 Panther Branch, 100 feet upstream of
confluence with Spring Creek.
S-09 Spring Creek, 100 feet upstream of
confluence with Panther Branch.
S-10 Spring Creek, vicinity Interstate
Highway 45 at U.S.G.S. Gaging Station.
-------
Panther
Ranch
Soil Sample Location
0 400' 800'
P-30
Figure 2. Locations of soil sampling sites and lakes A and B water quality
sampling locations (1/11/75 survey).
-------
Parameter Lake B Lake A
Length, ft. (m) 1,100(335) 1,400(427)
Width, ft. (m) 150-200(45.7-61) 300-500(91-152)
Depth, avg., 6(1.83) 8(2.44)
ft. (m)
Surf, area, 3.3(1.34) 12.6(5.10)
ac. (ha)
Volume, ac.ft. 20(24,669) 90(111,014)
(cu.-m)
Two stream locations within the Houston metropolitan area
were chosen to monitor storm events for the purposes of com-
parison of rural and urban storm runoff quality. The first
of those two, Hunting Bayou, is located in northeast Houston
and is designated as U.S.G.S. Station No. 08075760, Hunting
Bayou at Falls Street, Houston, Texas. A flood-hydrograph
partial record station is located in the drainage area of 3.50
square miles. Located in a heavily populated older neighbor-
hood, the waters in Hunting Bayou at that location always
showed signs of being heavily polluted with rubbish, free and
floating oil, and other materials.
The other urban stormwater monitoring site was the
Westbury site, located in southwest Houston in a drainage
ditch at Atwell Street. The area drained is 210 acres (85 ha)
and is composed of relatively newer homes, paved streets, and
established lawns. Measurements were made at that location
with a flow meter and staff gage.
OBJECTIVES
The period of time during which data were obtained from
both field and laboratory analyses and measurements was from
November, 1973 through May, 1976. That extent of time was
divided into three units for reporting purposes: November 1,
1973 - May 31, 1974; June 1, 1974 - May 31, 1975; and June 1,
1975 - May 31, 1976.
The following eight objectives were to be accomplished
during the project period: 1) monitor the bacteriological
quality of Bear Branch and Panther Branch during periods of
low flow as well as the low flow characteristics of Lakes A
and B; 2) monitor the bacteriological quality of stormwater
runoff within the primary test site (the Woodlands), in-
cluding the stormwater inflow and discharges from Lakes A
and B; 3) establish the relevance and usefulness of indica-
tor organisms (bacteria) as indicators of the hygienic
quality of stormwater runoff; 4) evaluate the aftergrowth
potential of bacterial indicator organisms found in storm-
water runoff; 5) evaluate the disinfection requirements of
stormwater runoff; 6) compare the densities of indicator
-------
bacteria in stormwater runoff with the densities of similar
groups in municipal wastewater; 7) determine the changes
imposed on stream bacteriological quality by urbanization
of a rural forested area; and 8) to evaluate the effects of
sedimentation on bacterial populations in stormwater runoff.
Those objectives included the enumeration of indicator
groups and groups or genera of potentially pathogenic bac-
terial populations. The groups and genera which were iden-
tified, along with the codes used in this report are as
follows: "TC" is total coliform; "FC" is fecal coliform;
"FS" is fecal streptococci; "SS" are presumptive members of
the Salmonella-Shigella group; "SA" is Salmonella sp.; "PS"
is Pseudomonas aeruginosa; and "ST" is Staphylococcus aureus
Other tasks were requested during the period of inves-
tigation due to questions concerning the sources of high
bacterial counts and other related reasons. They consisted
of the following: 1) assessment of the bacterial content
of Hunting Bayou and the Westbury test site stream; 2) de-
termination of the source of bacteria, i.e., stream sedi-
ment, soils, humus, or from swales; 3) determination of
variability in blended sample counts; 4) determination of
the variability in counts within a single sample, i.e. the
precision and accuracy; 5) identification of count variabi-
lity in a stream during low flow over a 24-hour period; and
6) determination of bacterial dieoff characteristics in re-
frigerated (4°+0.5°C) versus non-refrigerated (25°+l°C)
samples for up to 24-hours.
BACKGROUND AND RELATED RESEARCH
The importance of accurate assessment of bacteriologi-
cal water quality cannot be overemphasized. Elevated den-
sities of pathogens have been directly implicated with on-
set of disease, and as far as that subject is concerned
the literature is repleat with documentation of the public
health importance of the indicator groups and pathogenic
bacterial species. Several recommendations and sets of
standards have been reported in recent years for bacterial
quality of recreational waters. One report (2) suggested
the addition of iodine and chlorine to a recreational lake
for bacteria disinfection and algae control and recommended
a level of fecal coliforms (FC) not to exceed 200/100 ml.
Pathogenic content correlation with indicator bacteria is
gaining more attention (3) due to the fact that all too
often the total coliform (TC) counts do not represent the
potential for infection. The American Chemical Society's
official evaluation (4) of the TC test results was stated
as follows: "There is no significant epidemiological basis
-------
for the total coliform bacteria counts used today to es-
tablish standards for body-contact recreational waters."
Some numerical values for bacterial densities had to
be chosen for comparison with those found during this in-
vestigation. The standards chosen were recently adopted
by the Texas Water Quality Board (5). For contact recrea-
tion the fecal coliform geometric mean shall not exceed
200/100 ml with no fewer than 5 samples collected in 30
days and not more than 400/100 ml shall exist in more than
10% of the total samples in the 30-day period. For noncon-
tact recreation the fecal coliform standard is one order of
magnitude higher, i.e., 2,000/100 ml and 4,000/100 ml, res-
pectively. Not only are standards being changed from TC,
some authors recognize the need for specific pathogenic ge-
nera as water quality indicators. Pseudomonas aeruginosa
and streptococci are among those which have received such
attention (6,7). Pseudomonas aeruginosa is equally as im-
portant a pathogen as the Salmonella group. But testing
for pathogens is far less amenable to routine analysis than
TC, FC, and FS. The media available today are not as selec-
tive as they should be, especially for water bacteria.
Most media developed for pathogen identification have been
perfected employing clinical specimens. No doubt exists
that pathogens may occur along with total coliform bacteria,
Geldreich (8) reported that greater chances exist for Sal-
monella sp. isolation when fecal coliform concentrations
exceed 200/100 ml. Attention is hereby directed to the
data.presented herein in which virtually all FC counts
greatly exceeded that level. Geldreich further reported an
85.2% probability of SA isolation when the FC levels were
200 - 2,000/100 ml, and a 98.1% chance when they were
greater than 2,000/100 ml. Bott (9) expressed the situa-
tion quite adequately in stating that due to the fact that
many human pathogenic species are harbored in non-human
hosts, it is generally unwise to limit one's concern to the
subject of recent human fecal pollution as the sole source.
Earlier reports halve avoided estimating the numbers of
specific species which when ingested would result in the
respective disease. Only recently. Rosenberg, of the Cen-
ter for Disease Control, stated that it requires 10 orga-
nisms to result in shigellosis and between 105 and 108 for
salmonellosis to occur (10). And waterborne disease inci-
dences appear to be on the increase (10). A recent out-
break occurred at Crater Lake National Park (11) when 288
park employees and over 1,000 visitors were infected with
the first reported toxigenic E. coli, serotype 06:H16. Ap-
plication of the most bactericidal agent may not be the fi-
nal answer in protecting the public health. Englebrecht
-------
(12) reported an order of resistance to chlorine of the fol-
lowing organisms or groups: acid fast organisms > yeasts
> polio type I virus > Salmonella typhimurium > E. coli.
The implication being that other pathogens may survive chlo-
rination attempts. With improved culture techniques, patho-
gens are finally receiving the attention they rightfully de-
serve. Highsmith and Abshire (13) found Pseudomonas to
have occurred in creek waters in as large numbers as FC
and in greater numbers than FS in all samples they tested.
Pseudomonas sp., of course, can cause severe ear, eye, and
nasopharyngeal infections. Furthermore, of all places to
find large numbers of that pathogen, Carson, ejt al. (14)
reported Pseudomonas sp. existence in hospital distilled
water mist therapy units.
10
-------
SECTION 2
CONCLUSIONS
A grab sample represents the stream condition only at
one point in time. The diurnal coefficient of variability
of indicators and pathogens in a small rural stream can be
as much as 126% (o/X). Normal variability in indicator and
pathogen growth using standard culture media for a single
sample ranges from 9% for ST (a/X) to 35% for SA (a/X).
The indicator bacteria groups and pathogenic genera
which were tested settled at different rates in quiescent
stream water samples which contained suspended solids con-
centrations above 200 mg/1. ST settled effectively in 12
hours in secondary treated municipal wastewater but PS and
SA populations tended to remain in suspension. TC, FC, and
FS populations also did not settle effectively in the waste-
water.
Blending of highly turbid samples up to 10 seconds can
produce higher indicator bacteria counts. Additional blend-
ing time resulted in reduced counts probably due to cellular
disruption by heat and shearing action. Some counts can be
increased 56% but most groups increase to levels within sin-
gle sample variability percentages. Blending is not feasi-
ble for storm water analysis due to time constraints in
plating large numbers of samples and sterilization of blend-
ing equipment between samples.
Preservation of field samples for bacteriological ana-
lysis should be accompanied if the time between sampling
and analysis exceeds 12 hours. Samples kept at 4°C and
2Q°C showed no significant differences in counts between
TC, FC, FS, ST, and SA for up to 12 hours. However, PS in-
creased in density in the refrigerated sample within 12
hours. In no cases were variances in samples greater than
the diurnal field sampling variances quoted above.
Chlorinated secondary treated municipal wastewater ex-
hibits FC/FS ratios (- 1.5) which suggest contamination of
the water by non-human sources. Percent coefficient of _
variation variances in treated municipal wastewaters (o/X)
11
-------
were significantly higher than in small urban stream over a
24-hour period and within settled versus agitated water
samples containing high solids and turbidity.
Irrigated high humus soils can yield greater counts of
TC, FC, FS, PS, ST, and SA than rural streams receiving no
point source contributions. Stream sediment, surrounding
soils, and impoundment sedimentary materials can contain
greater densities of TC and ST on number/unit weight or
volume basis than stream waters.
Under low flow conditions, small rural streams can
exhibit differences in upper and lower reaches insofar as
bacterial quality is concerned. On the average, station
P-30, downstream, contained FC densities of 2.38 (log
value) while the upper reaches had a value of 2.13. Com-
pared to a contact recreation standard of 2.3 the lower
reach was not acceptable for contact recreation. Statisti-
cally, this upstream log geometric mean was not different
from the value for the waters downstream. But on the basis
of the not to exceed 2.3/100 ml standard, the downstream
waters exceeded the value. Bacterial concentrations under
low flow conditions exceeded single sample variability and
diurnal variability. TC were not reliable indicators of
pathogen, FC, or FS densities under low flow conditions.
The impoundments investigated in the primary test site
showed reduced FC, FS, and pathogen concentrations compared
to inflow storm water concentrations. Dieoff and settling
both contributed to reduced bacterial populations. TC con-
centrations, on the other hand, remained as high as in
storm water, suggesting that that group of indicators may
not be a reliable water quality index of pollution.
Excessive doses of chlorine or ozone are required for
disinfection of storm waters which contain high suspended
solids. Solids concentrations greater than 30 mg/1 effec-
tively reduce the disinfectant efficiency similarly. Im-
poundment waters during storm events exhibited a chlorine
demand greater than 8 mg/1 and an ozone demand greater than
32 mg/1. P-10 waters had a chlorine demand of 8 mg/1 and
P-30 waters had one of 8 mg/1 to 16 mg/1, depending on the
quality of the water. Given 2, 4, 8, and 16 mg/1 chlorine
doses, TC bacterial populations in storm waters behave
similarly, i.e., decreases at thirty minutes followed by
increases to 107/100 ml in eight days. FC, FS, and some
pathogens decrease to 10/100 ml or lower and will remain at
that-level for up to eight days. ST can reestablish their
populations after four days. All indicators and pathogens
respond the same to ozone as to chlorine.
12
-------
Urban storm runoff contains bacteria concentrations
which may be two orders of magnitude higher than those in a
small rural stream. Extreme variability can occur between
samples in urban areas. Also, indicators and pathogens can
continue to increase in concentration throughout a storm
event in urban runoff. TC ranges of 3.1 x 10^/100 ml to
7 x lO^/lOO ml can be found in urban runoff. In a well de-
fined storm event in a rural watershed, TC concentrations
show less clearly defined pollutographs compared to FC and
FS. The term "pollutograph" conforms with the terminology
established by the U.S.E.P.A. in usage of the Storm Water
Management Model. Indicators and pathogens peak prior to
the hydrograph peak during the first flush phenomenon and
closely correlate with solids and turbidity peaking times.
FC/FS values show more constancy in lower reaches of a small
rural stream than in its upper reaches. FC, FS, and ST can
peak earlier in the lower watershed than in upper stream
areas. Rainfall duration and intensity can exert an in-
fluence on physical stream factors. Solids peaks during
storm events ranged from 0.3 cfs to 1,250 cfs, and turbidity
ranged from 4 J.T.U. to 2,000 J.T.U. Log mean value ranges
for all Woodlands storm events were TC, 4.20-7.76; FC, 0.85-
3.63; FS, 2.19-5.12; PS, 0.47-4.66; ST, 1.50-4.42; and SA,
0.28-1.94.
In thirteen out of twenty-four monitored hydrographs,
the mean FC values exceeded a non-contact recreation stan-
dard of 2,000/100 ml and 22 of the 24 exceeded the contact
recreation standard of 200/100 ml. 22 of 123 data exceeded
a 4,000/100 ml FC standard. The exceptions were the im-
poundment effluent data, signifying the usefulness of stor-
age reservoirs in stormwater bacteria reductions. Non-lin-
ear relationships exist between bacterial populations and
physical factors in storm waters. Log transformation of
bacterial data is essential to effective interpretation of
pollutographs. FC and FS yielded the highest correlations
with physical factors (Q, SS, T) for predicting stormwater
quality. FC can increase as the turbidity and solids while
FS decrease. FS can increase as the discharge and FC. FS
was found to have been the most useful quality predictor.
