EPA/600/2-79/050b
tates
lental Protection
Municipal Environmental Research EPA-600 2 79-050b
Laboratory July 1 979
Cincinnati OH 45268
and Development
maximum
Utilization of
Water Resources in a
Planned Community
Stormwater Runoff
Quality: Data
Collection,
Reduction and "
Analysis
-------
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-600/2-79-050b
July 1979
MAXIMUM UTILIZATION OF WATER RESOURCES
IN A PLANNED COMMUNITY
Stormwater Runoff Quality: Data Collection,
Reduction and Analysis
by
William G. Characklis, Frank J. Gaudet,
Frank L. Roe and Philip B. Bedient
Department of Environmental Science and Engineering
Rice University
Houston, Texas 77001
Grant No. 802433
Project Officers
Richard Field
Anthony N. Tafuri
Storm and Combined Sewer Section
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 pro-
ducts constitute endorsement or recommendation for use.
-------
FOREWORD
The U.S. Environmental Protection Agency was created because
of increasing public and governmental 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
require 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 Environ-
mental 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 of public drinking water supplied and to minimize the
adverse economic, social, health, and aesthetic effects of pollu-
tion. 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 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 storm water 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 neces.sary, to maximize its effective utilization through altera-
tions 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 recrea-
tional and aesthetic purposes was a major feature of the water re-
sources plan at The Woodlands. Control of downstream flooding
was also of great importance and so storage reservoirs, in the
form of recreational lakes and wet weather ponds, were created
by the developers. Water quality was a concern if the impound-
ments were to be aesthetically appealing and/or suitable for re-
creation. Therefore, a major sampling and analytical program was
designed to monitor water quality and quantity at different loca-
tions in the developing area. The Storn Water Management Model
(SWMM) provided the focal point for combining the water quality
and quantity data into a predictive tool for design and manage-
ment purposes.
SWMM was originally developed for highly urbanized areas and,
therefore, was calibrated for this project in an urban watershed
(Hunting Bayou). Subsequently, SWIM was modified to model runoff
and water quality from natural drainage areas, such as The Wood-
lands. Because of the lag in the construction schedule at The
Woodlands, the dense urban areas were not completed during the pro-
ject period. Consequently, Hunting Bayou and other urban water-
sheds 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. Patho-
genic bacteria were also enumerated since the role of traditional
bacterial indicators in stormwater runoff was not clear. Algal
bioassay tests on stormwater were conducted to assess the eutrophi-
cation potential that would exist in the stormwater impoundments.
The source, transport and fate of chlorinated hydrocarbons in storm-
water runoff was also investigated.
Several of the large Woodlands impoundments will receive re-
claimed wastewater as the major input during dry weather. Besides
their use as a source of irrigation water, the lakes will be used
iv
-------
for non-contact recreation — primarily fishing and boating.
Because the reclaimed wastewater must be disinfected, there was
a concern about disinfectant toxicity to the aquatic life in the
lakes. Consequently, comparative fish toxicity tests were con-
ducted with ozone and chlorine, the two alternatives available
at the water reclamation plant.
Porous pavement was considered by the developers as a method
for reducing excessive runoff due to urbanization and an experi-
mental 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 deter-
mined.
Hopefully, the results of this project will contribute in a
positive way to the development of techniques to utilize our urban
water resources in a manner more compatible with our cherished
natural environment.
v
-------
ABSTRACT
An ecologically planned community (The Woodlands, Texas)
has adopted a unique water management plan designed to avoid
adverse water quality and hydrological effects due to urbaniza-
tion while benefiting from the existing natural drainage. The
initial years of development were monitored by a comprehensive
sampling and analytical program in an effort to evaluate the
innovative new water resources concept. Data on water quantity
and quality were collected during dry weather and during storm-
water runoff. To supplement the prime study site, stormwater
samples were also collected at watersheds in the Houston area.
Parameters monitored during the reporting period were as follows:
rainfall, streamflow, chemical oxygen demand (COD), soluble
organic carbon (SOC), biochemical oxygen demand (BOD), ammonia
(NHo), nitrate (NO3), nitrite (NO2), total Kjeldahl nitrogen
(TKN), orthophosphates (ortho-P), total phosphorus (TP), dissolv-
ed oxygen (DO), pH, turbidity, total suspended solids (TSS), and
specific conductance. Data were analyzed for water quality re-
lationships in an effort to predict pollutant loads according to
land use. Comparisons were made to wastewater and rainwater
quality.
Significant relationships were observed between total
volume of runoff and total load of various pollutants. The load-
runoff relations are a function of the type of land use activity
in the watershed and have been used to simulate stormwater
quality responses.
This report was submitted in fulfillment of Grant No.
802433 by Rice University under the sponsorship of the U. S.
Environmental Protection Agency. This report covers the period
July 16, 1973, to May 31, 1976, and work was completed as of
December 31, 1976.
VI
-------
CONTENTS
Foreword iii
Preface iv
Abstract vi
Figures viii
Tables x
Abbreviations xii
Acknowledgment xiii
1. Introduction 1
2 . Conclusions 2
3 . Recommendations 5
4 . Site Descriptions 7
5 . Sampling and Monitoring Programs 29
6. Experimental Methods and Procedures 35
7 . Results and Discussion 50
References 96
Vll
-------
FIGURES
Number Page
1 Location of study sites 8
2 Average monthly air temperatures at Conroe, Texas ... 11
3 Existing drainage network for The Woodlands 13
4 panther Branch near confluence with Bear Branch .... 14
5 Panther Branch near confluence with Spring Creek. ... 15
6 Panther Branch travel time discharge relationship ... 17
7 Schematic water balance for The Woodlands 20
8 photographs of typical channel conditions for Hunting
Bayou 22
9 photographs of typical secondary drainage system for
Hunting Bayou 24
10 Hunting Bayou watershed land use 25
11 Westbury Square watershed 26
12 Location of sampling sites and rain gauges within the
panther Branch watershed 31
13 The Woodlands man-made lake system with locations of
stormwater monitoring sites 33
14 Flow chart for storm samples 39
15 Generalized pollutographs observed for stormwater
parameters 65
16 Comparison of total suspended solids pollutographs at
P-10, Westbury and Hunting Bayou during similar
storm flow 67
17 Load-runoff relationships for total Kjeldahl nitrogen
and total phosphorus 70
18 Load-runoff relationships for nitrate and ammonia. . . 71
viii
-------
FIGURES (continued)
19 Load-runoff relationships for suspended solids and
chemical oxygen demand (COD) 72
20 Load-runoff relationships for soluble COD and soluble
organic carbon 73
21 Comparison of p-10 and P-30 temporal distribution of
streamflow, total suspended solids and total COD
for the storm event of April 8, 1975 73
22 Comparison of P-10 and P-30 temporal distribution of