Relationships of FC and FS with highest correlations were
Log FC = .719 log FS - .002Q + .003T + .002SS + .612
(R2 = .77)
Log FS = .720 log FC + .002Q + .151 log TC + .001 T -
.001SS - .170
(R2 = .82)
Multiple regression analysis and other statistical analysis
approaches showed FC/FS values to have virtually no corre-
lation for storm water compared to low flow conditions when
examined together or separately- P-30 storm events yielded
13
-------
much higher correlations than the waters in the upper reaches
(P-10) . The most significant indicator was
Log ST = .000668T 4- .779 Log FS + .000625SS -
.489 Log FC + .167 Log TC + 1.051
(R2 = .79)
Combining data from four stations in a representative storm
event in the Woodlands (3/12-14/75) with 97 data points,
showed PS and ST to have had the highest correlations.
Log PS = .000862T + .665 Log TC - .0233Q - 1.757
(R2 =8.5)
Log ST = .259 Log TC + .000612T + .968 Log PS +
.0278Q - .209 Log FC - .589
= .90)
In virtually all of the statistical analyses involving
correlations of FC/FS values with pathogen densities, the
R2 values were less than 0.20 and none were significant at
the 0.5 significance level. SA were the most difficult
bacteria to predict in storm waters. Pathogens can be pre-
sent in surface waters when indicators are either not pre-
sent or in such low numbers that detection methods did not
record them. Considering all data, when Salmonella sp. were
detected, 6 samples contained no FC, 7 samples contained no
FS, 29 samples contained no PS, and 3 samples contained no
ST. Salmonella sp. were not detected when TC were 4,500/
100 ml.
A graphical analytical approach can be a most useful
tool in water quality analysis. This application showed
distinct increases in FC and FS with increased urbanization,
considering 5 land use areas (Figure 16). FC and FS dis-
tributions were not constant for different land uses. Sta-
tistical analysis (LSD test) determined that no significant
differences occurred in FC/FS values for the following
water sources: storms at Lake B (FC/FS = .53); low flow,
Lake A (.58); chlorinated wastewater (.73); low flow at
P-10 (.92); storms at P-30 (.97); low flow at Lake B (.99);
low flow at P-30 (1.26); storms at Westbury (1.47); low
flow at all Woodlands stations (1.68); storms at P-10
(2.11); and raw sewage (2.42).
The two-factor analysis of variance may be the most
applicable approach to comparing bacterial concentrations
between upper stream reaches (P-10) and lower reaches (P-30)
No significant differences in TC concentrations were found
between the two locations. But, FC were 2.7 times as con-
centrated in storm waters at P-30 (downstream) than at P-10
(upstream): FS were 5.2 times as concentrated, PS were 5.7
14
-------
times as concentrated, and ST were 5.0 times as concentra-
ted.
No statistical differences were observed in FC and FS
concentrations on a seasonal basis during low flow condi-
tions at the Woodlands. ST concentrations showed less varia-
tion between seasons than did PS or SA. TC concentrations
were significantly higher in summer than winter and spring
(statistically equivalent), and in turn those were signi-
ficantly higher than fall and winter (statistically equiva-
lent) .
15
-------
SECTION 3
RECOMMENDATIONS
Future investigations on water resources for bacterial
assessment and management of storm water runoff should es-
tablish base flow diurnal variability in indicator bacteria
concentrations prior to evaluating storm events. Concur-
rent identification of point sources should be accomplished
in order that accurate decisions can be made relative to
the need for specific treatments of the resource. In es-
sence, the variables which exist in each individual area
will dictate the quality of surface water runoff and they
must be dealt with only on a one time basis.
If disinfection of storm waters is a measure which is
being considered for pollution control in a specific area,
several factors must be taken into account prior to appli-
cation. The disinfectant demand of the waters should be
established or adequate disinfection cannot be accomplished.
Following that, it is essential that the need for disin-
fection on an area by area basis be justified, taking into
account whether the public health would be endangered if
disinfection were not carried out. Cost effectiveness
should be considered if stormwater containing not domestic
wastewater is to be disinfected.
Fecal coliform/fecal streptococci ratios have been
proposed as reliable indicators of the sources of bacterial
pollution. That concept is applicable only if an FC/FS
value is established at the source. In view of the marked
differences in concentrations of indicator bacteria in any
water source with time, a factor which should also be taken
into account is land use. Indicator bacteria as well as
pathogens may demonstrate trends between low flow condi-
tions and during storm events which are peculiar to each
area taken into consideration. Total coliform bacteria
were not reliable in relating pathogen content of the water
sources. The high correlations of fecal coliforms shown
earlier, again strongly indicate that they are a primary
indicator of fecal pollution. Additional analyses should
also be conducted for pathogen groups such as Salmonella
sp., Pseudomonas sp., and/or Staphylococcus sp.
16
-------
Through the years, physical and chemical tests for
determining water quality have been refined to the point
where most are highly accurate and precise. The science of
aquatic microbiological quality (and quantity) assessment has
lagged far behind in technique refinement. Membrane filtra-
tion techniques have made an appreciable improvement in this
area. Yet, when using that technique with storm waters,
the high solids content impedes some bacterial enumeration
efforts due to the inability of the membrane to filter
adequate amounts of water for proper enumeration. Clearly,
alternative methods must be developed yielding equal or
greater reliability for bacterial indicator and pathogen
quantification. Additionally, most of the commercially
available culture media were developed from clinical settings.
Water quality assessment requires the use of more selective
and more rapid identification types of media.
Settling and dieoff patterns of bacteria in impoundments
with relatively short detention times suggest that if dis-
infection is to be applied to a particular storm water source,
if at all feasible, the equivalent of stilling basins should
be employed. That will significantly reduce the amount of
disinfectant required and, of course, it will reduce the
cost.
The time between when a sample is taken and when aliquots
are cultured for bacterial analysis should not exceed twelve
hours if the storage temperature does not exceed 20°C.
Increased times may result in significantly different counts
than were originally present in the water. That, in turn,
could lead to erroneous conclusions regarding the quality of
the water under consideration.
Chlorine is significantly longer lived than ozone in
storm waters containing elevated solids concentrations. If
disinfection is considered that factor should be taken into
account prior to application.
Examination of bacterial data for comparison with other
water quality parameters for analysis or resource management
purposes should not be accomplished using a simple arithmetic
approach because non-linear relationships exist between
variables. Bacterial data should be transformed to Iog10
basis.
17
-------
SECTION 4
MATERIALS AND PROCEDURES
Field sampling was accomplished by personnel from
The Environmental Science and Engineering Department, Rice
University, Houston, Texas and the University of Texas at
Houston, School of Public Health, Environmental Science
Discipline.
BACTERIOLOGY
The methodology incorporated in the laboratory research
associated with this project involved standard techniques
for assaying defined groups of bacteria, genera or species.
Detailed descriptions for routine proven methods such as
autoclaving media and glassware and membrane filter counting
procedures are given in Standard Methods (15) and will not
be detailed herein.
Water samples taken during periods of low flow were
iced-down immediately and maintained in that condition until
analyses could be conducted on each sample. All low flow
analyses were initiated well within three hours after the
samples were gathered in the field. During storm events,
sample batches of six-hour duration were returned to the
laboratory, at which time the dilution, filtration, plating,
and incubation processes were started immediately. Reported
bacterial numbers (colonies developed on the respective
media) are the result of replicate plating in all cases.
During the time period of November, 1973 through May,
1974, methods which were applied for routine analysis for
total bacteria populations, total coliform (TC), fecal coli-
form (FC), and fecal streptococci (FS), were those described
in Standard Methods (15). Total coliforms, fecal coliforms,
and fecal streptococci were counted on membrane filters. ,,
FS were identified on M-enterococcus medium. The Salmonella-
Shigella group was identified with the aid of Bismuth Sul-
fite Agar (16, 17, 18). Pseudomonas sp. were identified
and quantified with Pseudomonas Agar F (Difco) (19). Pseu-
domonas Agar F was deleted soon after the earliest phases
of this investigation and a new medium, m-PA agar, was uti-
18
-------
lized (20). Pathogenic Staphylococci were identified with
Mannitol Salt Agar (18) only in the early phases of the
investigation.
Confirmation of colony taxonomy as appeared on selec-
tive media such as Bismuth Sulfite Agar was checked by
using a relatively new biochemical analysis system (21).
The "API 20 ENTERIC" was employed in this regard. Briefly
this system exposes a test organism (colony) to twenty bio-
chemical tests. Positive, negative, or other reactions are
recorded after the appropriate procedure is applied for
analysis. The results are coded and the test organism
keyed-out by means of established diagnostic hardware which
differentiates between the species of bacteria in the En-
terobacteriaceae.
Following May, 1974, all data for Salmonella were de-
veloped using the following procedure. XLD Agar (Xylose
Lysine Deoxycholate Agar) (BBL) plates were inoculated with
a precise volume of test water which was allowed to evaporate
to near dryness at room temperature (25 + 1°C). Incubation
at 37°C for 24 hours followed. Typical colonies were
counted and recorded. All typical colonies were then trans-
ferred to TSI (Triple Sugar Iron Agar) (BBL) slants which
were then incubated for 24 hours at 37°C and their reactions
recorded. Slants yielding positive reactions were subjected
to further confirmation by use of API test strips which em-
ploy twenty-one separate biochemical reactions (21). A
description of reactions of Salmonella sp. on XLD Agar and
TSI Agar can be found in Standard Methods (15). At least
10% of the colonies were picked for confirmation.
Vogel-Johnson Agar (BBL) was used for the identifica-
tion of Staphylococcus sp. It isolates coagulase-positive,
mannitol-positive Staphylococci, Staphylococcus aureus, a
proven pathogen, is easily isolated on this medium. It has
been reported that Staphylococcus aureus is capable of iso-
lation on that medium to the point of nearly excluding other
forms (22, 23). Additipnal testing of Staphylococcus sp.
was conducted using coagulase and catalase tests (22, 23).
An example of the selectivity of the principal medium used
for Staphylococcus sp. quantification lies in one series of
picked colonies (60 total) of which 57 were coagulase-posi-
tive and catalase-positive, thereby confirming their iden-
tity. Attention is hereby directed to the pathogenic bac-
teria genus, Arizona. It was formerly named Salmonella sp.
and is included in the numbers of Salmonella sp. reported
herein. Citrobacter sp. was found"to have been the prin-
cipal false positive.reaction organism on XLD plates but
the, total numbers found were less than 5% of the total
positives.
19
-------
As time progressed, the elevated numbers of Salmonella
sp. which were being reported, especially from stormwaters
in the Woodlands, began to be suspect. Further confirma-
tion was accomplished using the schemes presented in Table
2. Schemes II and III were preferable due to the fact that
quantification could be achieved by their use. Under
crowded conditions many Salmonella colonies on XLD did not
exhibit black centered colonies. Only well isolated colo-
nies yield typical colonial morphology. On plates that are
overcrowded, numerous black globs were sometimes observed
but recovery efficiency of Salmonella sp. or Arizona sp.
from those was low. These data are similar to those re-
sults presented by Kenner and Clark (24), conducted in
U.S.E.P.A. laboratories. Their confirmation steps were not
as extensive as those outlined above and followed through-
out this investigation.
DISINFECTION EXPERIMENTATION
Methodology applied to water samples taken from the
Woodlands sites for special studies of disinfection require-
ments during storm events and during periods of low-flow
included the following procedures and equipment. Chlorine
addition was accomplished by use of standardized solutions
of sodium hypochlorite. Studies of break-point chlorination
used the amperometric titration methods described in Stan-
dard Methods (15) for free available, total residual, and
combined residual chlorine. Quantification of the amount
of bromine added to water samples was accomplished by ac-
curately weighing elemental bromine and immediately dis-
solving same in appropriate volumes of distilled water to
yield the required disinfection dosage. Ozone dosing was
accomplished through May, 1974 with a Welsbach Ozonator,
Model T416. The remainder of ozonation experiments were
conducted using a Sander Ozonizer, Type III. Dosing of
representative water samples with ozone was calculated to
equal the chlorine and bromine dosaqes. Over periods of
time, such as up to 21 days duration in some of the disin-
fection experiments, one set of water samples were left
standing at room temperature and the other set was placed
in a Precision Scientific Freas Incubator at 20°C + 1°C.
Roam temperature was considered to have been 25°C + 1°C.
All vessels in which the disinfection series sample's were
placed were sterilized prior to addition of the water.
Each vessel was covered with sterilized aluminum foil to
retard evaporation and allow oxygen transfer to maintain
aerobic conditions.
20
-------
TABLE 2. SCHEMES USED FOR ISOLATION, IDENTIFICATION, AND
ENUMERATION OF SALMONELLA.
Scheme I;
Qualitative
Scheme II:
Quantitative
Scheme III:
Quantitative
10 ml or 25 ml
sample membrane
filter
tetrathionate
broth tubes
(24 hrs.)
streak
XLD (18-24 hrs.)
1 or 2 ml sample
evaporate
on XLD
XLD for 18-2,4 hrs
1 ml, 2 ml, or 5 ml
sample: membrane
filter
>_10% black centered colonies
streaked for isolation
J,
XLD
tetrathionate broth
pad (7-12 hrs.)
XLD for additional
12-24 hrs.
black centered colonies
TSI slants
I
oxidase test
API
I
serological testing,
salmonella polyvalent antisera
false positive
XLD - xylose lysine desoxychlate agar
TSI - triple sugar iron agar
API - battery of 21 biochemical tests for classification of
the family Enterobacteriaceae
21
-------
BLENDING
Blending of samples was performed to determine the effect
on colony counts. The commercial laboratqry blender which
was used was a Waring Model 7010S (formerly Model 5010S) , 480
watt, 120 volt, 60 Hz and 17,500 rpm base, with four blades.
Two blades pitched up at 30° with the horizontal were 2.562
inches and the two blades pitched down at 25° to the hori-
zontal plane were 2.375 inches in length. Each blade was
designed to serve as an inverted air foil having a 0.008 inch
invert from the leading edge.
STATISTICAL ANALYSES
Multiple regression analyses appeared to have been the
most appropriate method for data analysis. To that purpose,
a package included in Statistical Package For The Social
Sciences (25) was used. The program incorporated a stepwise
multiple regression for selecting the dependent variable with
the highest predictability combined with the least number of
independent variables. Prediction equations were developed
and are presented below.
Other programs used included REGRESSION, ANOVA, ONEWAY,
SGATTERGRAM, FREQUENCIES, and an analysis of covariance
(courtesy Dr. Arthur Littell, Professor of Biometry, U.T.
S.P.H.). The University of Texas at Houston Education and
Research Computer Center facilities were used with a Control
Data Corp. Model No. CYBER 73 having been the instrument
employed.
STORM SAMPLING EQUIPMENT
Four stations in the Woodlands were used for storm
analyses. The stations and equipment located at each are
described below:
1. Station P-10: U.S.G.S. Station No. 08068400.
Discrete sampler; Stevens Type A recorder for
stage height and manometric water level sensor.
2. Station P-30: Similar to those described for
P-10. U.S.G.S. Station No. 08068450.
3. Lake B: U.S.G.S. Float gage with stilling well
for stage height, on bridge over Swale 8, Manning
S-4000 discrete sampler.
4. Lake A: U.S.G.S. float gage with stilling well
for level and a Sigmamotor WM24-R sampler on a
pier adjacent to the Lake overflow-weir.
22
-------
A more detailed description of stormwater sampling procedures
can be found in the water quality report to the U.S. Environ-
mental Protection Agency by Rice University, Environmental
Science and Engineering Department: Dr. W. G. Characklis,
Principal Investigator (Contract No. R802433).
SETTLING EXPERIMENT
To test the effects which the settling of suspended
solids had on bacterial densities, equal volumes of a single
sample were placed in 5-gallon carboys. One was shaken con-
tinuously throughout the test period and the second was left
unagitated. Samples were drawn from a depth of 15 cm in all
instances.
23
-------
SECTION 5
RESULTS AND DISCUSSION
SINGLE SAMPLE VARIABILITY AND DIURNAL STREAM STUDY
Ten replicates were plated from a single water sample
obtained from station P-02 and analyzed for TC, FC, FS, PS,
ST, and SA. The purpose was to determine variances within
populations within a shaken water sample. The results of
those analyses are £ummarized below. Units are number/100 ml
except FC/FS and cr/X. Number of replicates = 10.
Mean (X) Std. Dev., (o) o/X
Total Coliform 60,836 12,805 21
Fecal Coliform 484 127 26
Fecal Streptococci 796 107 13
FC/FS 0.63 0.22 35
Salmonella sp. 5,310 1,208 23
Staphylococcus sp. 22,718 2,012 9
Pseudomonas
aeruginosa 420 86 20
These data are significant in that they represent the vari-
ability which can be anticipated within any water sample
subjected to analysis for the indicated species and groups
of bacteria and at those concentrations. Variability in
bacterial populations in a stream over a period of 24 hours
during low flow conditions are summarized below. The values
were obtained from station P-10; number of £amples = 18.
Units are number/100 ml except FC/FS and a/X.
Mean (X) Std. Dev., (a) a/X
Total Coliform 56,611 40,478 71
Fecal Coliform 164 127 77
Fecal Streptococci 227 65 29
FC/FS 0.8 0.9
Staphylococcus sp. 287 195 68
Pseudomonas
aeruginosa 258 325 126
Salmonella sp. 554 532 96
24
-------
Concurrently, station P-30 was sampled and the results of
those analyses are summarized below. Units are the same as
those above. Number of samples = 23.
Mean (X) Std. Dev., (g) g/X
Total Coliform 80,696 20,379 26
Fecal Coliform 213 66 31
Fecal Streptococci 213 237 111
FC/FS 1.3 1.2
Staphylococcus sp. 792 582 73
Pseudomonas
aeruginosa 185 123 66
Salmonella sp. 279 188 67
Those data suggest that interpretation of data derived from
a single grab sample should reflect the conditions of a
stream only at the point in time and location from which_
the sample was obtained. The percentage variability (o/X)
within all of the indicated groups and species was higher
at both stations throughout the 24-hour period than within
a single sample.
BLENDING OF SAMPLES
The results of blending samples of prechlorinated se-
condary treated municipal wastewater indicate that all three
indicator groups exhibited more than double the counts at
the ten second blending interval but order of magnitude
differences were not in evidence. In some cases, longer
blending times resulted in lower counts than the controls.
Blending of samples for disruption of bacterial clumps is
impractical for stormwater samples because of the large num-
ber of samples taken during each storm event and the time
limitation of sterilizing a blender between each sample and
the time limitation for plating the samples before each be-
comes invalid. This blending experiment was requested by
U.S.E.P.A. personnel. /The data are presented below.
Blending Time, Seconds
Q 10 20 40 60 90
Total Coliform, q
No./lOO ml x 10 1.4 3.2 1.3 1.3 2.0 1.9
Fecal Coliform, ft
No./lOO ml x 10a 2.8 6.4 1.9 2.9 3.1 5.1
Fecal Streptococci,
No./lOO ml x 104 4.9 7.3 4.7 6.0 5.4 4.2
25
-------
SEDIMENTATION EXPERIMENTS
Settling experiments were conducted to determine the
extent of reduction in bacterial densities in the water
column over time. Data obtained for a settling time of 30
minutes using water samples taken from stations S-10 and
P-20 are summarized below for the three indicator groups.
Units are as number/100 ml.
Station S-10
Agitated
Settled
Station P-20
Agitated
Settled
TC
2.17x10;
0.5 xlO'
3.1 xlO
1.6 xlO'
FC
7.37x10
0.36x10'
3.0 xlO:
2.8
FS
7.3x10:
6.8x10'
9.7x10!
5.1x10'
The settling rates of the indicator groups differed between
the two samples taken from different locations. Suspended
solids concentrations probably influenced the results.
In a repeated experiment which included pathogen quan-
tification and settling times up to 12 hours, indicator bac-
teria did not demonstrate clearly defined settling patterns
and FC/FS values were irregular. Pseudomonas (PS) and Sal-
monella (SA) did not effectively settle whereas Staphyloco-
ccus (ST) decreased in numbers within the water column.
The results suggest that water samples containing elevated
quantities of settleable solids (>200 mg/1), will exhibit
low counts if not shaken immediately prior to sampling and
plating.
MUNICIPAL WASTEWATER BACTERIOLOGY
Fourteen samples were obtained over a period of as
many days of untreated, secondary treated, and chlorinated
secondary treated municipal wastewater. They were taken
from two of Houston's largest and most efficiently operated
treatment plants. These tests were requested by U.S.E.P.A.
personnel in order that comparisons could be made with
stormwater bacteria densities in the same geographical lo-
cale. The data obtained from the analysis of those samples
are summarized below. Units are number/100 ml except FC/FS
and a/X (coefficient of variation).
26
-------
Mean (X) Std. Devi., (a) c/X
Total Coliform
Fecal Coliforra
Fecal Streptococci
FC/FS
Total Coliform
Fecal Coliform
Fecal Streptococci
FC/FS
Total Coliform
Fecal Coliform
Fecal Streptococci
FC/FS (n=7)
Untreated Wastewater
41.24x10.
30.05x10
101.11x10.
68.86x10
18.83x10 20.72x10
12.8 9.3
Prechlorinated Secondary
Treated Wastewater
16.26x10'
27.66x10'
13.56x10^
44.88x10:;
75.56x10
30.69x10
126 64
Chlorinated Secondary
Treated Wastewater
245.7
18.1
6.7
1.5
318
24.9
9.5
0.7
245
229
110
83
162
246
130
137
142
Perhaps the most striking result is the indicated change in
FC/FS. The chlorinated effluent FC/FS ratio of 1.5 suggests
the pollutant source to have been of mixed origin (man and
other animals). Prechlorinated secondary treated indicator
bacteria densities were within levels observed in storm water
populations. Ten additional samples were analyzed in order
to compare pathogen densities with indicator bacteria densi-
ties. The data for these ten analyses are summarized as
follows; values are for number/100 ml.
Mean (X)
Std. Dev., (a)
Total Coliform
Fecal Coliform
Fecal Streptococci
Pseudomonas aeruginosa
Staphylococcus sp.
Salmonella sp.
Total Coliform
Fecal Coliform
Fecal Streptococci
Pseudomonas aeruginosa
Staphylococcus sp.
Salmonella sp.
Untreated Wastewater
59.07xl08
32.66xl05
5.1 xlO5
1.69xl05
2.79X105
7.98x103
135.76x10®
45.75xl05
1.51x105
1.41xl05
2.13xl05
19.72xl03
Prechlorinated Secondary
Treated Wastewater
44.87x10
3.17xl05
1.35xl04
2.66xl03
3.54x104
1.40xl02
94.80X106
1.68xl05
0.49xl04
1.71xl03
4.92xl04
2.00xl02
27
-------
Mean (X)
Std. Dev., (o)
2.73xl02 '
41.8
140.0
2.8
3.74xl02
ND
6.19X102
125.9
263.3
4.1 j
4.68x10
ND
Chlorinated Secondary
_ Treated Wastewater
Total Coliform
Fecal Coliform
Fecal Streptococci
Pseudomonas aeruginosa
Staphylococcus sp.
Salmonella sp.
(ND = none detected)
Salmonella sp. were not detected in 50 ml of sample. The
chlorination apparently resulted in the virtual elimination
of Salmonella sp. but permitted substantial numbers of total
coliforms to survive. It was least effective against
staphylococci.
SAMPLE PRESERVATION STUDY
The effect of storage temperature on the bacterial
content of samples was important for data analysis due to
travel time -and sampling intervals during storm events. The
samples were obtained from station P-30 on January 11, 1976.
The data derived from the analyses of all bacterial groups and
species are summarized below as the mean values of five
replicates at each of the three times. Units are number/100
ml.
Total Coliform
Fecal Coliform
Fecal Streptococci
Pseudomonas aeruginosa
Staphylococcus sp.
Salmonella sp.
Total Coliform
Fecal Coliform
Fecal Streptococci
Pseudomonas aeruginosa
Staphylococcus sp.
Salmonella sp.
Time, Hrs.
6
Stored at 20°C
1.65xl05
2.29xl02
2.45x103
3.48x10!?
1.16xlOJ
1.9 xlO2
1.71xl05
2.20xl02
2.56xl03
3.46xl04
l.lSxlO3
1.7 xlO2
Stored at 4°C
1.67xl05
2.51xl02
2.62xl03
3.53xl04
1.36xl03
2.0 xlO2
1.62xl05
2.66xl02
2.37xl03
3.52xl04
1.38xl03
1.9 xlO2
12
1.63xl05
2.34xl02
2.73xl03
3.10x10
l.llxlO3
1.8 xlO2
l.SOxlO5
2.40xl02
2.39xl03
3.02xl04
1.42xl03
2.2 xlO2
28
-------
A statistical analysis (t-test) indicated that the only
significant difference was found with Pseudomonas aeruginosa.
The standard deviation was lower, however, than those cal-
culated for the variance within a single sample; the data for
which have been presented above. It is important to note
that the greater average decrease in Pseudomonas aeruginosa
concentration occurred in the refrigerated sample. In all
probability the differences are not significant when compared
to field sampling variations.
SOIL AND SEDIMENT ANALYSES
An important factor in the bacterial contamination of
waterways during and following storm events is the stream
sediment bacterial content, content of soils in adjacent
areas and stream velocity. Other investigators have reported
similar sources of contamination.
Contaminated soils yield erratic leaching concentrations
of indicator bacteria. This variance depends on amount and
duration of rainfall. A 90% reduction for FC was reported
to have been attained in 3.3 days in summer and 13.4 days in
the fall (26) . For the fecal streptococci the same report
indicated a 90% reduction in numbers in a time of 2.7 days
in the summer and 20.1 days in winter. Heavier and prolonged
rains yielded greater numbers of isolates than short duration
rainfall events. Longer survival periods have been reported
by Mailman and Litsky (27). Their investigation involved
soils with elevated organic contents. At up to eleven weeks
duration the log number/100 grams of soil ranged from 5.6 at
time zero to 3.2 at eleven weeks. Livestock were observed
grazing in some undeveloped areas of the primary test site
periodically throughout this investigation. Therefore it is
entirely possible that considerable numbers of indicator
bacteria and pathogens were contributed to the water courses
from those sources. Soils and stream sedimentary materials
were analyzed for their bacterial content in order to deter-
mine whether contributions to overlying waters during periods
of heavy rainfall or elevated stream volumes were made from
those sources. Von Donsel and Geldreich (28) demonstrated
that distinct relationships exist between pollution and FC
and FS in the top two inches of mud. FC were found to have
been 100-1,000 times as concentrated in the mud as in the
water. A 19% probability for isolating SA in muds existed
when the FC were at concentrations of 1-200/100 ml in the
water. The probability increased to 80% for SA in the mud
when the FC were at levels exceeding 2,000/100 ml in the
water. Human population density also influences runoff
coliform densities. Soderlund and Lehtinen (29) found that at
a density of 35 persons/ha, the stormwater runoff contained
3.7 x IQll/lOO ml and at 100 persons/ha the coliforms were
9.2 x 1011/100 mi.
29
-------
The effect of suspended sediment as a major source of
bacterial contamination during such periods was determined
by sampling the stream sediment and soils in the vicinity
of the established sampling stations during low flow periods.
The results of those analyses are summarized below. Units
are number/gram of sediment or soil.
Stream Lakes A&B
Sediment Soils Sediment
Total Coliform 100-31xl04 200-3,500 1,000-25,000
Fecal Coliform <20-20 <20-40 20-60
Fecal Streptococci <20-20 <20-340 45-270
Pseudomonas aeruqinosa <20-20 <20-40 100-500
Staphvlococcus sp. 20-240 160-1,000 15-2,600
Salmonella sp. <25 <25 <25
Soil samples from the golf course area, swales, wooded
areas, and near roads exhibited the following bacterial con-
tents. Units are number/gram.
Ranges
Total Coliform 1,000-543,000
Fecal Coliform <20-280
Fecal Streptococci <20-1,000
Pseudomonas aeruginosa <20-1,000
Staphylococcus sp. 40-1,040
Salmonella sp. <25
Since the samples were obtained from the first two inches
(5 cm) of soil or sediment cover, leaching and scour velo-
cities obviously could have contributed to elevated bacterial
concentrations during storm events.
LOW FLOW MONITORING
The principal value of the data derived from monitoring
low flows is for comparison purposes with either urban stream
quality or storm event data developed from the primary test
site. More specific results for those low flow data are in-
cluded in statistical analysis results which are included at
the end of this section.
Concentrations of indicator bacteria and pathogens
ranged far beyond the diurnal low flow variation values and
single sample variations during the period of survey. The
ranges are summarized below. Units are number/100 ml.
Range
Total Coliform 200-lOxlO6
Fecal Coliform <10-214,000
30
-------
Range
Fecal Streptococci <10-1,580
Pseudomonas aeruginosa <10-53,600
Staphylococcus sp. 1-23,400
Salmonella sp. 1-5,800
The maximum values are comparatively high for a rural
forested area receiving no direct sewage discharges.