streamflow, total phosphorus and Kjeldahl nitrogen
for the storm event of April 8, 1975 79
23 The Woodlands construction activity in relation to the
P-10 and P-30 sampling sites 80
24 Hydrograph and predicted solids load for the P-30
hydrograph period of 10/28/74 to 4/12/75 84
25 Hydrographs and observed and simulated mass flow
curves for P-30 storm events 85
26 Fitted curves for storm runoff and pollutant mass
flows observed at P-10 on 4/8/75 88
27 Hydrograph and cumulative hyetograph at the Lake B
gauging station for the April 8, 1975 storm event . . 92
28 Reduction of total suspended solids through The Wood-
lands lake system 93
IX
-------
TABLES
Number Page
1 Summary of Land Use Allocation for The Woodlands. ... 10
2 Slope Characteristics of Bear and Panther Branches. . . 18
3 Hunting Bayou Watershed Characteristics 25
4 Westbury Square Watershed Characteristics 26
5 Watershed Characteristics 28
6 Field Equipment Operated and/or Maintained by Rice
University and Used for Sampling and Monitoring ... 36
7 Annual Storm Sampling Costs per Site for Twelve Storms. 40
8 Hydrological Definitions and Calculations 44
9 Water Quality parameters for Stormwater Runoff and
Low Flow 46
10 Leaching in Membrane Filters 46
11 Precision, Accuracy, and Preservation 48
12 Summary of Low-Flow water Quality Parameters, Average
Concentrations 51
13 Storm Event Hydrology Summary 53
14 Rainwater Quality Analysis 55
15 Comparison of Rainwater and Runoff Quality in Houston
and at The Woodlands 56
16 Comparison of Groundwater Quality After Rainfall to
Runoff Water Quality, The Woodlands 57
x
-------
TABLES (continued)
17 Runoff Water Quality Summary—Mass Flow and Weighted
Average 59
18 water Quality for Stormwater Runoff, Untreated Sewage,
and Treated Sewage 64
19 Comparative Data of First Flust Effect for the 3 Storm
Events at Site P-10 68
20 Pollutant Load Ranking of the Four Study Area Water-
sheds 74
21 Annual Mass Loads from P-10 and P-30 Watersheds
(October 1974 - September 1975) 76
22 Comparison of Stormwater Quality at P-10, P-30 and
Developing Areas During Storm #10 81
23 Comparison of Simulated and Observed Results for
Three Storms 86
24 Summary of Water Quality Parameters for Sites Lake A
and Lake B During the April 8, 1975 Storm Event ... 94
25 Stormwater Sediment Removal at Lake Harrison 95
XI
-------
LIST OF ABBREVIATIONS
BOD Biochemical Oxygen Demand
COD Chemical Oxygen Demand
DO Dissolved Oxygen
g mass flow rate
HB Hunting Bayou watershed
ortho P Orthophosphate
P mass of pollutant
PDS slope of the load-runoff curve at some point in time
PVC Polyvinyl Chloride
Q volumetric flow rate
r rate of runoff
S total storage
SOC Soluble Organic Carbon
SWMM EPA Stormwater Management Model
TKN Total Kjeldahl Nitrogen
TOC Total Organic Carbon
TP Total Phosphate
TSS Total Suspended Solids
W3 Westbury watershed
Xll
-------
ACKNOWLEDGMENTS
This project was supported by the U. S. Environmental
Protection Agency, The Woodlands Development Corporation and
Rice University.
The Project Director expresses his sincere gratitude to
the EPA Project Officers, Anthony Tafuri and Richard Field and
to James Blackburn, Skip Christy, Peter D'Alessandro, Ralph
Everhart, Bill Kendricks, Ken Kimbrough, Robert Heineman, Plato
Pappas and James Veltman of The Woodlands Development Corpora-
tion for their time and valuable contributions.
Acknowledgments also go to Robert Gabrysch, Jim Hutchinson,
Steve Johnson, Emil Kamanski, and Robert Smith of the Water
Resources Division, U. S. Geological Survey, for collecting and
compiling essential hydrologic data at The Woodlands.
Credit is also due to the members of this multidisciplinary
project. They are as follows:
Rice University
P. B. Bedient
E. Birch
J. Bishop
K. Carter
W. G. Characklis
J. Coffey
J. A. Conner
M. Curtis
S. Davis
J. Delia
F. M. Fisher
G. For tenberr y
F. J. Gaudet
D. Gee
P. Graves
J. H. Hall
B. Hammond
T. D. Hayes
D. Harned
J. D. John
M. A. Kessick
J. M. King
J. LeBlanc
M. Lee
H. M. Liljestrand
W. L. Lloyd
K. Manchen
D. C. Marks
R. Morr ison
L. P. Metzgar
T. Miller
P. McSherry
L. Price
F. L. Roe
J. B. Smith
C. S tagg
M. Walker
C. H. Ward
J. C. Weismiller
L. Wong
A. Yarletts
J. S. Zogorski
Kill
-------
University of Texas School of Public Health - Houston, Texas
D. Casserly P. Mittlemark
E. M. Davis D. Moore
J. Greene H. Tamashiro
P. Mattox
Espey, Huston and Associates, Inc. - Houston, Texas
W. H. Espey, Jr. T. Remaley
E. Diniz F. Sofka
D. Holloway D. E. Winslow
S and B Engineers - Houston, Texas
L. Chandler
W. Davis
J. Matson
Franklin Institute Research Laboratory - Philadelphia, Pennsyl-
vania
R. Ho 1linger
E. Thelen
xiv
-------
SECTION 1
INTRODUCTION
Surface waters are maintained at a desirable natural quality
standard through the control of pollutant discharges. Histori-
cally, abatement procedures have focused upon control of indus-
trial and municipal effluents, the point sources. Increasingly
stringent point source effluent standards have generated concern
for non-point source pollution control. Urban stormwater runoff
is considered a major problem in this regard.
The hydrological characteristics of natural watersheds change
with urbanization. Replacement of flow-retarding vegetation with
impervious surfaces, such as roads and buildings, increases the
amount of stormwater runoff. Removal of the water is tradition-
ally implemented by the use of an urban drainage system consist-
ing of storm sewers and deep, concrete-lined drainage ditches,
designed specifically for prompt drainage. Increased runoff
volumes and peak flow rates result, creating problems of down-
stream flooding and channel erosion.
Infiltration of stormwater is a major groundwater recharge
source, however the urban emphasis of surface removal minimizes
the infiltration rate, resulting in a lowered water table and
possible urban land subsidence problems. Water quality in the
area becomes generally poorer because the natural purification
which infiltration provides is compromised.
The urban environment typified by industry, highways, solid
waste and high population density provides a major pollutant
source for runoff waters (1). Recent investigations recognize
the significance and magnitude of pollution problems from urban
stormwater runoff. In terms of specific pollutants, the sediment
yield problem is the most dramatic. Due primarily to urban con-
struction, urban sediment loads were found to be as much as 75
times greater than loads in agricultural regions (2, 3). Other
runoff pollutants reported higher in urban regions include dis-
solved solids (4), coliforms (5), biochemical oxygen demand and
chemical oxygen demand (BOD and COD) (6), polychlorinated bi-
phenyls, heavy metals, pesticides and fertilizers (6-9). To
verify these findings, the present study proposed to monitor
urban development activities and define stormwater pollution
characteristics.
-------
SECTION 2
' CONCLUSIONS
Stormwater runoff from an undeveloped forested watershed is
relatively low in pollutants and pollutant indicators. Typical
values are: Total Phosphorous (TP) 0.06 mg/1, Total Kjehdahl
Nitrogen (TKN) 1.24 mg/1, Total Suspended Solids (TSS) 36 mg/1,
Total Chemical Oxygen Demand (COD) 42 mg/1, and Dissolved Oxygen
(DO) 6 mg/1.
Development in the forested watershed has significantly in-
creased TSS and nutrients in runoff. COD and other organic para-
meters were not affected. Increased TSS values are a result of
sediments washed from construction sites, some located within the
floodplain. Development of The Woodlands area will increase sur-
face water turbidity. Dredging will probably be necessary to re-
move excess sediments from lakes and ponds. Increased nutrient
loads from developed areas will create algal and macrophyte growth
problems in The Woodlands lake system.