Figure 3 presents some of the earlier data representing
the relationships between TC vs FC and FS for values obtained
from analyzing station P-20 waters. Statistically, the slopes
of each are not that dissimilar. Visual analyses of those
least square data fit lines indicate a comparatively low rate
of either FC or FS with corresponding increases in TC.
Comparisons were made of low flow FC data with existing
contact and non-contact recreational standards. Recalling
that the geometric mean values not to be exceeded for contact
recreation are 200/100 ml (log^o = 2.3) and for non-contact
recreation are 2,000/100 ml (logio = 3.3) on the basis of 5
samples taken in 30 days, the following relationships are
present in the low flow data. At station P-10 the mean value
for all low flow data was 2.13 (log^o basis). That implies
that, on the average, no values for FC exceeded the contact
recreation standard. For stations P-30, Lake A and Lake B,
the values can best be discussed by the following summary.
Station P-30 Lake B Lake A
Geometric Mean 2.38 1.98 1.53
Waters at station P-30, therefore, were unacceptable for con-
tact recreation whereas the lakes seemed to effect reduction
in FC values.
CHARACTERISTICS OF IMPOUNDMENTS IN THE PRIMARY TEST SITE
Field measurements for dissolved oxygen, pH, temperature
and light penetration were made in Lakes A and B when top
(0.3m) and bottom water samples were taken on five dates:
1/11/75; 10/19/75; 12/15/75; 1/24/76; and 2/15/76. Sampling
points are represented in Figure 2. The numbers of sample
sites varied, but was never fewer than four in Lake B and
six in Lake A.
Physically/ both impoundments remained highly turbid
during periods of low flow. Runoff from swale 8, entering
Lake B, thence to Lake A always contained colloidal materials
which tended not to settle. Early in the survey of the lakes
it became obvious that both impoundments were operating
31
-------
LO
K)
o
o
^1 -H
o o
i»-l U
•H 0
rH U
O O
U -l-i
04
rH 03
a M
O 4J
Q) W)
en nj
o o
ij (P
•a
c
FC
FS
FC
FS
Figure 3.
Log Total Coliform/100 ml
Total coliform vs fecal coliform and fecal streptococci relationships
during low flow periods. Station P-20.
-------
hydrologically as sedimentation basins. This can be justified
by comparing most of the data for the upper reaches of Lake B
with the data for Lake A (Table 3.) Turbid conditions can
easily be verified by the quoted Secchi Disc values; those
generally were less than a foot. Dissolved oxygen concen-
trations were always found to have been more than adequate
to allow survival of the fish species which were stocked in
the impoundments. pH values usually ranged from 7.8 to 8.4.
A peculiar bacteriological characteristic of both impound-
ments was the relatively low densities of FC, FS, and all
pathogens whereas the TC values were high. That pattern of
behavior implies that TC enumeration, used as a water quali-
ty evaluation tool, is almost meaningless if considered
alone. Although the FC mean log values were 1.98 and 1.53
during low flow conditions for Lakes B and A, respectively,
for FC, another view of the data on a direct mathematical
basis merits other attention. Thirty-nine of 104 FC data
or 37.5% exceeded the 200/100 ml standard.
STORMWATER DISINFECTION REQUIREMENTS
Disinfection of combined sewers and stormwater has been
used but may not be, in itself, the final answer to pollution
control. Glover and Herbert (30) found that 5 mg/1 chlorine
additions in high rate chambers reduced "coliforms" by four
or more orders of magnitude. Initial FC and FC levels re-
ported were 0.1 x 10° to 3 x 106/100 ml and 1 x 103 to 1 x
105/100 ml respectively. Final FC levels were 5-10/100 ml.
Some regrowth may occur in those instances if complete die-
off is not attained. Reports from 'other areas of the coun-
try (31, 32) showed combined sewer overflows to have con-
tained indicator bacteria at the following levels: TC, 11
x 106/100 ml and FC, 1.6 x 106/100 ral-3 FC levels in com-
bined sewer flows ranged from 7.5 x 10 to 1.6 x 107 in dry
weather and 1.2 x 101 to 2.0 x 107 in wet weather (32). It
must be remembered though that the character of combined
sewers and stormwater runoff alone is not necessarily the
same.
Disinfection of stormwater is feasible and it is gener-
ally agreed that large dosages will be required to achieve
adequate reduction in indicator and pathogen densities.
Chlorine or chlorine dioxide has been reported to be the
most effective disinfectant, and in many cases the least
expensive (33, 34, 35). Davis, et al. (36, 37) discussed
current disinfection research and practices along with en-
countered problems which occur in combined sewer disinfec-
tion and stormwater disinfection.
Samples were obtained from different locations during
low fl'ow periods and during storm events to determine the
disinfectant demand and the disinfection effectiveness.
33
-------
TABLE 3. PHYSICAL, CHEMICAL, AND BACTERIAL RANGES IN LAKES
A AND B. VALUES ARE FROM SURVEYS OF 1/11/75, 10/19/75,
12/15/75, 1/24/75, and 2/15/76.
Lake B
Diss. Oxygen, mg/1
PH
Temp., °C
Secchi Disc, inches
Total Coliform,
No./lOO ml
Fecal Coliform,
No./lOO ml
Fecal Streptococci,
No./lOO ml
Pseudomonas
aeruginosa,
No./lOO ml
Staphylococcus sp.,
No./lOO ml
Salmonella sp.,
No./lOO ml
Diss. Oxygen, mg/1
PH
Temp., °C
Secchi Disc, inches
Total Coliform,
No./lOO ml
Fecal Coliforra,
No./lOO ml
Fecal Streptococci,
No./lOO ml
Pseudomonas
aeruginosa,
No./lOO ml
Staphylococcus, sp.
No./lOO ml
Salmonella sp.,
No./lOO ml
Surface
8.0-8.5
7.8-8.4
13.5-22.5
3-14
Ixl03-1.09xl06
60-97xl03
100-2.85xl03
<10-9.0xl04
<10-600
25-100
8.0-8.5
7.7-8.3
13.0-23.8
3-12
2xl03-1.4xlO!
80-820
10-1.35xl03
<20-1,980
300-700
<10-20
Bottom
7.6-8.0
7.6-8.0
13.0-23.8
2xl03-2.34xl05
90-1.09xl03
100-1.7xl03
<10-5.5xlOJ
400-1.6xl03
25-500
Lake A
6.3-7.0
7.5-8.2
13.5-21.5
3xlO'3-2.9xlO-
160-400
100-2.47xl03
<10-1,930
100-800
<10-30
34
-------
Locations from which the water samples were taken and other
pertinent data are summarized below.
Date
12/6/73
1/15/74
4/22/74(storm)
12/5-9/74(storm)
3/12-14/75(storm)
(Multiple Times)
3/13/75
4/7-10/75(storm)
(Multiple Times)
4/7/75
5/8-9/75(storm)
(Multiple Times)
8/14/75
11/26/75(storm)
(Multiple Times)
Disinfectant
C12,
C12
Br-
0-
C12, 03
C12, 03
C12, 03
C12, 03
ci2/ o3
C12, 03
C12, 03
Station(s)
P-20
P-30, S-10
(and C12 demand)
P-30
P-30
Lake A
Lake A (and C12
demand)
Lake A
Lake A (and C12
demand)
Westbury
Woodlands (Sew,
Treat. Plant)
Westbury
Those samples termed "multiple times" in the above list
signify that disinfection was applied to separate samples
taken at different times throughout storm events. The purpose
for that approach was to determined whether the qualities of
water discharged throughout a storm event required different
amounts of disinfectant to achieve the same end product.
Examination of those data revealed that little, if in fact
any difference in the disinfectant requirement between times
in storms occurred. The reason for that was the large dis-
infectant demand of the storm waters. That demand was due
for the most part to high solids (and turbidity) values;
values exceeding 200 mg/1 solids. Data from disinfection
experiments were averaged for those dosages and points in
time which were common to each set. They are presented in
graphical summary form following these commentaries.
Chlorine demand varied from station to station. The data
indicated the waters at station S-10 had a demand of about
8 mg/1 and those at station P-30 had a demand of between 8 and
16 mg/1. The chlorine demand of Lake A waters on 3/13/75 was
10 mg/1 and the ozone demand was in excess of 32 mg/1. Other
data indicated the same demand for Lake A outflow waters from
4/7/75 and for 3/13/75 samples. Demands such as those may
result in the disinfection of storm waters being an econo-
mically unwise and unfeasible venture except in specific
instances where impairment of the public health is foreseen.
35
-------
It therefore becomes a matter of choice on a location and
need basis.
Summary disinfection data presented in Figures 4 through
7 relate basically six complete disinfection sequences with
chlorine. TC concentrations tended to increase from 5.0
(log value) to 5.1 in the control over the eight day test
period. FC and FS concentrations decreased from 2.4 to 1.2
(FC) and 3.6 to 1.8 (FS) in 8 days. Staphylococci proved to
be the longer lived of the three pathogens which were enumer-
ated; decreasing from 2.9 (log value) to 2.2 in eight days.
PS and SA demonstrated viability at eight days but were
present at time zero in relatively low concentrations.
Pathogens at log No./lOO ml values of 1.0 were at the lower
detection limit. Some data for those groups showed values
of <1 x 101.
With chlorination at 2, 4, 8, and 16 mg/1 dosages the
TC all responded similarly, i.e., a decrease at 30-minutes
followed by regrowth of from about 2 (log value) to about 7-
All chlorine dosages effectively reduced the FC and FS
populations to about 10/100 ml within one day. They remained
at that level for 8 days. ST showed initial decreases with
all chlorine dosages to minimum values at one day, then re-
growth to about 4 days, followed by dieoff similar to the
controls. PS and SA demonstrated almost no propensity for
aftergrowth following all chlorine doses. Figures for SA
values are not included. They were at or below detection
concentrations.
Indicator and pathogen figures are not included for
ozone due to the fact that they would appear to be near
duplication of the chlorine dosages. For all practical
purposes the behavior of the bacteria in the presence of
ozone was similar in all but one case. That exception was
some regrowth of FS from 4 to 8 days following 8 mg/1 03
dosage.
TC populations which reestablished themselves to higher
concentrations than the control at 8 days, did so, perhaps,
due to decreases in competitive populations of other bacteria,
FC and FS were unable to compete at such low population
densities and therefore did not demonstrate aftergrowth
potentials. Staphylococci on the other hand demonstrated
the ability to increase in numbers up to 4 days time. That
pattern was followed by normal dieoff in the reestablished
populations. Graphs were not constructed for PS and SA
because of their relatively simplistic responses. FS, for
example, decreased to about 10/100 ml at all chlorine doses
and remained at that level, along with the control population
to 8 days. SA dropped to or below detection levels (1/100 ml
36
-------
U)
-J
30 min. Id
Figure 4
2d 4d Time 8d
Total coliform responses to chlorine; Lake A (n=6).
-------
00
§ 4
o 3
r-4
tn
O
Control
2 , 4 , 8, and 16 mg/1
30 min . Id
Figure 5
2d 4d Time 8d
Fecal coliform responses to chlorine; Lake A (n=6).
-------
vo
o
o
o
z
CP
0
4 . 8, and 16 mg/1
30min. Id
Figure 6.
2d 4d Time
Fecal streptococci responses to chlorine; Lake
8d
(n=6)
-------
o
O
o
a
30min. Id
Figure 7.
4 mg/1
2d 4d Time 8d
Staphylococci responses to chlorine; Lake A (n=6).
-------
or less) almost immediately following chlorination and re-
mained at those densities throughout the 8-day test period.
STORM EVENT MONITORING
Twenty-one storm events were monitored for microbiolo-
gical content. The dates and locations from which samples
were obtained are summarized below.
Date
1/19-20/74
3/20-21/74
3/27-28/74
4/11-12/74
4/22-23/74
4/22-23/74
10/28-31/74
12/5-9/74
3/4/75
3/12-14/75
4/7-10/75
5/8-9/75
5/8-9/75
6/30-7/1/75
6/30-7/1/75
9/5-6/75
10/25-27/75
11/26/75
3/7-8/76
3/8-9/76
4/5/76
Location
Woodlands
Hunting Bayou
Hunting Bayou
Hunting Bayou
Woodlands
Woodlands
Summary
Woodlands
Woodlands
Woodlands
Woodlands
Woodlands
Hunting Bayou
Westbury
Hunting Bayou
Westbury
Woodlands
Woodlands
Westbury
Woodlands
Woodlands
Woodlands
Stations
Monitored
P-30
Single
Single
Single
P-30
P-30
P-30
P-10, P-30
P-30, Lakes
P-10, P-30,
P-10, P-30,
Single
Single
Single
Single
P-10, P-30,
P-10, P-30,
Single
P-10, P-30,
P-30, Lakes
P-10, P-30,
A&B
Lakes A&B
Lakes A&B
Lakes A&B
Lakes A&B
Lakes A&B
A&B
Lakes A&B
In examining the data representing the storm events
monitored at the Hunting Bayou site it was discovered that
the TC values were not dissimilar from values reported from
other test sites. The TC range of all reported values for
Hunting Bayou water samples was from 3.1 x 10^ to 7 x 108.
That upper limit was an order of magnitude higher than most
values from all other locations. Those data are included in
the statistical evaluation of different land use areas which
is discussed in a following subsection. Data obtained from
Hunting Bayou are summarized below.
Storm
3/20/74
3/27/74
5/8/75
6/30/75
TC, log No./lOO ml
Min. Max.
4.49
4.86
7.00
4.78
5.48
6.45
8.84
6.70
Time at
Max., Hrs.
1.75
9.5
ND
ND
No. of
Samples
13
67
27
35
41
-------
Peaking times were not discernable for the latter two storm
events because of high variances in the water samples. For
the most part, the TC concentrations were at high levels and
remained at those levels throughout the period of time the
waters were monitored.
Westbury site waters demonstrated similar trends in TC
concentrations as well as within the other groups and species,
Data representing the three storm events monitored at that
site are as follows (TNTC = Too Numerous To Count):
Storm
5/75
log No./lOO ml
Min. Max. Mean
6/75
(n=12)
11/75
(n=5)
TC
FC
FS
PS
ST
SA
TC
FC
FS
PS
ST
SA
TC
FC
FS
PS
ST
SA
4.70
3.86
2.60
2.30
3.07
4'.47
2.93
2.83
2.97
3.43
<1.40
6.26
3.96
4.39
4.31
TNTC
2.00
8.70
4.60
4.61
4.40
4.30
1.51
8.70
4.60
4.61
4.40
4.30
1.51
6.67
4.50
4.68
TNTC
TNTC
4.17
7.48
4.34
4.12
3.88
3.91
1.51
7.48
4.34
4.12
3.88
3.91
1.51
6.49
4.28
4.51
4.54
3.19
Points in time at which peaking occurred for TC, FC, and FS
were impossible to determine because the concentrations tended
to demonstrate continual increases throughout the time of
monitoring each storm event. In the cases of PS, ST, and SA,
extremely high variances occurred between samples. Generally,
most of the groups or species measured were at or above those
minimum base level concentrations and maximum (peak) concen-
trations for Woodlands storm events.
Woodlands sites demonstrated more constancy in peaking
trends for both the bacterial concentrations and physical
factors. This can be observed in the data presented in Tables
4 and 5. Pathogens are not included for the 1/18/74 and
4/22/74 storms at station P-30 because the identities of those
which needed the greatest amount of attention (PS, ST, SA)
were not determined at that time. Times at which bacterial
peaking occurred are indicated in Table 4 for comparison with
hydrograph peaks, and solids and turbidity peaks in Table 5.
The notation "ND" signifies discernable peaks were not detec-
table from the available data. That happened most often
42
-------
TABLE 4. SUMMARY OF INDICATOR AND PATHOGEN BASE LEVELS, PEAK LEVELS AND TIMES DURING
STORM EVENTS.
Station
1/18/74
P-30
4/22/74
P-30
10/28/74
P-30
Indicator
TC
PC
FS
PS
ST
SA
TC
FC
FS
PS
ST
SA
TC
FC
FS
PS
ST
SA
Log of No./lOO ml
Min. Max. Mean
3.95
2.75
2.45
--•_
—
—
6.00
2.20
1.95
2.78
—
—
4.78
2.00
2.00
2.30
2.00
1.70
5.37
4.32
4.30
—
—
—
7.76
4.13
4.33
4.42
—
—
6.90
4.65
5.15
4.45
5.52
3.40
4.92
3.63
3.61
—
—
—
7.15
3.89
3.73
—
—
—
5.98
3.63
3.73
3.46
3.77
—
Time @
Max.,
Hrs.
5
8
8
-
-
-
13
8
14
8
-
-
6
9
5
3
10
4
Station Indicator
12/5/74 TC
P-30 FC
FS
PS
ST
SA
12/5/74 TC
P-10 FC
FS
PS
ST
SA
3/4/75 TC
P-30 FC
FS
PS
ST
SA
Log of No./lOO ml
Min. Max. Mean
4.48
2.60
2.00
2.00
2.90
1.52
4.00
2.00
2.00
1.00
<1.00
<1.39
4.68
2.58
2.85
1.48
2.99
<1.40
6.02
4.28
4.24
3.00
3.80
2.42
6.32
4.06
3.46
2.77
3.73
2.52
5.61
3.32
4.03
2.00
4.16
1.40
5
3
3
2
3
1
5
3
2
2
2
1
5
3
3
1
3
<*•
.38
.86
.48
.87
.45
.94
.13
.20
.65
.15
.77
.88
.17
.00
.37
.80
.40
.40
Tine @
Max.,
Hrs.
3
3
3
3
4
2
5
18
15
18
12
20
-
-
-
-
-
-------
TABLE 4 . SUMMARY OF INDICMDR AND PATHOGEN BASE LEVELS, PEAK LEVELS AND TIMES DURING STOBM
(GONT'D.) EVENTS.
Station Indicator
3/4/75 TC
Lake B FC
FS
PS
ST
SA
*. 3/4/75 TC
it-
Lake A FC
FS
PS
ST
SA
3/12/75 TC
P-30 FC
FS
PS
ST
SA
Log of No./lOO ml
Min. Max. Mean
5.36
1.60
3.46
2.67
3.67
<1.40
3.85
<1.00
2.60
<1.00
1.00
<1.40
4.78
1.00
1.00
<1.00
2.38
<1.40
6.
3.
4.
3.
4.
<1.
4.
1.
3.
2.
2.
<1.
6.
3.
3.
1.
3.
2.
03
90
38
83
45
40
70
30
72
43
74
40
72
65
92
85
89
00
5.58
3.38
4.21
3.39
4.12
<1.40
4.52
.85
3.16
1.52
2.14
<1.40
5.99
2.41
2.49
1.58
2.98
1.74
Time @
Max.,
Hrs.
1.5
1.5
4
3
1
19
19
19
19
22
22
Station Indicator
3/12/75 TC
P-10 FC
FS
PS
ST
SA
3/12/75 TC
Lake B FC
FS
PS
ST
SA
3/12/75 TC
Lake A FC
FS
PS
ST
SA
Log of No./lOO ml
Min. Max. Mean
4.30
<1.00
<1.00
<1.00
1.00
<1.40
4.60
<2.00
<2.00
<2.00
2.85
<1.40
4.00
<1.00
1.88
<1.00
1.00
<1.40
6.57
3.44
3.39
2.69
3.40
2.10
6.75
3.46
3.93
3.58
5.23
2.24
5.49
2.15
3.53
1.95
1.95
2.40
5
2
2
1
2
1
6
3
3
2
4
1
4
1
3
1
1
.93
.93
.87
.55
.52
.53
.18
.12
.53
.73
,04
.25
.91
.44
.12
.47
.50
.47
Time @
Max. ,
Hrs.
18
19
21
ND
20
ND
15
18
15
18
18
20
20
21
20
ND
17
19
-------
TABLE 4 . SUMMARY OF INDICATOR AND PATHOGEN BASE LEVELS, PEAK LEVELS AND TIMES DURING STORM
(OONT'D.) EVENTS.
tn
Station
4/7/75
P-30
4/7/75
P-10
4/7/75
Lake B
Indicator
TC
FC
FS
PS
ST
SA
TC
FC
FS
PS
ST
SA
TC
FC
FS
PS
ST
SA
Log of No./lOO ml
Min. Max. Mean
3.77
2.00
2.00
<2.00
2.30
<1.51
4.32
<2.00
2.00
<2.00
2.30
<1.51
4.00
2.00
2.00
2.00
2.48
<1.51
6.80
4.40
4.42
3.51
3.85
2.12
6.79
4.40
4.24
3.26
3.45
2.36
5.80
4.39
4.40
3.57
3.83
2.51
5.48
3.38
3.47
2.67
3.26
1.72
5.18
3.36
3.05
2.35
2.88
1.89
5.25
3.37
3.23
2.81
3.17
.28
Time @
Max. ,
Hrs.
21
14
16
ND
13
ND
20
20
18
21
19
23
21
16
14
14
13
ND
Station Indicator
4/7/75 TC
Lake A FC
FS
PS
ST
SA
9/5/75 TC
P-30 FC
FS
PS
ST
SA
9/5/75 TC
P-10 FC
FS
PS
ST
SA
Log of No./lOO ml
Min. Max. Mean
4.00
2.00
2.00
<2.00
2.30
<1.51
6.44
3.00
4.59
3.00
3.70
<1.00
3.90
<1.00
1.60
<1.00
1.00
<1.00
5.62
3.60
4.06
3.36
3.36
2.00
6.70
3.60
5.47
4.46
4.43
<1.00
5.18
2.78
2.63
1.90
2.23
1.00
5.06
3.03
3.50
2.61
3.29
.35
6.58
3.39
5.12
3.72
4.10
<1.00
4.30
2.41
2.19
1.60
1.69
1.00
Time @
Max. ,
Hrs.
ND
14
14
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
-------
TABLE 4. SUMMARY OF INDICATOR AND PATHOGEN BASE LEVELS, PEAK LEVELS AND TIMES DURING STORM
(CONT'D.) EVENTS.
Station
9/5/75
Lake B
9/5/75
Lake A
10/25/75
P-30
Indicator
TC
FC
FS
PS
ST
SA
TC
FC
FS
PS
ST
SA
TC
FC
FS
PS
ST
SA
Log of No./lOO ml
Min. Max. Mean
6.34
3.48
4.41
4.11
3.78
<1.00 <
3.78
2.00
2.04
1.46
1.50
<1.00 <
4.00
3.17
3.14
3.11
2.77
<1.40
6,
4.
4.
5.
5.
^ •
4.
2.
2.
3.
2.
6.
4.
4.
4.
5.
2.
86
26
99
12
03
00
97
85
88
40
82
00
04
62
94
77
05
17
6.55
3.90
4.74
4.66
4.42
<1.00
4.28
2.36
2.52
2.62
2.05
<1.00
5.52
4.17
4.26
4.34
3.72
1.62
Time @
Max.,
Hrs.
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
6
9
7.5
7.5
7.5
12.5
Station Indicator
10/25/75 TC
P-10 FC
FS
PS
ST
SA
10/25/75 TC
Lake B FC
FS
PS
ST
SA
10/25/75 TC
Lake A FC
FS
PS
ST
SA
Log of No./lOO ml
Min. Max. Mean
5.20
2.60
2.20
1.47
2.04
<1.40
4.90
3.20
3.30
2.60
2.47
<1.40
3.84
1.00
1.69
1.00
1.00
<1.40
6.02
4.81
4.39
3.97
3.43
2.17
6.87
4.44
4.85
4.63
5.02
2.30
4.99
3.00
3.21
3.66
3.06
2.30
5.41
3.48
3.36
2.99
2.68
1.54
6.01
4.22
4.16
3.62
4.03
1.40
4.20
2.41
2.88
2.04
2.29
.62
Time @
Max. ,
Hrs.
4
4.5
3.5
3.5
3
2
1.5
2
1
10
1
34
1
13
10
16
3
56
-------
TABLE 4 . SUMMARY OF INDICATOR AND PATHOQ2N BASE LEVELS, PEAK LEVELS AND TIMES DURING
(CCNT'D.) STOFM EVENTS.
Station
3/7/75
Lake B
3/7/76
Lake A
3/7/76
P-10
Indicator
TC
PC
FS
PS
ST
SA
TC
PC
FS
PS
ST
SA
TC
PC
FS
PS
ST
SA
Time
Log of Nb./lOO ml Max. ,
Min. Max. Mean Hrs.
5.30
3.04
3.23
2.30
3.70
<2.00
3.60
1.48
2.04
<1.00
2.32
<1.00
4.20
2.64
2.70
<1.00
2.45
<2.00
5 a? _
. y£ ~~ ~
4.80
3.79 —
3.04 —
4.41
2.70
4.04
1.95
2.34
1.70
3.28
<1.00
4.88 — —
3.77
3.40
2.45 —
3.15
2.30
Station Indicator
3/7/76 TC
P-30 FC
FS
PS
ST
SA
3/9/76 TC
Lake B FC
FS
PS
ST
SA
3/9/76 TC
Lake A FC
FS
PS
ST
SA
Time
Log of Nb./lOO ml Max. ,
Min. Max. Mean Hrs.
4.30 5.18 — —
2.78 3.53 —
3.18 3.72 — ~
3.15 3.41
2.36 3.04 — —
<1.00 <1.00 — —
4.36 5.11 — —
2.78 3.91
2.78 3.83
<1.00 1.95 — —
2.28 4.50
<2.00 <2.00
3.63 4.43
2.20 2.47
2.28 2.86
<1.00 1.70
2.70 3.28
<1.00 <1.00
-------
TABIJE 4. SUMMARY OF INDICATOR AND PATHOGEN BASE LEVELS, PEAK LEVELS AND TIMES DURING STORM
(CONTD.) EVENTS.
oo
Station
3/9/76
P-30
4/5/76
P-10
4/5/76
P-30
Indicator
TC
PC
FS
PS
ST
SA
TC
PC
FS
PS
ST
SA
TC
FC
FS
PS
ST
SA
Tine @
Log of No./lOO ml Max. ,
Min. Max. Mean Hrs.
4.08
3.20
3.56
1.00
3.95
<2.00
3.72
1.90
2.30
<2.00
1.48
<1.00
3.00
<2.00
<2.00
<2.00
2.00
<2.00
4.96
3.86
4.03 — —
2.11
4.48 — —
<2.00 —
4.26
2.50 — —
2.76
2.30
290 _ _—
. ^O ^^
1.00
5.77 —
4.00 — ~
4.59
3.00
4 ^ 59
2.30 ~
Station Indicator
4/5/76 TC
Lake B PC
FS
PS
ST
SA
4/5/76 TC
Lake A FC
FS
PS
ST
SA
Time @
Log of No./lOO ml Max.,
Min. Max. Mean Hrs.
4.81
2.90
2.49
<2.00
3.45
<2.00
3.76
1.48
2.69
1.00
2.00
<1.00
6.51 — —
3.78
3.11
2.70
4.45 —
2.00 — —
3.97 —
2.79
3.09
1.90
2.89 —
<1.00 —
-------
TABLE 5. SUMMARY OF PHYSICAL FACTOR PEAK VALUES AND TIMES DURING STORM EVENTS,
VD
Time to
Station
1/18/74
P-30
4/22/74
P-30
10/28/74
P-30
12/5/74
P-30
P-10
3/12/75
P-30
3/12/75
P-10
Lake B
Lake A
P-30
3/4/75
P-30
Lake B
Lake A
4/7/75
P-30
P-10
Lake B
Lake A
Q
19.5
8
3.6
5.3
3.6
20
19.5
26
21
20
0
1.5
-
40
32.5
14
23.2
Peak, Hrs. (Max.
(cfs)
(1
(
(
(
(
(
(
(
(
(
(
(
(
(1
(1
(
(
,250
9.
107
—
279
195
25
56.
11.
2.
25
4.
0.
0.
,060
,225
96
113
)
6)
)
)
)
)
8)
7)
0)
)
9)
3)
0)
)
)
)
)
ss
4.5
10
7
2
1.5
-
17
16
—
-
—
2
-
13
18
20
26
(mg/1)
(2,090)
(1,600)
(1,208)
( 667)
( 120)
( 222)
( 170)
(5,358)
( 164)
( 222)
( 62)
( 326)
( 194)
( 672)
( 226)
(2,660)
( 356)
Value)
T(J
4.5
11
8
1
1.5
-
17
16
-
-
-
2
-
20
18
13
26
.T.U.)
( 500)
( 500)
(1,460)
( 350)
( 70)
( 132)
( 85)
(2,000)
( 108)
( 132)
( 45)
( 490)
( 125)
( 285)
( 100)
( 900)
( 210)
Rain
Amt., in.
2
0
3
1
1
0
0
0
0
0
0
0
0
2
2
3
3
.02
.45
.46
.59
.52
.75
.69
.81
.81
.75
.33
.26
.26
.43
.43
.97
.97
No. of
Samples
46
49
58
17
57
5
32
20
27
5
8
13
11
26
34
38
38
-------
TABLE 5. SUMMARY OF PHYSICAL FACTOR PEAK VALUES
(CONT'D.)