Houston urban runoff contains higher nutrient and TSS loads
than forest runoff. Nutrient concentrations (ammonia, TKN,
nitrate, nitrite, TP, ortho phosphates) are as much as 10 times
greater in urban areas. TSS concentrations are 4 times greater.
Higher concentrations combined with increased runoff coefficients
(the amount of runoff for a given amount of rainfall) provide re-
ceiving waters with heavy pollutant loads in urban areas. Sedi-
ment buildup and algal growth problems will result in impoundments
receiving such flows.
A man-made lake serves as an effective trap for excessive
sediments transported by construction site runoff. During seven
separate storm events totaling 10.2 inches (26 cm) of rainfall,
180 tons (1.6 x 105 kg) of sediment entering 110 ac-ft (13.56ha-m)
Lake Harrison was reduced to 34 tons (3.08 x 104 kg) in the ef-
fluent. This was an 81% reduction in sediment load.
A definite first flush was observed for urban and undevel-
oped watershed runoff, most commonly for TSS and turbidity para-
meters. The flush is related to transport of stream bed sedi-
ments. Urban drainage systems have increased transport potential
and therefore exhibit higher flush concentrations.
-------
Rainwater contain phosphates, nitrogen and COD which account
for a significant portion of runoff pollutant loads.
Natural soils are capable of removing nutrients found in
rainwater. Disturbed soils in developing areas lose this capa-
bility.
Municipal wastewater would require advanced treatment to
meet nitrogen and phosphorus concentrations in stormwater runoff.
Secondary treatment of wastewater will lower suspended solids and
COD concentrations below that in stormwater runoff.
A linear relationship exists between total pollutant loads
and total stormwater runoff which is useful in comparisons be-
tween watersheds and analytical prediction of stormwater pollu-
tant loads.
A statistical ranking of four watersheds, on a Ib/ac/in
(kg/ha/cm) of runoff basis, indicates that urban watersheds are
clearly the greatest producers of TSS and nutrient loads. Loads
from the forested and developing watersheds are lower by as much
as an order of magnitude.
The load-runoff curves may be used to sequentially simulate
mass flow curves. Simulation of a six month period, containing
three measured storm events, produced reasonable comparisons of
observed and simulated curves.
Stormwater treatment at The Woodlands should be restricted
to sedimentation in the man-made lakes. Costs prohibit the con-
struction of facilities which would result in no significant ef-
fect on the use of lakes.
Stormwater sample preservation in battery operated samplers
by ice refrigeration is not feasible.
Representative sampling of TSS (with high clay content)
by automatic samplers was excellent.
Recording monitors (pH, DO, temperature and turbidity) re-
quire weekly maintenance for recalibration and antifouling
measures.
Cross-sectional homogeneity of stream parameters should be
verified before choosing stream sampling points.
Discharge proportional sampling is an efficient method lead-
ing to characterization of stormwater runoff.
Glass fiber filters were not suitable for removing fine or
coloidal particles from natural water samples.
-------
Refrigeration plus mercuric chloride preservation was not
adequate for stabilization of ortho-phosphate (ortho-P) at low
levels in surface water samples.
-------
SECTION 3
RECOMMENDATIONS
Continued research at The Woodlands is needed to assess its
planned water resources system. These studies should be con-
ducted at a future date when development is more advanced.
Research is needed to establish the relationship of air
pollution to rainwater and runoff quality. Special attention
should be focused on the water quality effects of projected
ambient air criteria.
Sediment discharge from construction area stormwater runoff
should be controlled. Suggested methods are:
1. Prevention; i.e., prohibiting construction in flood
plain areas.
2. Treatment; i.e., by sedimentation basins or more
sophisticated methods.
3. Control of erosion by various techniques.
More stormwater quality data should be obtained on Hunting
Bayou and Westbury watershed so that unit loadographs can be
simulated for single storm events. Investigations can then de-
termine relationships between the gamma distribution shape para-
meters (n, k) and land use or physiographic factors in the water-
shed.
Simpler methods for determining annual loads, partially de-
veloped during this research, should be refined and verified with
field data. The methods show great promise for differentiating
between point source and non-point source pollution loads.
Automatic discrete sampling of streams is most efficiently
carried out by a high speed (2 ft/sec or 60 cm/sec) peristaltic
or vacuum chamber type, battery operated sampler such as Isco,
Manning or Sirco models. Some mechanical and electronic mal-
functioning was encountered with the Manning S-4000 sampler.
-------
Sample filtration through Millipore HATF membrane filters
was adequate, however, Nucleopore membrane filters should be
used due to the added advantage of low tare weight and a hydro-
phobic response allowing rapid drying to constant weight.
Analyses of TKN, TP and COD on refractory compounds by the
Technicon automated methods (helical digestor for TKN and TP and
digestion bath for COD) were inferior to manual methods because
of incomplete digestion.
Efficiency of pumping in Sigmamotor samplers is improved by
replacing existing pumping mechanism with a "Masterflex" pump-
head.
Better field preservation methods for discrete samples
should be developed.
-------
SECTION 4
SITE DESCRIPTIONS
Several sites in the Houston metropolitan area were chosen
to determine if stormwater runoff quality is dependent on land
use and development activities. The primary study site is The
Woodlands, Texas, a planned satellite city selected for a com-
prehensive investigation of runoff quality during all phases of
development. Two other watersheds were chosen to supplement data
collected from The Woodlands. Hunting Bayou is a developed
watershed with strong industrial influences and deteriorating
residential areas. Westbury Square is a middle class residential
area chosen because of the absence of construction in the water-
shed. The locations of these study sites are shown in Figure 1.
Each watershed is comprehensively described in the following
text.
THE WOODLANDS
General Description
The Woodlands is a newly planned community being developed
in southern Montgomery County, Texas. The community is situated
in a heavily forested tract about 35 niles (56 kin) north of
Houston, directly west of Interstate 45 (see Figure 1). The
Woodlands encompasses 17,776 acres (7194 ha) and will be devel-
oped over a twenty-year period beginning September, 1972. In
contrast to a residential subdivision, The Woodlands will contain
all services of a modern city, including facilities for social,
recreational, educational, commercial, institutional, business
and industrial pursuits. The community concept is committed to
high standards for environmental and lifestyle quality. The
phased, long range development places priority on ecological pre-
servation and balance, as well as social and habitational quality.
This objective is to be accomplished through a comprehensive en-
vironmental preservation and management program, including plan-
ning and design controls. The water resource system in The
Woodlands, including its drainage system, is a good example of
such planning and was the primary subject of this research.
Development plan
The prime objective of The Woodlands is to provide the
finest urban environment in the Houston metropolitan area in
-------
INTERCONTINENTAL
HUNTING BAYOU
WESTBURY
SQUARE
Figure 1. Location of study sites.
1 mi = 1.6 km
-------
terms of physical setting and quality of human life and community
services. The basis for all aspects of development in The Wood-
lands is a unique ecological inventory conducted from 1971-1973.
The geology, soils, drainage, water resources, weather, vegeta-
tion, and wildlife endemic to The Woodlands were evaluated by
specialists. The findings of these studies were the basis for
developing criteria to locate roads, homes, offices, and other
physical structures to be built at The Woodlands.