Time to Peak, Hrs. (Max.
Station
Q
(cfs)
SS
(mg/1)
AND TIMES DURING STORM EVENTS.
Value)
T(J.T
.U.)
Rain
Arat. , in.
No. of
Samples
9/5/75
P-30
P-10
Lake
Lake
B
A
21
-
( 9,0)
( 0.3)
None
None
21
-
30
-
( 158)
( 9)
(2,766)
( 34)
29
-
30
-
(
(
(1,
(
162)
4)
440)
24)
0
0
1
1
.35
.25
.11
.11
4
12
8
16
10/25/75
P-30
P-10
Lake
Lake
B
A
17
20
15
(56 )
(22 )
(15 )
17
2
1
T-
( 398)
( 42)
(1,276)
( 78)
17
2
1
-
(
(
(
(
265)
28)
680)
28)
—
—
—
—
14
30
12
23
Ul
o
-------
when too few samples were obtained to portray the complete
storm, or the amount of flow was of such a small order of
magnitude as to render additional data evaluation meaningless.
That is also true for missing mean values in the data presented
in Table 4. Those data substantiate the belief that a first
flush effect occurs during storm events, regardless of the
magnitude and duration of the hydrograph. In the majority
of instances both indicator and pathogen populations peaked
in concentration prior to or very near the point in time
through the storm event at which the flow volume demonstrated
peaking. That was also true for turbidity (T) and suspended
solids (SS) concentrations. By visual inspection alone, there
appeared to have been little correlation of the amount of
rainfall and peak SS, T, or bacterial levels. Perhaps the
soil conditions prior to each storm event influenced the
runoff quality.
The storm event of 4/7/75 was chosen as that which
represented the most typical event which was monitored.
Figures 8 through 15 present some of the microbiological data
derived from stations P-10 and P-30 during the storm. Total
coliform concentrations demonstrated a less clearly defined
curvature when graphed than did the PC and FS concentrations.
All three of those groups peaked well in advance of the
hydrograph peak. FC and FS densities peaked earlier at
station P-30 than at the upstream station (P-10) (Figures
8-11). A similar trend can be observed in the data presented
in Figures 12 and 13 for staphylococci. Data for the other
pathogens (PS and SA) are not included in graphical format
due to their highly erratic concentrations versus time through
the storm. Attention should be given the FC/FS values for
both stations plotted in Figures 14 and 15. Those occurring
through the storm at P-10 were obviously extremely erratic
whereas virtually all of those at station P-30 held near or
below 1.0 throughout the storm. Data such as these obviate
the need for a statistical approach to analysis rather than
simple examination.
\
TEXAS WATER QUALITY STANDARDS AND WOODLANDS STORMWATER
BACTERIA CONTENT
Conditions existing at P-10, P-30 Lakes A and B during
low flow periods with regard to FC contact and noncontact
recreation standards have been discussed earlier in this
report. The Texas standards apply even more directly to
stormwaters because in virtually all sequences more than
five samples were obtained within the thirty day prescribed
time period. Recall the standard for contact recreation;
FC not to exceed 200/100 ml (log =2.3) as a geometric mean
and in less than 10% of the samples they should not exceed
400/100 ml. For noncontact recreation the figures are one
51
-------
7r-
cn
to
u
cr>
o
_L
_L
20
60
80
Figure 8.
40
Hours
FC densities during 4/7-10/75 storm event; the Woodlands station P-10
-------
ui
oo
CP
o
I
I
20
60
40
Hours
Figure 9. Fecal coliform densities during 4/7-10/75 storm event? the Woodlands station P-30.
80
-------
7r
Ul
• •
I
20
40
Hours
60
80
Figure 10. FS densities during 4/7-10/75 storm event; the Woodlands station P-10,
-------
7r
in
in
20
40
Hours
60
80
Figure 11. Fecal streptococci densities during 4/7-10/75 storm event; the Woodlands station P-30.
-------
EH
CO 4
in
2-
1 -
20
60
Figure 12.
40
Hours
Staphylococci densities during 4/7-10/75 storm event; the Woodlands station P-10.
80
-------
w
(Jl
-J
20
60
80
Figure 13.
40
Hours
Staphylococci relationships during 4/7-10/75 storm event; the Woodlands station P-30.
-------
en
oo
10
CO
Figure 14
20
40
Hours
60
80
FC/FS ratios during 4/7-10/75 storm event; the Woodlands
station P-10.
-------
U1
VD
ior
• •
\
_L
20
60
Figure 15.
40
Hours
FC/FS relationships during 4/7-10/75 storm event; the Woodlands
station P-30.
80
-------
order of magnitude higher, i.e., 2,000/100 ml and 4,000/100 ml
in less than 10% of the samples taken in a 30-day period.
Comparisons of storm event data can be easily accomplished
with the idea of estimating the sedimentation effect of the
lakes (referred to earlier) and distance downstream from P-10
to P-30. Table 6 summarizes the data for the storms and in-
dicates whether the contact and noncontact standards were
exceeded based on geometric mean FC values for each storm
indicated.
The only source of the four listed in Table 6 which did
not exceed contact recreation FC values was the Lake A effluent
during two of five events. All others exceeded that standard.
Thirteen of 24 storms, considering all stations, exceeded the
noncontact recreation standard of 2,000/100 ml.
MATHEMATICAL APPROACHES TO DETERMINATION OF BACTERIOLOGICAL
WATER QUALITY
The first phase of data analysis by computer analysis in-
cluded data obtained through March, 1975. Multiple regression
analyses were conducted using several sets of low flow and
storm event data. Those analyses employed a step-wise
regression technique which determined the relationship between
the bacteria count, set as the dependent variable, and other
bacteria counts and the other water quality factors set as
independent variables. Data which were used in those analyses
were from the following low flow monitoring and selected
storm events:
Low Flow - Inclusive of Stations B-09, P-09, P-10,
P-20, P-25, P-30, and P-40 for 9/7/74,
2/7/75, 2/18/75, and the diurnal low
flow of September 21-22, 1975 at
Stations P-10 and P-30.
Storm Events - 10/28-31/74 (P-30), 12/5-9/74 (P-10
and P-30), 3/4/75 (P-30), and 3/12-14/75
(P-10 and P-30).
231 data were employed for those analyses.
The cross correlation matrix developed from using
bacterial counts in their usual numerical form is presented
in the following table. Those data show the highest R2 values
to have been between fecal streptococci with discharge
(R2 = 0.75) and suspended solids with discharge (R2 = 0.71):
(T = turbidity, J.T.U.; Q = discharge, cfs; SS = suspended
solids, mg/1).
60
-------
TABLE -6. FECAL COLIFORM GEOMETRIC MEANS DURING STORM EVENTS.
Storm:
12/5/74
3/14/75
4/8/75
9/5/75
10/25/75
1/18/74
4/22/74
10/28/74
12/5/74
3/4/75
3/13/75
4/8/75
9/5/75
10/25/75
3/4/75
3/12/75
4/8/75
9/5/75
10/25/75
3/4/75
3/12/75
4/8/75
9/5/75
10/25/75
Geo.
Geo. Mean Exceeds
Std. For:
Mean Contact Rec. Noncontact Rec. Supp. Data
3.20
2.93
3.36
2.41
3.48
3.63
3.89
3.63
3.86
3.00
2.41
3.38
3.39
4.17
3.38
3.12
3.37
3.90
4.22
0.85
1.44
3.03
2.36
2.41
Station P-10
Yes
Yes
Yes
Yes
Yes
Station P-30
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Lake B
Yes
Yes
Yes
Yes
Yes
Lake A
No
No
Yes
Yes
Yes
No
No
Yes
No
Yes
Yes
Yes
Yes
Yes
No
No
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
No
No
No
No
No
15/58>4,000
1/5>4,000
1/21>4,000
— — —
5/39M,000
61
-------
ST .33471
SA .18274 .37348
TC .43297 .46562
FC .34707 .30922
FS .63224 .68690
SS .42237 .44566
Q .37058 .84900
T .15363 .52922
ST
20925
.17502 .37188
,24429 .56734
,10987 .30088
38854 .43463
.46259 .27847
SA
TC
,54247
,49152
.44541
,29076
FC
.57878
.75013 .70998
.35788 .03861
FS
SS
55439
Q
Plotted relationships of bacteria counts and SS, T, and Q
were, for the most part, nonlinear. When transformed to
semilog plots, the relationships were much more useful in data
manipulation. Log base-10 values for bacterial densities were
then rerun using the same computer program for multiple re-
gression analysis. SS, Q, and T values were not entered as
log values. The results are presented in the following summary
matrix tabular format.
ST
SA
TC
FC
FS
SS
Q
T
.48310
.45079
.33548
.48222
.53829
.50524
.52644
.38069
.32302
.53282
.59134
.76638
.59214
.73667
.40761
.27482
.29133 .47659
.40474 .55472
.32590 .34821
.40444 .41138
.36398 .32760
PS
ST
SA
TC
,81061
,50260
,46280
,43997
FC
,61729
.70600
,44268
FS
,70998
,03861
SS
,55439
Q
The highest R2 values shown in those results were FS and FC
(R2 = 0.81) and ST and FS (R2 = 0.76). Finalized models for
the period of the investigation for which data were analyzed
through May, 1975 are presented in Table 7. In reviewing
those models, several important factors become evident. For
example the FS was used as an independent variable in three
of the models and was, in fact, the independent variable
entered in the step-wise regression. That proved it to have
been the variable which explained the most variance in the
dependent variable when used above. Additionally, in the
FC and FS models, their coefficients were nearly equal, i.e.,
.719 vs .720. Their densities therefore were influenced by
runoff factors (discharge, turbidity, and solids) but with
opposing effects. FC increased with increases in turbidity
and solids and FS decreased. On the other hand FS increased
with increases in discharge and FC counts decreased. Those
models (Table 7) strongly suggest the FS group to have been
the most useful indicator group. Since the FC were not used
in the models for pathogenic bacteria, they cannot be con-
sidered to have been adequate indicators of fecal pollution
in the test site. It is important to note that due to the
high numbers of data used for those analyses, all R2 values
presented in Table 7 are statistically significant.
62
-------
TABLE 7. SUMMARY TABLE OF MODELING EQUATIONS
R2
Pseudomonas aeruginosa PS = .195 FS - 13.608 Q + .983 SS + 6.645 T - 3.669 .47
Pseudomonas aeruginosa(log) PS = .206 FS + .002 SS + .002 T + 1.063 .40
Staphylococcus aureus ST = 153.276 Q - 59.167 SS + .002 TC - 23.122 T - 1,344.787 .79
Staphylococcus aureus (log) ST = .363 FS + .002 Q + .183 TC + .439 .68
—=
Salmonella SA = .560 T + .826 Q + 79.476 .24
Salmonella (log) SA = .240 PS + .002 T + .001 SS + .594 .27
Total Coliforms TC = 39.122 FS + 543,719.80 .32
Total Coliforms(log) TC = .408 FS + 4.102 .31
Fecal Coliforms PC = .141 FS + 17.731 SS + 16.271 T - 15.487 Q + .494 .42
Fecal Coliforms (log) FC = .719 FS - .002 Q 4- .003 T + .002 SS + .612 .77
Fecal Streptococci FS = 58.010 Q + .003 TC + .704 FC - 19.975 T - 2,798.547 .67
Fecal Streptococci (log) FS = .720 FC + .002 Q + .151 TC - .001 T - .001 SS - .170 .82
-------
An additional analysis of variance was conducted on
storm event and low flow data for 12/5-9/74, 3/12-14/75,
4/7-10/75, 9/5-6/75, 10/25-27/75, the diurnal study of
9/21/74, and low flow data. The procedure assumed sampling
of normally distributed populations. Since P-30 and P-10
represented major categories, separate frequency diagrams
were constructed. The distributions of raw bacteriological
data were markedly skewed. The log transformed data elimi-
nated most of that skewness, however the diagrams appeared
to represent the two overlapping populations of low flow
and storm data. Therefore, storm and low flow frequency
diagrams were constructed separately for the transformed
data. This effectively reduced skewness and kurtosis on all
variables.
In all the bacteriological data presented herein the
geometric mean closely approximates the median value, proving
a normal distribution occurred in the populations, thus
validating the applied statistics. Differences between, for
example, upstream and downstream stations can be cited as
multiplicative differentials. This is important in denoting
urbanization effects and land use effects on runoff water
quality-
Separate Applications of Multiple Regression Analyses
Analyses similar to those described for the earlier work
were applied to all data from January, 1974 through June,
1975, plus a comparison of low flow and storm data was in-
dicated. Regression analyses were conducted with computer
control cards similar to those used in the earlier study
except that an option was used to affect paired deletion
of cases which contained missing-data values. Seventy-two
regression equations were generated using log-transformed
data. The usage comparisons were as follows:
Dependent Independent
Variable Variable
Log TC Q, SS, T, log FC, log FS
Log FC Q, SS, T, log TC, log FS
Log FS Q, SS, T, log TC, log FC
Log PS Q, SS, T, log TC, log FC, log FS
Log ST Q, SS, T, log TC, log FC, log FS,
log PS
Log SA Q, SS, T, log TC, log FC, log FS,
log PS
Independent variables were entered into the regression equa-
tions in a stepwise fashion by virtue of their partial cor-
relation to the dependent variable; the highest correlated
independent variable having been entered first, followed by
64
-------
the next independent variable that best improved the R2. The
first independent variable entered into the equations always
indicated the highest percentage of the explained variance
of the entire equation; the second variable entered would
improve the R2 by a lesser amount. The maximum change in R2
for the second independent variable entered in those equations
was 0.21. Most of the changes were considerably lower. A
comprehensive listing of all the regression equations would
be pointless. The variables entered into the equation, their
order, and the percent variance in the dependent variable
explained by the equations were altered with different
analytical treatments (sequences) of the data. Some devia-
tions were observed in the correlation matrices. The
examination of the results which follows explains some of
the anomalies mentioned above. The data were subdivided
into the following 11 individual sequences.
Sequence 1: All storm event and low flow data through June,
1975 (n = 669)--
Log FS produced the highest correlations, and was
apparently the best predictor of the other bacterial popula-
tions which were examined. For all the data, Log FS gave
the highest correlations with Log TC (R2 = .31), Log FC
(R2 = .45), and Log ST (R2 = .39). The FC/FS ratio yielded
low correlations many of which were insignificant at the
5% level. SS and T yielded correlations greater than
R2 = .75.