A summary of the land use allocation for The Woodlands is
shown in Table 1. Residential areas will occupy 6,820 acres
(2,760 ha) of The Woodlands site. A total of 33,000 dwelling
units are programmed. Housing units planned include single
family detached, townhouses and patio, and apartments. The pro-
jected population in 1992 is 112,000. Another 1,699 acres (688
ha) is being designated for restricted industrial use. Additional
area has been allocated for retail, commercial, office, open
space, and other land sales. Approximately one-third (30.2%) of
The Woodlands has been designated open space. The majority of
this space will be located within the floodplain of Panther
Branch and its major tributary, Bear Branch.
The Revised General Plan for The Woodlands includes all of
those elements essential to modern living. Social, recreational,
educational, commercial, institutional, business, cultural and
industrial elements are planned within The Woodlands. A concern
for nature and convenience for man were two of the major criteria
used in the development of the General Plan for The Woodlands.
Climate
The macroclimate of the Houston metropolitan area is domi-
nated by the Gulf of Mexico. Winters in the region are normally
mild, while summers are hot and humid. The mean daily winter
temperature is about 50° F (10° C) , whereas the mean daily summer
temperature is 82° F (27.8° C). The average maximum daily tem-
perature is 97° F (36° C) and occurs in August. The average
minimum daily temperature is 38° (3.3° C) and occurs in January.
Freezing temperatures occur about seven days per year. Figure 2
summarizes the variation of ambient temperature for the Spring
Creek basin, in which The Woodlands is located.
Average yearly rainfall in The Woodlands totals about 46
in (117 cm) and is evenly distributed throughout the year.
Annual extremes range from 17.66 in (44.86 cm) in 1900 to 77.43
in (196.67 cm) in 1973. However, a majority of years of record
(75%) have recorded annual rainfall between 30 and 60 in (76 and
152 cm). April, May, November and December are usually the wet-
test months, while March is the driest month. The areal varia-
tion of rainfall for specific storm events can be significant
especially during summer months. The majority of rainfall
occurring during June, July, August and September is associated
9
-------
TABLE 1. SUMMARY OF LAND USE ALLOCATION FOR THE WOODLANDS
General Breakdown
1 ac = .405 ha
Acres percent
Residential 6,820 38.4
Retail and Commercial 470 2.6
Office 724 1.5
Industrial (including roads) 1,699 9.6
Open Space 7,803 43.9
Other
_ phase one* 276 1.6
- health 39 .2
- churches 72 .4
- schools 322 1.8
17,776 100.0
Open Space (developed uses)
University 350 2.0
protective Services 22 .1
Neighborhood, mini, playlot 90 .5
District parks 90 .5
Community Center Library 16 .1
Village Centers 19 .1
Pathway 354 2.0
Townwide Road 990 5.5
Golf (public) 300 1.7
Golf (private) 200 1.1
Equestrian center 10 1
Open Space (undeveloped uses)
2,441 13.7
Floodways 3,635 20.4
Drainage 1,328 7.5
Wildlife Corridors 175 1.0
Miscellaneous 2 25 1.3
5,363 30.2
*Initial development - Conference Center, model
residential areas, and office buildings.
10
-------
E
•rH
X
I
•rH
rf
P
fl
nJ
(U
/ I I
/ '
\rtJ
0)
*-1
\ S i X
\ \ s
\ \I \
\ \ \
\ \ \
\ \ \
\ \ \
\ \ \
\ \ \
\ \ \
\ \ \
\ \ \
I I I I
o o o o o o o
o
-------
with thunderstorms. Precipitation during these months is unpre-
dictable. Frequently an inch or more (a few centimeters or more)
of rainfall can be recorded in one part of a watershed, while a
short distance away no precipitation occurs.
Snow rarely occurs in the Houston metropolitan area, al-
though three separate snowfalls occurred in 1972. Prevailing
winds are northerly in January and southeasterly during the rest
of the year. Destructive winds are not frequent, although ex-
cessive rain and high winds normally accompany tropical depres-
sions which move inland from the Gulf of Mexico.
Existing Drainage
The natural drainage for The Woodlands community is shown in
Figure 3. Approximately 80 percent of the development is drained
by Panther Branch, a tributary of Spring Creek. The remaining
portion of the development drains directly into Spring Creek,
which has a total drainage area of 750 square miles (1942.5 sq.
km). Because Panther Branch and its tributaries represent the
major existing drainage for the development site, the hydrologic,
morphologic and transport characteristics of this stream are im-
portant.
Panther Branch is approximately 14.6 miles (23.5 km) in
length and has a total drainage area of 36.2 square miles. It
originates north of FM road 1488 and travels in a south south-
easterly direction. The only major tributary to Panther Branch
is Bear Branch, which drains the northwestern portion of the
watershed. Bear Branch is 9.0 miles (14.48 km) in length and
drains about 42% of the Panther Branch basin. The headwaters of
Bear Branch are located north of FM 1488, near Egypt, Texas, and
water flows in a southeasterly direction.
Both Bear Branch and Panther Branch meander extensively and
have well-defined low-flow channels. Representative examples of
the streams' morphology are shown in Figures 4 and 5. Alluvial
sediments, small riffles, and slow moving pools are commonplace
within Panther Branch and Bear Branch. The width of the low flow
channel is highly variable but is normally between 5 and 20 feet
(1.5 and 6.1 m). The depth of the established channel increases
from about 2 to 4ft (.61 to 1.22 m) in the headwaters to approxi-
mately 8 to 12 ft (2.44 to 4.88 m) near the confluence with
Spring Creek. When the capacity of the defined channel is ex-
ceeded, storm runoff discharges into a very broad, flat flood-
plain. Presently, the floodplain has a heavy brush cover. The
width of the floodplain varies along the length of the stream
but is typically 1000 to 2000 ft (304.8 to 609.6 m) for a 100
year storm event. Flood runoff is characterized by low veloci-
ties and shallow depth because (a) a large land area is inundated,
(b) flow resistance is high and (c) hydraulic slope is low.
Excluding those areas presently under construction, essentially
12
-------
PANTHER BRANCH
DRAINAGE AREA
OUTLINE
SCAIE IN MHES
SAWDUST ROAD
Figure 3. Existing drainage network for The Woodlands
(numbers shown indicate elevation above mean sea level).
1 mi = 1.6 km
13
-------
Figure 4. Panther Branch near confluence with Bear Branch,
14
-------
Figure 5. Panther Branch near confluence
with Spring Creek.
15
-------
no evidence of any serious erosion can be found anywhere in
Panther Branch watershed.
Time of travel measurements within the Panther Branch water-
shed yielded the asymptotic curve relationship presented in
Figure 6. The relationship between travel time and discharge was
developed for a 6.8 miles (10.9 km) reach of Panther Branch using
fluorescent dye studies during low flow and time for hydrograph
passage during storm runoff. Dye studies indicate water veloci-
ties below 0.1 ft/sec (.03 m/sec) are typical for dry weather
flow. Consequently it is estimated that during low-flow condi-
tions approximately 10 to 14 days of travel time is required for
an element of water originating in the headwaters of Panther
Branch to reach Spring Creek. The long travel time is a direct
consequence of low channel slopes present throughout the drainage
network. The elevation of the bottom of the stream channel above
mean sea level for Panther Branch, Bear Branch and their tribu-
taries is shown in Figure 3. Table 2 summarizes the change in
elevation and slope for several stream reaches. The total change
in channel elevation across the drainage basin is about 120 ft
(37 m), with an average rate of change of 8.2 feet/mile (1.6 m/km)
(0.16%). The slopes reported were calculated using river mileages
delineated from USGS topographic maps (scale 1:23,000), however
this technique normally underestimates the actual length of stream,
especially when considerable stream meandering is present. The
actual slope of Panther Branch is, therefore, more nearly 5 to 7
feet/mile (.95 to 1.33 m/cm).
Higher velocities of stormwater flow permit shorter travel
times through the watershed. The hydrograph crest of a large
storm event, flow greater than 100 cfs (2.8 m3/sec) will traverse
the watershed in less than 24 hours with surface water velocities
approaching 1 ft/sec (.31 m/sec).
The impact of low channel and land slopes within The Woodlands
is reflected in a relatively low surface runoff coefficient.
U.S. Geological Survey data for the 1973 and 1974 water years
show that only 23% of total rainfall ended up as surface runoff.
The remaining 77% either evaporated, transpired or infiltrated
into the ground. It should be noted that rainfall was heavy during
the 1973 and 1974 water years, respectively 77 in (195.6 cm) and 51
in (129.5 cm). Thus it is estimated that only 10-15% of rainfall
will run off during a year of average rainfall, 45 in (114.3 cm)
(14) .
Runoff from the Panther Branch watershed is not evenly dis-
tributed throughout the year. During the summer months of May
through September little discharge occurs, except immediately
following an intense and prolonged rainfall. The average daily
low-flow discharge at Sawdust Road, including summer months, is
1 to 2 cfs (.03 to .06 m3/sec). An average daily discharge of 100
cfs (2.8 m3/sec) at this site is exceeded approximately 5% of the
time.
16
-------
U 8 .
-I
or
D
O
I
UJ
6 .
4 .
LU 2.
or
0 100 200 300 400
STREAMFLOW, CFS
Figure 6. Panther Branch travel time
discharge relationship.
I mi = 1.6 km
1 cfs = .0283 nr/sec
17
-------
H
CQ
Pn
O
CO
u
H
EH
CO
H
UJ
H
E-i
U
U
W
&i
s
CO
CN
OQ
r-J
0) ^
d
d
a
r-t
r1-!
u
• H
Pi
00
r-l
i
O
d d
O fl
o cu
m
CN
•
o
CO
CM
r-l
cn
^D
U3
in
r|
r*n
4J
O -H
-P >
,d
to
d
rd
Oj
r-<
r-l
(
o
r»
in
r-l
m
o
CTi
rj J3
5 o£
•H 4-1 -i-l
? ? ^
,d a)
(1) O 0)
U d U M
d fd d o
0) M (U
P CQ P en
r-l r-l d
4-1 S-i U-l -H
d rd d M
O 0) O Q<
O CQ U CO
•g
d
rd
M
CQ
M
(U
rd
4-)
d
rd
&)
VD
1— 1
•
o
5 rVJ
0)
0) 0) (1)
MUM
(u d u
4J 0)
rd 3 iJ>
? r-l d
•73 <4-l • r-l
rd d M
a) o cu
K U CO
•s
d
fd
M
ffl
M
0)
rd
4-)
d
rd
a<
g e
^
m
U3 O
• m
H
II II
•H 4J
6 IM
r-l rH
18
-------
WATER RESOURCE SYSTEM OF THE WOODLANDS
An approximate water balance for The Woodlands is presented
schematically in Figure 7. The annual rainfall at The Woodlands
is partitioned as runoff into existing lakes and streams and in-
filtration into the ground. Losses result from evapotranspira-
tion, evaporation and subsurface transport to streams.
The maintenance of a satisfactory groundwater reservoir
above the perched water table is critical for the continued
growth of vegetation. Any drainage system for The Woodlands must
consider the detrimental consequences of disrupting the movement
of water within this shallow aquifer. Deeper aquifers (1800 ft or
549 m) are used for community water supply.
A series of wet weather ponds and variable volume lakes will
serve as recreational"centers, wildlife preserves and, more impor-
tantly, storage of stormwater runoff. This system of waterbodies
also contributes to the maintenance of an adequate perched water
table for plant life. Lake water will be lost primarily through
surface evaporation and irrigation. Inflow to the lake system
will result primarily from stormwater runoff and reclaimed waste-
water (approximately 20 mgd or . 88 m3/sec)., If necessary, treated
sewage effluents can be discharged directly into Panther Branch.
"Natural Drainage System"
The Woodlands Development Corporation has specified that the
basic drainage system for their new community will utilize "natu-
ral drainage" concepts. Related design principles have been re-
ported (15) and will not be described in detail herein. Rather,
a brief summary is presented on the "natural drainage" concept
and reasons for its selection in The Woodlands.
The normal procedure for disposing of stormwater runoff within
the Houston Metropolitan Area is to enlarge the natural drainage-
ways by deepening and widening existing stream channels and pro-
viding supplementary lateral drains. In the City of Houston,
this approach generally results in storm sewers for the lateral
drainage and deep, wide, concrete-lined ditches for the major
drainage. This solution to stormwater disposal, although widely
used and approved by the City of Houston, was incompatible with
one of the major criteria used in developing The Woodlands--pre-
serving and enhancing the natural environment. "Natural drainge"
concepts adopted by Woodlands Development Corporation are envi-
sioned as a method of providing adequate drainage and yet mini-
mizing disruption of natural processes. Primary objectives of
the drainage approach are to impede movement of surface runoff
and to recharge stormwater runoff into the ground where feasible.
Impediment and storage are provided by modifying existing drain-
ageways, where necessary, with wide shallow swales, check dams,
storage lakes and wet weather ponds. In comparison with the
19
-------
45'
TOTAL PRECIPITATION
5M80
CONSUMPTIVE USES
WATER DEMAND OF
THE WOODLANDS
(20 MOD}
EVAPOTRANSPMATON
TRANSPORT AND
TREATMENT OF
SEWAGE (I5-2O M6D)
GROUND WATER
RESERVOIR a
SOIL MOISTURE
EVAPORATION
IRRIGATION
WET
WEATHER
PONDS.LAKES
AND
RESERVOIRS
SURFACE RUNOFF
PANTHER BRANCH
OR
SPRING. CREEK
Figure 7. Schematic water balance for The Woodlands.
20
-------
normal approach, benefits of the "natural drainage" approach in
managing stormwater runoff are as follows: (a) maximizes re-
charge, (b) minimizes runoff, (c) minimizes erosion and siltation
problems, (d) minimizes vegetation removal and (e) minimizes cost
of the drainage system. The "natural drainage" concept is essen-
tially a recharge and containment approach to managing stormwater
runoff and is designed to achieve the following goals: (a) re-
duce legal entanglements resulting from excessive runoff leaving
the property, (b) sustain existing plant life by retention of a
stable, high water table, (c) sustain planned perennial lakes and
(d) minimize clearing and grading costs for swale and storm sewer
trenching.
CLC LAKE SYSTEM
The man-made lake system at The Woodlands Commercial, Lei-
sure and Conference Center (CLC) was filled during March 1974.
The system, known as Harrison Lake, is comprised of two lakes
separated by a decorative waterfall. The upstream smaller lake,
designated Lake B, is constant volume. In dry weather there is
no streamflow into the lakes, and the water level of Lake B is
maintained by recirculating water from the lower lake, Lake A,
or by the inflow of tertiary treated sewage. Sewage flow from
the first phase of The Woodlands community is not uet a major
source of lake water, however, projected sewage flow from Phase
I is 6 mgd (.26 m3/sec). During wet weather, Lake B is designed
to receive stormwater runoff and serve as a sedimentation basin
for the 337 acre (136.4 ha) watershed under construction at this
time. Upon completion, it will include an 18 hole golf course
meandering through a residential area.