Sequence 2: All storm data through June, 1975 (n = 486)—
The storm event correlations tended to be higher and
more significant than the low flow data and all the combined
data. With the storm event data, Log ST gave the highest
correlations and Log FS the second highest. Their R2 ranges
were as follows: Log ST = .06 - .46; Log FS = .17 - .46.
The correlations involving the FC/FS ratios were low and
only one, FC/FS with Log SA (R2 = .02) was significant at the
5% level. i
Sequence 3: All low flow data through June, 1975 (n = 181)—
The low flow data exhibited 22 of 3-7 correlations that
were insignificant at the 5% level. The correlations in-
volving FC/FS ratios were low for all bacterial densities
and most were insignificant at the 5% level.
Sequence 4: Complete storm data—
Regressions were conducted with storm event data which
contained no missing values to determine if inclusion of
65
-------
cases with missing values affected the correlation co-
efficients. This necessarily decreased the number of cases
and the degrees of freedom. The resulting correlations
varied. The correlations of FC/FS ratios with other bacterial
densities were not significant at the 5% level.
Sequence 5: Lake B storm data through June, 1975 (n = 74)--
.For Lake B storm event data, the best predictor was
Log FS. It gave the highest correlations for Log ST
(R2 = .49). All correlations involving Log SA were not
significant at the 5% level. The FC/FS ratios yielded low
and/or negative correlations with other variables.
Sequence 6: Lake A storm data through June, 1975 (n = 79)—
Lake A storm event correlations were difficult to inter-
pret. Log PS gave the highest correlations with other
bacterial densities but all were below R2 = .50. Log FS
ranked third after Log FC. significant correlation involving
Log SA was with Log PS which was -.39 (R2 = .16). The FC/FS
ratios yielded some significant correlations at the 5% level.
Sequence 7: All P-30 storm data through June, 1975
(n = 208) —
This and sequence 10, below, yielded R2 values of greater
than- 0.75. The model which did that was
Log ST = .000668T + .779 Log FS + .000625 SS -
.489 Log FC + .167 Log TC + 1.051
and had an R2 = .79.
Sequence 8: P-10 storm data through June, 1975 (n = 125)—
For station P-10 storm events, the correlations involving
FC/FS ratios were not significant at the 5% level for all
bacterial densities except Log SA with R2 = .20. Log FS
yielded high correlations (R2 = .28) for Log TC, (R2 = .63)
for Log FC, and (R2 = .50) for Log ST. SS and T values
yielded R2 values greater than .75.
Sequence 9: Storm event of L2/5-9/74 (stations P-10 and
P-30) (n = 74); Sequence 10: Storm event of 3/12-14/75
(station P-10, P-30, Lakes A&B)(n = 97); and Sequence 11:
Storm event of 4/7-10/75 (station P-10, P-30, Lakes A&B)
(n = 14) —
For each storm event tested, the Log FS values appeared
to have been the best predictors of bacterial densities
measured. FC/FS ratios yielded low and/or not significant
correlations at the 5% level for most variables.
66
-------
The three storm events were chosen for analysis because
they were thought to represent the most typical hydrographs.
Q, SS, T, Log TC, Log FC and Log FS were always specified as
independent variables. The following table reinforces the
earlier conclusion that FS were the best predictors for
Woodlands bacterial water quality data.
A B C D
Q 1 7 66
SS 2 2 66
T 5 11 66
Log TC 10 5 58
Log FC 17 3 58
Log FS 28 4 58
A = Independent variable.
B = No. of times the independent variable was the
first variable entered into an equation.
C = No. of times the independent variable was the
second variable entered into an euqation.
D = No, of equations being examined.
The models developed in sequence 10, and summarized
below, yielded the highest R2 values.
Log PS = .000862T + .665 Log TC - .0233Q - 1.757
(R2 = .85)
Log ST = .259 Log TC + .000612T + .968 Log PS +
.0278Q - .209 Log FC - .589
(R2 = .90)
2
SS and T values in sequences 10 and 11 also yielded R values
greater than .75.
In summary, the FC/FS ratios were of little use in
predicting the presence of pathogens, based on correlations
derived for the 5% leve.l. Log FS appeared to have been the
most promising predictor, yielding higher correlations for
most bacterial densities measured and most data treatments.
However, Log TC gave some higher correlations for stations
P-10 and P-30 storm event data. P-10 and P-30 data rep-
resented approximately 50% of all the data. Salmonella sp.
densities were difficult to predict. Some correlations
were encountered but they were not consistent between the
data sequences.
Inspection of all the data developed through June, 1975
revealed that, 1) six samples did not contain Fecal Coliforms
when Salmonella sp. were detected; 2) seven samples did not
contain FecalStreptococci when Salmonella sp. were detected;
3) twenty-nine samples did not contain Pseudomonas aeruginosa
67
-------
when Salmonella sp. were detected; and 4) three samples did
not contain Staphylococcus aureus when Salmonella sp. were
detected. However, no Salmonella sp. were detected when the
Total Coliform densities were below 4,500/100 ml.
FC/FS RATIOS AND LAND USE
An important comparison employed in examining data is
the relationship between FC and FS concentrations. A FC/FS
ratio 4 suggests the presence of human wastes. Between
0.7 and 3.9. the FC/FS ratio may suggest human wastes mixed
with other source materials. If values fall at or below
0.6, a predominance of livestock or other warm-blooded ani-
mal waste is suggested. Absolute conclusions derived from
FC/FS values require a knowledge of sample storage history,
point source FC/FS values, and the length of time the or-
ganisms were in the water prior to sampling.
Changes in natural forest cover to either clear cut
areas for cultivation or for construction projects have been
shown to result in measurable changes in water quality and
hydrological factors (40). Leopold (39) cited distinctly
measurable changes in lag times to peaking with urbanization.
Claudon (38) stated that construction sites regularly con-
tribute Vibrio cholerae, Salmonella typhi, Salmonella para-
typhi, and Shigella dysenteriae to runoff volumes and that
agricultural and urban runoff regularly contributes to Sal-
monella sp. in recreational waters. Geldreich, e_t al. (41)
found the greatest median value for FS to have been 790,OOO/
100 ml and occurred in winter. Tafuri (42) reported FC le-
vels of 23,000/100 ml from a 1.67 sq. mi. (4.32 sq. Km)
drainage area. Snowmelt and related agricultural runoff in
far northern climes in the continental U.S. have been shown
to contribute high densities of indicator bacteria and path-
ogens to runoff (43, 44). However, many urban area res-
ponses exceed rural agricultural runoff pollutant loadings
(45). To identify a universal relationship for anticipated
indicator or pathogen content in rural stormwater runoff
would indeed constitute an exercise in futility- Each site
must be examined and appraised on its own merit. Correla-
tion efforts between runoff volumes and other water quality
parameters generally result in meaningfully high values.
But biological systems are a different matter entirely.
This is further emphasized by the research conducted earlier
by Winslow and Epsey (48). They developed good correlations
in all analyzed factors against runoff volume except the
indicator bacteria. Yet some predictive values for bacteria
are possible on a site by site basis (46, 47). Physical
factors such as suspended solids yield highest correlation
values.
68
-------
FC/FS ratios in water or wastewater of one, or less,
point toward nonhuman pollution sources. Woodlands and
Westbury sites demonstrated similar ratios for stormwater
runoff and statistically related differences were not
observed in three other areas considered herein. Those
areas and/or stations were Westbury, Lakes A and B, P-10,
and P-30. The FC/FS ratios appeared to be relatively
constant regardless of sample site. This can easily be
observed in the plotted data presented in Figure 16. That
scalar approach to analysis shows two important factors.
During storm events, the densities increased significantly
but ratios remained relatively constant. The data points in
Figure 16 represent mean FC and mean FS values. Increases
occurred in FC and FS values with increased urbanization;
Westbury being the site which was completely developed and
P-10 and (to some extent) P-30 as rural areas. For further
confirmation of these results, analysis of covariance was
used to test the equivalence of least square regression lines
for each of the five land use sites. The results of those
analyses are presented in Figure 17 and Table 8. The lines
representing P-10, P-30 and Lake B (influent) were analyzed
simultaneously- Westbury and Lake A effluent data were
analyzed separately due to their marked differences. FC and
FS distributions were not constant for different land uses.
Figure 17 shows the slopes decreasing with increases in
urbanization and elevations increasing with increased urbani-
zation, except for Lake A, where settling effects in the
lakes resulted in decreased densities in Lake A outfall.
Differences between densities of FC in incoming waters (Lake
B) and Lake A effluent usually were at least one order of
magnitude. FS densities decreased to a lesser extent.
Increased FC densities with increased urbanization is more
than strongly suggested.
The least Significant Difference (LSD) statistical
analysis was applied to FC/FS ratios from all stations which
were analyzed during this investigation. The following
sequence resulted from that analysis:
Location, FC/FS (Geo. Mean)
Soil samples
Storms, Lake A
Storms, Lake B
Low Flow, Lake A
Chlorinated 2°
Low Flow, P-10
Storms, P-30
Low Flow, Lake B
Low Flow, P-30
Storms, Westbury
Low Flow Woodlands
Storms, P-10
Raw Sewage
2° Treated Sewage
Sewage
0.16
0.17
0.53
0.58
0.73
0.92
0.97
0.99
1.26
1,
1.
47
68
2.11
2.42
13.30
69
-------
o
Storms
O P-30
D P-10
Q Lake B
V Lake A
WWestbury
* 2° Sew;
•+• Raw Sew.
FC/FS=1.0
FC/FS=.1
I
J
Figure 16.
234567
Log FS
Scalar approach to FC/FS patterns in different land use areas
-------
0 3
CP
O
Median
FC/FS
Values
12345
Log FS
Figure 17. Fecal streptococcus vs. fecal coliform
regression lines for five land use areas
TABLE 8. MEANS, SLOPES, ELEVATIONS, AND CORRELATION CO-
EFFICIENTS FOR FS ON FC REGRESSION LINES FOR LAND USE AREAS
SHOWN IN FIGURE 17.
Log FS Log FC
4.39
3.31
3.47
3.00
2. 34
b = slope
a = intercept
Westbury
Lake B
P-30
P-10
Lake A
.56
.54
.81
.77
.51
.32
.30
.66
.59
.26
.40
.48
.68
.85
.87
2.
1.
1.
.
— * •
76
75
10
68
53
28
89
264
191
126
4
3
3
2
3
. 22
.56
.47
.75
.30
71
-------
Those groups which demonstrated no significant differences
between subsets are indicated by the vertical lines on the
side.
STORMWATER RUNOFF CHARACTERISTICS AT P-10 VS P-30
Increased bacterial densities for urban stormwater
runoff observed for most bacterial variables by a one-way
analysis of variance. The primary interest was focused on
bacterial densities between stations P-10 and P-30. Marked
variations in bacterial densities existed between storms.
To reduce those variations log transformed data were analyzed
and the sites were paired by storm event which indicated the
need for use of a two-way analysis of variance. Results
indicated that this model actually fit only the SA data. No
differences in SA densities were detected between stations
P-10 and P-30. The F-Statistic applied for the effects of
station locations for the other variables were high, so that
a one-way classification was used to determine differences
on an individual storm basis.
The results indicated that no differences in TC densi-
ties were observed between stations P-10 and P-30, but on
the average, the following relationships existed: 1) the FC
densities were 2.7X higher at P-30 than at P-10; 2) the FS
densities were 5.2X higher at P-30 than at P-10; 3) the PS
densities were 5.7X higher at P-30 than at P-10; and 4)
the ST densities were 5.OX higher at P-30 than at P-10.
ALTERNATIVE ANALYSES
Means and 95% confidence intervals were calculated for
all available low flow and storm event data derived from
stations P-10 and P-30. Figures 18 and 19 include those
data. Fewer monitoring dates were recorded at P-10 than at
P-30. Nevertheless enough data were obtained to permit this
type- of visual comparison. Trends in the concentrations of
bacteria can easily be recognized. At P-10 the concentrations
of FC, FS, and PS all followed the general concentration
trend for TC and SS but, of course, for different orders of
magnitude. ST and SA curves are not included but closely
resembled the curve representing FC concentrations. Only a
cursory examination of the data representing P-30 (Figure 19)
is required to verify earlier stated conclusions regarding
the water quality at that station compared to P-10.
Bacteriological trends were not the same during storm events.
For one, the values were generally higher at P-30 than P-10.
And ST and SA values were considerably more erratic at P-30.
As to the causes for such concentration trends, some
implication may be derived from the meteorological data
72
-------
en
CM
CO
a.
en
o
4
3
2
1
7r-
OL
4
1 _
5-
_ ,_• MI» "••" j—
log TC
log FC
_I
-2.2
1
I
I
LF 1/74 4/74 10/74 12/743/4/75 3/13/75 4/75 9/75 10/75
n=33 56 32 34 12 30
Figure 18. Indicators and pathogen relationships; low-flow
and storm events. Station P-10, The Woodlands.
73
-------
L
_L
_L
_L
J_
_L
LF 1/74 4/74 10/74 12/74 3/4/75 3/13/754/75 9/75 10/75
n=39 45 49 56 16 8 5 26 4 14
Figure 19. Indicators and pathogen relationships; low-flow
and storm events. Station P-30, the Woodlands.
74
-------
presented in Figure 20. A steady drying cycle can be noted
in the prior monthly rainfall data. That did not follow the
bacterial concentration trends. Prior week rainfall also
shows no correlatable trend to bacterial densities. However,
the rainfall amount and duration tended to show similarity
with bacterial concentration trends at P-10. The data also
refute the theory that a short term storm event of greater
intensity will result in a greater leaching effect with
respect to bacterial concentrations in the runoff. Note in
Figure 20, with the exception of discharge (cfs), the com-
paratively close relationships of rainfall data between
stations P-10 and P-30.
Indicator bacteria and PS mean values from 16 sites or
sources were calculated and plotted in Figure 21 along with
their 95% confidence intervals. The abbreviations used
on the Y-axis are; raw sew. (raw sewage), Sew 2°
(secondary treated sewage), Cl2 Effl. (chlorinated secondary
effluent), Sedmt. (stream sediment), LF Wood, (low flow, the
Woodlands), LF LA (low flow, Lake A), LF P-10 (low flow,
station P-10) , LF P-30 (low flow, station P-30), LF LB (low
flow, Lake B), S LA (storms, Lake A), S P-10 (storms, station
P-10), S P-30 (storms, station P-30), S LB (storms, Lake B)
S West, (storms, Westbury site), S Hunt, (storms, Hunting
Bayou).
Those data reveal some unanticipated similarities in
bacterial concentrations. For example, TC concentrations were
higher in 12 sources and sites than TC values for chlorinated
secondary treated municipal wastewater, excluding raw and
secondary treated wastewater. Within the 95% confidence
intervals many of the FC and FS levels were similar in orders
of magnitude; again excluding raw and secondary treated
wastewater. SA and ST trends appeared the same as that curve
presented for PS. It is significant to note that the trend
in the data in Figure 21 is for increased indicator and
pathogen concentrations with increased urbanization. Also,
the TC values indicated for Westbury site and Hunting Bayou
were equal or greater ithan those for secondary treated waste-
water .
All of the available low flow data for the Woodlands
were analyzed by season of the year to determine whether
temperature affected population densities. The results of
those LSD and F-statistic calculations are presented in
Figure 22. Statistically, no differences occurred between
seasons for FC and FS. The notation at the side of the TC
data indicates that the summer subset was statistically
greater in itself than either the winter or spring data
subset. But the winter and spring subsets were statistically
equivalent. Within the pathogens, ST provided the least
variation between seasons. FC/FS ratios appeared to have
been higher in summer and fall than spring or winter.
75
-------
c
-rH
n
o
•H
M
CU.
0 C
•H -H
M
-------
Log PS
Log FC or Log TC
I
3
I
4
1
4
1
6
I
8
24 Raw Sew,-
24 Sew 2°-
17 C12 Eff-
Log FS
Log TC
Figure 21. Comparison of low-flow, storm, sewage, soils, and stream
sediment indicator and Pseudomonas densities.
77
-------
u
H
U
t,
CTI
O
CO
fn
01
o
CO
6 r
5
4
3l_
2
1
2
1
3
lU
4r
CP
0
U _i
<
£ -2
_ -3
in
u
S-i
0-
_L
_L
LSD tost homogeneous
Subsets
Summer >
Winter - spring >
Fall - winter
F-stat. not sig.
F-stat. not sig.
Spring - summer >
Winter - fall
Wint. - spr. - summer >
Fall - wint. - spr.
Spring >
Wint. - summer - fall
Winter - fall >
Summer - spring
Fall - summer >
Winter - fall >
Spring
Season Spr. Summ. Fall Wint.
n= 38 21 40 52
Figure 22. Woodlands seasonal low-flow variations.
78
-------
REFERENCES
1. Zogorski, J. S., W, G. Characklis, R. H. Schumer, and
F. Gudet. Temporal Characteristics of Stormwater Runoff
An Overview". Pres. at Urbanization and Water Quality
Control Conference, Rutgers Univ., July, 1975.
2. "Conceptual Engineering Report - Kingman Lake Project".
R. F. Weston, Inc., F.W.P.C.A. Dept. of the Interior.
11023 FIX 08/70, 1970.
3. Allen, M. J. and E. E. Geldreich. "Bacteriological
Criteria for Ground Water Quality". (Presented at
Second Natl. Ground Water Quality Symposium, Denver,
Colo., Sept. 25-27, 1974) 21 p.
4. Cleaning Our Environment: The Chemical Basis For Action.
Amer. Chem. Soc., Washington, D.C., 1969.
5. Texas Water Quality Standards. Tex. Water Quality Board.
Austin, Texas, February, 1976.
6. Hoadley, A. W. Jour, of the New England Water Works
Assn., 83, 1964.
7. Mallmann, W. L. and A. Sypien. Amer. Jour, of Public
Health, 2_4: 681-688, 1964.
8. Geldreich, E. E. "Applying Bacteriological Parameters
to Recreational Water Quality". Jour. Amer. Water Works
Assn.. 62_(2): 113T120, Feb., 1970.
9. Bott, T. L. "Bacteria and the Assessment of Water
Quality". In. Biological Methods for the Assessment of
Water Quality". Amer. Soc. For Testing and Matls., STP
528, 1973.
10. Rosenberg, M. L. "How to Investigate a 'Waterborne1
Outbreak". Proc. Tech. Conf. on Water Quality. Amer.
Water Works Assn., December 8-9, 1975, Atlanta, GA.
79
-------
11. Koplan, J. P. "Turista Domestica Aquatica - Gastroin-
testinal Illness at Crater Lake". (Presented at 25th
Annual Epidemic Intelligence Service Conf., U.S. Dept.
H.E.W., P.H.S. Center for Disease Control, Atlanta,
April 5-9, 1976).
12. Englebrecht, R. S., M. Weber, P. Amirhor and B. Severin.
"New Microbial Indicators of Disinfection Efficiency".
Pub. Health Aspects of_ Viruses In Water. Ed. by N. A.
Clarke. Biological Cont. Br., Water Qual. Div., U.S.
E.P.A., Cincinnati, Ohio, 5: 7, April, 1976.
13. Highsmith, A. K. and R. L. Abshire. "Evaluation of
a Most Probable - Number Technique for the Enumeration
of Pseudomonas aeruginosa". Appl. Microbiol. 30 (4);
596-601, October, 1975.
14. Carson, L. A., M. S. Favero, W. W. Bond and N. J.
Petersen. "Factors Affecting Comparative Resistance
of Naturally Occurring and Subcultured Pseudomonas
aeruginosa to Disinfectants". Appl. Microbiol. 23(5):
863-869, May, 1972.
15. Standard Methods for the Examination of Water and
Wastewater, 13th Ed. A.P.H.A., W.P.C.F., A.W.W.A., 1971.
16. Biological Analysis o_f Water and Wastewater. Applica-
tion Manual AM 302. Millipore Corp., Bedford, MA, 1973.
17. Difco Supplementary Literature, Difco Laboratories, Inc.,
Detroit, Mich., 1968.
18. Difco Manual of_ Dehydrated Culture Media and Reagents
for Microbiological and Clinical Laboratory Procedures,
Ninth Ed., Difco Laboratories, Inc., Detroit, Mich.,1967.
19. Pseudomonas Agar F. Product of Difco Laboratories.
20. m-PA Agar. Ref. Dr. E. E. Geldreich, U.S.E.P.A., N.E.R.C.,
Cincinnati, Ohio, 1974. (Also, Standard Methods; 14th Ed.)
21. Analytab Products, lac., Carle Place, N. Y. (API-20
Enteric Strip and Coding Apparatus).
22. Burrows, W. Textbook of Microbiology, 19th Ed./ W. B.
Saunders Co., Philadelphia, 1969.
23. Davis, B. D. Microbiology, 2nd Ed., Harper and Roe
Publ., Inc., 1973.
24. Kenner, B. A. and H. P. Clark. "Detection and Enumera-
tion of Salmonella and Pseudomonas aeruginosa". Jour.
Water Poll. Control Fedn., 46(9): 2163-2171, September,1974.
80
-------
25. Statistical Package for the Social Sciences. Nie, N.
H. , et al. McGraw-Hill Book Co., New York, 1970.
26. Van Donsel, D. J., E. E. Geldreich, N. A. Clarke.
"Seasonal Variation in Survival of Indicator Bacteria
in Soil and Their Contribution to Stormwater Pollution".
Appl. Micro. 15_(6): 1362-1370, November, 1967.
27. Mallmann, W. L. and W. Litsky. "Survival of Selected
Enteric Organisms in Various Types of Soil". Amer. Jour.
of_ Pub. Health, 41; 38-44, January, 1951.
28. Van Donsel, D. J. and E. E. Geldreich. "Relationships
of Salmonellae to Fecal Coliforms in Bottom Sediments".
Water Research, 5_: 1079-1087, 1971.
29. Soderlund, G. and H. Lehtinen. "Comparison of Dis-
charges from Urban Storm-water Run-off, Mixed Storm
Overflow and Treated Sewage". Advances in Water Poll.
Res., Proc. 6th Inter. Conf. Jerusalem, Ed. by S. H.
Jenkins, Pergamon Press, N. Y. , 1973.
30. Glover, G. E. and G. R. Herbert. Off. of Res. and
Monitoring. U.S.E.P.A., Washington, D.C., EPA-R2-73-124,
January, 1973.
31. Stream Pollution and Abatement from Combined Sewer Over-
flows, Buchyrus, Ohio, U.S. Dept. Interior, Burgess and
Niple, Ltd., FWPCA Contract 14-12-041, November, 1969.
32. "Urban Storm Runoff and Combined Sewer Overflow Pollution".
Envirogenics Co., U.S.E.P.A., 11024, FKM 12/71, December,
1971.
33. Bender, R. J. "Ozonation, Next Step to Water Purifica-
tion". Power, 114_(8) : 58-60, August, 1970 (Abs.) .
34. Microstraining and Disinfection of Combined Sewer Over-
flows. U.S. Dept.'Interior, Fed. Water Qual. Adm. 11023
EVO 06/70.
35. Moffa, P. E., E. C. Tifft, Jr., S, L. Richardson and J. E.
Smith. Bench-Scale High-Rate Disinfection of Combined
Sewer Overflows: With Chlorine and Chlorine Dioxide.
Natl. Environ. Res. Ctr., U.S.E.P.A., EPA-670/2 75-021.
36. Davis, E. M., L. W. Whitehead and J. D. Moore. "Dis-
infection". Jour. Water Poll. Control Fedn., 46(6);
1181-1191, June, 1974.
37. Davis/ E. M., J. D. Moore, D. Casserly, J. Petros and
W. DiPietro. "Disinfection". Jour. Water Poll. Control
Fedn. £7(6): 1323-1334, June, 1975.
81
-------
38. Claudon, D. G., D. I. Thompson, E. H. Christenson, G. W.
Lawton and E. C. Dick. "Prolonged Salmonella Contamina-
tion of a Recreational Lake by Runoff Waters". Appl.
Microbiol., 2_1(5) : 875-877, May, 1971.
39. Leopold, L. B. "Hydrology for Urban Land Planning -
A Guidebook on the Hydrologic Effects of Urban Land
Use". Geological Survey Circular 554, Washington, D.C.,
1968.
40. "Methods for Identifying and Evaluating the Nature and
Extent of Nonpoint Sources of Pollutants". U.S.E.P.A.
Off. of Air and Water Programs. Washington, D.C.,
EPA-430/9-73-014.
41. Geldreich, E. E., L. C. Best, B. A. Kenner and D. J.
Van Donsel. "The Bacteriological Aspects of Stormwater
Pollution". Jour. Water Poll. Control Fedn., 40 (11);
Part 1, 1861-1872, November, 1968.
42. Tafuri, Anthony N. "Pollution from Urban Land Runoff".
"News of Environmental Research In Cincinnati". U.S.
E.P.A., Cincinnati, Ohio, April 11, 1975.
43. Harms, L. L., P. Middaugh, J. N. Dornbush, J. R. Andersen.
"Bacteriological Quality of Surface Runoff from Agricul-
ture Land - Part II". Water and Sew. Works, 122(11);
71-73, November, 1975.
44. Harms, L. L., P. Middaugh, J. N. Dornbush, J. R.
Andersen. "Agricultural Runoff Pollutes Surface Waters -
Part I". Water and Sew. Works, 122(10): 84-85, October,
1975.
45. Cherkauer, D. S. "Urbanization Impact on Water Quality
During a Flood in Small Watersheds". Water Resources
Bulletin, 11(5): 987-998, October, 1975.
46. Davis, E. M. "Community Health Aspects of Surface Waters
in the Woodlands". Rept. to U.S.E.P.A., Storm and Comb.
Sewer Tech. Br. The Univ. of Texas at Houston,Sch. of
Pub. Health, June, 1974.
47. Davis, E. M. "Microbiological Quality of Stormwater
Runoff in the Woodlands, Texas". Rept. to U.S.E.P.A.,
Storm and Comb. Sewer Tech. Br., The Univ. of Texas at ;
Houston, Sch. of Pub. Health, June, 1975.
48. Winslow, D. E. and W. H. Espey, Jr. "Storm Runoff
Analysis of Residential Settings for the Woodlands".
TRACOR, Inc., Austin, Texas, Rept. No. T-72-AU-9585-U,
1972.
82
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-79-050f
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
MAXIMUM UTILIZATION OF WATER RESOURCES IN A PLANNED
COMMUNITY - Bacterial Characteristics of Stormwaters
in Developing Rural Areas
5. REPORT DATE
August 1979 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Ernst M. Davis
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
The University of Texas
School of Public Health
Post Office Box 20186
Houston, Texas 77025
10. PROGRAM ELEMENT NO.
1BC822, SOS #2, Task 02
11. OOMTn*OT/GRANT NO.
802433
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory--Cin.,OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/14
volumes of one report. Project Officers: Richard Field and
Anthony N. Tafuri, Storm and Combined Sewer Section, FTS 340-6674, (201) 321-6674.
16. ABSTRACT
An investigation of low flow and stormwater runoff bacteria content from
rural and urban areas was conducted over a two and a half year period. Data were
obtained from total coliform, fecal coliform, fecal streptococci, Salmonnella sp., Pseu-
dompnas sp., and Staphylococcus sp. for comparison to densities in known polluted source
such as secondary treated chlorinated municipal wastewater. The usefulness of the
currently employed indicator groups of bacteria was evaluated with respect to the
accompanying densities of pathogens. The hygienic quality of water when compared
to new bacteriological water quality standards for contact and noncontact recreation
was considered. Settling of stormwater suspended solids was closely associated with
bacterial reductions in the water column. The most useful indicators of pathogen
content in stormwater runoff were fecal coliforms. Total coliforms and fecal strepto-
cocci were poor indicators of pathogenic bacteria densities. Chlorine and ozone doses
for disinfection of stormwater containing high (=200 mg/1) suspended solids may exceed
8 mg/1 and 32 mg/1, respectively. Regrowth of total coliforms occurs following disin-
fection. Indicator group densities in urban stormwater runoff can easily exceed rural
runoff densities with continual increases occurring throughout a storm event. Fecal
coliform densities exceeded 2,000/lOOml in 13 to 24 monitored hydrographs and exceeded
200/lOOml in 22 of those hydrogra'phs. Fecal coliforms and fecal streptococci yielded
the highest correlations with the physical factors, flow, suspended solids, and turbidi'1
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Microorganisms, Bacteria, Viruses, Storm
sewers, Streams, Urban areas
Urban stormwater micro-
organisms, Pathogenic
microorganisms enumera-
tion, The Woodlands
13B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. OF PAGES
95
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (Rex. 4-77)
83
* its. ommian mams «n£ ts* -6$T~060/^<>?
------- |