Lake A is a variable volume lake to be used for non-contact
recreation. Lake dishcarge is controlled by an outlet box at an
elevation of 121.8 ft (37.1 m) above sea level. Water from Lake A
irrigates The Woodlands Golf Course bordering the lake's eastern
and northern shores. In dry weather, the water level drops due
to evaporation and grounwater is pumped into the lake to compen-
sate. A clay bottom serves as an effective seal so that water is
not lost to groundwater recharge.
HUNTING BAYOU WATERSHED
The Hunting Bayou watershed is located in Northeast Houston
near the intersection of Highways 59 and 610. The 1,976 acre (800 ha)
watershed is characterized by low land slopes and impermeable
soils with high clay content. Primary drainage channels are
trapezoidal in shape and lined with vegetation which varies in
dnesity from moderate to very heavy depending upon the season
and maintenance schedules. Typical channel conditions are
showns in Figure 8. The majority of the secondary drainage is
21
-------
Figure 8. Photographs of typical channel
conditions for Hunting Bayou.
22
-------
provided by roadside, grass-lined swales (Figure 9) comparable
to the drainage design at The Woodlands. A fourth of the area
is drained by storm sewers. There are no known effluents en-
tering the drainage system, however, point sources are probable
due to the age of the residential districts (illicit sanitary
sewer connections to the storm system and the presence of in-
dustrial influences). The area is poorly maintained and stream
channels are sometimes used as dumping areas for waste materials
such as oil and grease, old tires, and other refuse.
Land use in the watershed is mixed (Figure 10) with resi-
dential areas comprising the largest segment. Residences are
mostly single family dwellings of low value. Table 3 gives de-
mographic information regarding the indigent population.
Industrial activity in the watershed includes meat packing
and rendering plants, wrecking yards and mechanical contractors.
Various commercial establishments in the watershed support the
residential population. Construction in the watershed is cen-
tered around completion of Interstate Highway 610.
WESTBURY SQUARE WATERSHED
The "natural drainage" system utilized by The Woodlands is
an innovative method for controlling stormwater runoff and pre-
venting water quality deterioration. A Houston watershed with a
conventional drainage system was selected as a comparative study
site. The residential land use of Westbury is similar to that
being constructed at The Woodlands.
The 210 acre (85 ha) watershed is located in Southwest
Houston and is comprised exclusively of single-family residen-
tial dwellings. No commercial or industrial influences are
present in this area. The watershed is completely developed and
contains no construction sites, empty lots, and no undeveloped
land. Figure 11 and Table 4 provide watershed information.
Demographic information from the 1970 census presents the
area as upper-middle class, median annual income of $19,000.
Population density is approximately 11 persons/acre (27 persons/
ha), or 4 persons/household.
The separate stormwater drainage system consists of lateral
drainage provided by concrete pipe, 18 to 54 in. (45.7 to 137.2
cm) diameter, connecting with a main collecting channel at road-
way intersections. The channel is an open grasslined ditch,
often choked with vegetation in the summer months. The ditch
passes through culverts beneath roadways, and at these points
ponding occurs upstream and downstream providing slight storage
of runoff and a reduction of flow velocity. Low runoff coeffi-
cients can be expected due to the low land slope of only 0.8%.
Impervious cover is estimated at 35.4%. No dry weather flow is
present in this watershed.
23
-------
Figure 9. Photographs of typical secondary drainage
system for Hunting Bayou.
24
-------
(U
CD
S-l
O
-P
(0
rd
ffl
in
o
0) .
to
c 3 ii
•rt
-P T3 -P
C C IH
3 rt
0)
5-1
3
tr>
•H
Pn
01
§
to
O
o
o
o
n
3
in
c
o
U
W
J
m
<
EH
8.
C —I to
ro —
•H >, O
H 4J i-t E
Q C -P O
o to u
"O O C
•H S -H
to —
&
a
a
1 01
OTJ
3 3
•P 0
to C
e -H
o — •
u c O 0
0) -H i-l
13 4J *O
C
D
01
0
o
to
3
0
•f-t
a
E
M
S*^
-P
•H
U)
C
Q
C
0
•H
10
i— 1
a
0
Cu
c
<0
I
>1
a
e
U
0
o
u
in
c
10
'H
•a
o
£
CI
to
0
to
c
o
to
>H
ID
Cu
O
£
D
ID
C
(U
U
o
•C l<
O 01
IS
10
4J r-
10 o
Q rxi
25
-------
TABLE 4. WESTBURY SQUARE WATERSHED
CHARACTERISTICS
Total Drainage Area
Single Family Residential Areas
Impervious Cover
•Population Density
Family Median Income
Median School Years Completed
Home Values - Mean
Persons/Household
Rooms/Housing Unit
210 acres (518ha)
100%
35.4%
7,040 persons/mile
$19, 000/year
14.7 years
$29,000.00
3.8 persons
7.5 Rooms
* Data derived from 1970 Census: Census tracts 427 and
428 are averaged.
WILLOWBEND
ULUt
N
Figure 11.
Westbury Square watershed
(scalelin= 1000 ft).
1 ac = .405 ha
1 mi = 1.6 km
26
-------
Soils in the area are predominately dark clays and loams,
characterized by low permeability and high available water
capacity. The soils tend to be mildly alkaline or neutral.
Rainfall in the area averages 39.5 in/yr (1.3/yr).
Table 5 is a comparison of characteristics from the six
major watersheds monitored.
27
-------
CO
u
H
EH
CO
H
Pi
W
EH
U
rr
l-M
u
p
w
ffi
Pi
w
<
p*
in
H
1— 1
CQ
EH
t7
D
5
tn
0)
3
Cr
C 3
•rt O
4-1 >
C 10
3 03
M
CJ
?«
IB
r^ 0)
O 10
Q
2
1C
•a
c <
rH ID
•a A;
0 *0
o »J
s
in
'D
C
CO O
i— 1 rH
•0 1
o &i
S
en
^
(0 O
rH m
O (X
1^
..
•O CD
a -H
w
•o
D Si
U; -rl
I-l rH
ro
Vi 01
0) M
p, <|
E
H
tffi
O
1 0
1 rH
CO
OJ
0
o
0
Q
E
Vi
o -a
4J 0)
M Vi
d
ra s
o; o)
dw
OO
O
.
OP
rH
OP
ro
UP
dP
0
OJ
•
<*>
1C
^
01
Qj
rH
to
•o
£
5
II
rH
10
JJ
•rl E 3 01
Bi E "0 T3
CD C C C
Bi CJ I-H S
- t>*>
C -i
1 O rH to
•H ro c
CD -P -H O
rH CJ 4J -rl
J3 3 C -P
a VI 0) 18
•H JJ -D 0)
i-J tl) -rl VI
(0 C in cj
> o 01 a>
U K Oi
C * 02S rH
1 O -H (0
•H ro —I C
0) +J -H It O
r-l CJ -P -rl -rl
X! 3 C 0 -P
10 Vi 0) Vl IB
-H 4-1 ^O 0) 0)
Vi U) -H g Vl
ro c in E u
> O 0) O 0)
u a cj o;
1C
Cb
Vl •
•^ jy, — ^^ J
C1 QJ
rH DJ
Ifl II O II
VI rH
3 "O 0) -O
•P C > C
fO (0 QJ (0
Z rH Q rH
Bl 1
CD CJ
H • 4J 3 •
O <* C VI *>
UH O QJ -p O
C^ E B) rH
rH Q. C
(0 II O 0 II
VI rH CJ
3 "D 0) C
4J C > TJ O
fO (0 01 C -H
Z rH Q 10 4J
C
0
•H
OJ 4-1
in ic
D CJ
•H
•O in
C -rl
to in
J B)
10
rH
CJ
in 0)
0! rH Vl
10 rH O
rH (0
O It \
>i tn
QJ -PC
rH -rH O
T3 Bl B)
•C C Vl
•H Q) 0)
s Q a.
1
ro ii
O -H >,
E 3 4->
O QJ-H
o o in
c a. c •
M 0) 0)
V- U
o) c ro
3 o \
O -H OJ
rj -P rH
' .
D O) Bl
> D. B) •
0) O. a w
CD &, rt; ui C
Ti 3 M a; o
U O -H
T3 •• C -u
CD -P CD QJ fO
C C >H T3 (U
C CD TJ -H ^
ra E *o tn O
i-l tX-H QJ QJ
i-U O £ V-( r-i
-J
U-4
•H
TJ
•M
3:
C)
-H
ro
1— (
•rt
S
.,
"J
x: tn
a 01
10 4J
M 0
Q
1
crt
(-<
ri— 1
O
•*.
II
o
rH
23
-------
SECTION 5
SAMPLING AND MONITORING PROGRAMS
This chapter summarizes the field sampling and monitoring
programs initiated in September 1973 and concluded in April
1976. The programs were designed to meet the following specific
objectives:
1. Establish a sampling program for Bear Branch, Panther
Branch and Spring Creek to determine the effects of
urbanization on receiving waters.
2. Define the temporal characteristics of stormwater run-
off quality and changes due to urbanization.
The sampling programs were divided into two distinct seg-
ments. First, surface waters within The Woodlands were sampled
during periods of no overland runoff (dry weather flow condi-
tions) for the occurrence and concentration of selected physical
and chemical constituents. These measurements comprised a low-
flow data bank to determine (a) the effect of urbanization on
the water resources of The Woodlands, and (b) the water quality
criteria of the lakes within The Woodlands to insure their use
for recreational and aesthetic purposes. Sampling points were
located for comparison of undeveloped and developing areas with-
in The Woodlands. Sampling sites downstream provided data for
establishing the impact of The Woodlands development upon water
quality of the receiving body (Spring Creek).
The second phase of the sampling program concerned itself
with quantifying quality of overland runoff and its subsequent
impact on water resources of The Woodlands. Sampling sites were
located within the Woodlands and at two developed watersheds,
Hunting Bayou and Westbury Square. The storm sampling program
involved development of pollutographs for at least 25 different
water quality parameters per storm event. Not every storm event
was evaluated but, rather, selected storms were monitored such
that various seasonal and hydrologic conditions were defined.
A hydrologic network was established within the study
areas, centered around continuous discharge recording stations
operated and maintained by the Water Resources Division,
U. S. Geological Survey (USGS), Houston, Texas, at the request
29
-------
of The Woodlands Development Corporation. The purpose
of this network was to accurately delineate movement of water,
especially surface flows. For this purpose, a weather station,
rain gauges, streamflow stations and groundwater observation
wells were established.
DRY WEATHER FLOW SAMPLING
The dry weather sampling program resulted in a data bank
of chemical, hydrological, and physical characteristics of sur-
face water resources at The Woodlands. All sampling sites are
located in or near the Panther Branch watershed. As shown in
Figure 12, these sites were located throughout the watershed
from headwaters to receiving waters. Dry weather flow samples
were collected only within The Woodlands watershed.
Sampling frequency at a particular site was determined by
its importance in relation to monitoring effects of development.
For example, Lake Harrison was sampled frequently to aid in the
lake management study, while sites in Panther Branch headwaters
were seldom sampled.
Quality determinations were conducted on water samples and
included: temperature, dissolved oxygen (DO), pH, turbidity,
total suspended solids (TSS), soluble chemical oxygen demand (COD),
total COD, soluble organic carbon (SOC), total Kjeldahl nitrogen
(TKN), total phosphorous (TP), orthophosphate (ortho P), ammonia
(NH2), nitrite (N02)/ nitrate (N03), biochemical oxygen demand
(BOD), and specific conductance. Flow measurements were made
where feasible, in addition to four time-of-travel measurements,
covering the entire reach of Panther Branch.
STORM EVENT PROGRAM
The stormwater monitoring program used a hydrologic network
combined with intensive sampling to characterize runoff quality.
Runoff samples were collected at 6 sites, four located within
The Woodlands, a fifth in Hunting Bayou, and a sixth at Westbury.
The Woodlands
The Woodlands sampling locations and their designations are
listed below:
1. Panther Branch at the Confluence
with Bear Branch P-10
2. Panther Branch at Sawdust Road P-30
3. Outflow of Lake Harrison Lake A
4. Inflow of Lake Harrison Lake B
A USGS gauging station measures streamflow at each of these sites.
30
-------
W
CD
tn
P
nJ
cn
.s-g
(tJ ,£
M CQ
n
-O Q)
C -P
ftf (U
^
CO
0) ,C
4J O
--1 fi
to nJ
M
tnpq
C
•H M
0)
e-5
nJ C
w rO
cu
m
O 0)
c5
O
•H C
4J .H
rtJ ,C
O -P
O -r»
(U
M
P
tn
•H
S
31
-------
Station P-10 is located on Panther Branch 200 yards (182.9m)
downstream of the confluence with Bear Branch (see Figure 12) . The
•watershed at this point measures 16,050 acres (6,495 ha) of
predominately undeveloped pine-oak forest. Data collected from
P-10 represents runoff from a natural area devoid of urban in-
fluences and, when compared to other sites, serves to determine
the effects of urban influences on runoff quality. U.S.G.S.
established the streamflow gauging station at P-10 in July 1974.
P-30 is located downstream on Panther Branch and includes
the P-10 drainage area in its 21,606 acre (8,744 ha) watershed
(Figure 12). The sampling site is in an advantageous position
for monitoring runoff quality from development areas immediately
upstream. U.S.G.S. has operated a streamflow gauging and
monthly sampling station at this site since April 1972. Because
construction at The Woodlands was a continuing process, water
quality changes were expected.
Two stormwater sampling stations, shown in Figure 13, were
located at Lake Harrison to measure stormwater inflow and out-
flow. The upper station, situated in a swale at the head of
Lake B, sampled runoff from the major source of stormwater flow
into the lake system. During dry weather no flow occurs in the
swale. The 337 acre (136 ha) watershed was under extensive con-
struction at the time of study and had little completed develop-
ment. The Lake A gauging station, adjacent to the lake outflow
box, was sampled to assess effect of detention of runoff quality.
The drainage area at Lake Harrison outflow is 483 acres (196
ha) .
Rainfall data was available at the five locations in or near
The Woodlands (see Figure 12). The precipitation data collected
at these sites determined the average hyetograph for the Panther
Branch watershed during each storm event which defined the in-
tensity, duration, time period and total amount of rainfall.
The rainfall network provided data on the spatial and temporal
characteristics of rainfall within the study basin.
Runoff samples were analyzed for the same parameters listed
in the dry weather flow program. Chemical data was used in con-
junction with flow data to derive mass flow relationships for
each constituent during each storm.
Hunting Bayou
The stormwater sampling program at Hunting Bayou was simi-
lar to The Woodlands, except that only one sampling site was
monitored. Stormwater samples were collected from the U.S.G.S.
Hunting Bayou at Falls Street gauging station (see Figure 10).
precipitation data was available from two recorders located south
of the drainage basin.
32
-------
r
AREA SHOWN IN DETAIL
4.1 ACRES
20 ACRE FEET
NORMAL WATER LEVEL
EL. 128
AVERAGE DEPTH
6 FEET
12.5 ACRES
90 ACRE FEET
NORMAL WATER LEVEL
EL. 122
AVERAGE DEPTH
8 FEET
LOCATIONS OF U. S. G. S. GAUGING STATIONS
0 400' 800'
SCALE
Figure 13. The Woodlands man-made lake system with
locations of stormwater monitoring sites.
1 ft = .305 m
1 ac = .405 ha
33
-------
Westbury
Runoff from this watershed was sampled above the Atwell St,
bridge crossing the primary drainage channel (see Figure 11).
Flow measurements were determined using a flow meter and preci-
pitation measurements were made using a portable volumetric
gage located at the sampling site.
34
-------
SECTION 6
EXPERIMENTAL METHODS AND PROCEDURES
The sampling and analytical program is described in detail
in this section to enable the reader to evaluate data reliability.
The project team developed practical, effective methods for
sampling stormwater runoff simultaneously at multiple sites.
The methods described will be valuable to anyone planning such a
program. Finally, commercial equipment utilized represents a
small, but representative, sample of available instrumentation
and their inherent problems.
MAJOR SAMPLING STATIONS
The location of the major sampling stations has been dis-
cussed in Section 5. Each station was equipped to varying de-
grees for sampling and monitoring as indicated in Table 6. p-30
and Lake A had a llOv AC power source which simplified the opera-
tion of samplers and monitors. The manometric water level sen-
sors and recorders were maintained by the U.S.G.S. through a
grant from The Woodlands Development Corporation. All flow data
was compiled by the U.S.G.S. and transmitted to Rice University
for integration with other stormwater data.
SAMPLING
Equipment
Samplers—
A problem inherent in all automatic samplers used, except
Sigmamotor WM-1-24-R, was no provision for effective sample pre-
servation. The U.S.G.S. samplers left the sample compartment
exposed allowing access to insects and debris. The Manning
S-4000 sampler provided a center compartment for filling with
ice, but a test failed to cool tap water in the bottles below 59° F
(15° C). Filling the compartment with dry ice would change the
chemical nature of the samples (e.g., pH). Sigmamotor WM-4-24-R
sampler provided space for ice also, but cooling was insuffi-
cient.
The measured flow in sample tubing from Sigmamotor WM-4-24
with 8 ft (2.44 m) head was 0. 3 ft per second (. 09 m/sec) , less than
sufficient velocity to maintain suspension of particles of 0.09 y
and specific gravity of 2.65 (16). Replacement of the pumphead
35
-------
TABLE 6. FIELD EQUIPMENT OPERATED AND/OR MAINTAINED
BY RICE UNIVERSITY AND USED FOR SAMPLING
AND MONITORING
Site
P-10
near p-10
P-10 and P-30
P-10 and P-30
P-30
P-30
P-30
Lake B
Lake B and Lake A
Lake A
Lake A
P-10, P-30, Lake B,
Lake A
Hunting Bayou
Westbury
P-30, Lake A
Equipment
Discrete sampler (Sigmamotor #WM-4-24)
Rain gage (Belfort #5-780)
U.S.G.S. V-notch weir with manometric
water level sensor (Stevens)
Dissolved oxygen/temperature analyzer
(Delta #3610) with remote stirrer
(Delta #1010-10)
pH Analyzer (Delta #3412-01)
Suspended Solids (Turbidity) Analyzer
(Ecoloqic Instrument Corp., Model 204)
U.S.G.S. discrete pumping sampler
(#PS-69, Federal Interagency
Sedimentation project)
Discrete sampler (Manning #S-4000)
U.S.G.S. Mailbox Gage (Float type with
Stilling Well)
Discrete Sampler (Sigmamotor #WM-1-24R)
Weather Station (Weather Measure #M701
and #W123) including rain gage (Weather
Measure #P501)
Float switch and latching relay for
automatically starting samplers with
rise in stage height
Manual sampling, discharge from
U.S.G.S. rating curve
Manual sampling, discharge from Rice
University rating curve
llOv AC which powered samplers and
monitors
36
-------
-assembly and tygon tubing did not improve the pumping rate. For
this reason, a pumphead (Masterflex #7018-20) for 0.44 in (1.1 cm)
O.D. tygon tubing was adapted to replace the standard pumphead
assembly. This required filing down plastic nipples on the pump
housing, drilling and tapping two holes in the sampler case and
fabricating an adapter sleeve between pump and motor. The resul-
ting fill time of 3 minutes and a pumping velocity of 1.2 ft per
second (.37 m/sec) with the original 1/8 in (0.3 cm) I.D. tygon
sampling tubing was satisfactory for sediment suspensions at The
Woodlands. A test of the sampler on Woodlands water with a 9 ft
(2.7 m) head resulted in samples with turbidities within 2 FTU
of grab samples (220 FTU). The Manning S-400 sample velocity
was over 3 ft per second (0.92 m/sec).
Although operating characteristics of the Manning sampler
were the best of the battery-operated samplers, it presented the
worst breakdown record. Problems included (1) defective counter
integrated circuit in the clock circuit, (2) defective capacitive
probe collar, and (3) slipping stepping motor collar (sample spout),
The integrated circuit was replaced, the capacitive probe was re-
placed with a probe of different design, and the stepping motor
collar was secured with a set screw.
Float switches were fabricated which were used to activate
the samplers at a predetermined rise in stream level. The float
switch consisted of a reed switch activated by a magnet mounted
on a disc of styrofoam, floating freely in a short length of 2 in
(5.1 cm) diameter plexiglass tubing. The reed switch operated a
latching relay in series with the battery and sampler. Once on,
the sampler remained on even when the stream level dropped, opening
the float switch. This ensured a reference back to the first
sample initiating the operation.
On two separate occasions, discharged batteries connected to
an 8 amp battery charger failed to recharge. This was not obvi-
ous until the sampler went dead after a few samples. Subsequently,
specific gravity of the electrolyte was checked in recharged
batteries. Battery problems were also encountered v/ith the USGS
sampler. Weak batteries were not able to operate pumps, however
current drain under this low voltage condition blew fuses in one
circuit of the sampler. A battery charger installed to operate
periodically required installation of an automatic cutout switch
to prevent charger overload when the sampler was operating.
Monitors—
Fouling on monitor probes and drift in calibration were the
most prevalent problems encountered with the monitors. Cleaning
and recalibration were necessary before each storm. Calibration
of the turbidity monitor (Ecologic #204) presented unwarranted
problems. Two turbidity ranges were obtainable on the monitor,
but only one range had zero and calibration controls on the front
37
-------
panel. Internal potentiometers mounted on the printed circuit
board controlled the other range independently.
Rainwater Quality Monitor—
A rainwater sampler was constructed to collect samples at a
rate proportional to rainfall intensity. A tipping bucket rain
gage (one/pulse/.01 in or .025 cm) was interfaced with a chromato-
graphic fraction collector (18 x 25 ml test tubes). The rain
gage also activated an event marker on the recording of accumu-
lated rainfall. A polyethylene funnel (6 in or 15.2 cm diameter)
collected rainwater and discharged it into the test tubes. A
spring loaded cover protected the funnel from dust accumulation
during dry weather. The first pulse from the rain gage opened
the dust cover.
Sampling Procedures
Frequency—
Both temporal and spatial concentration variations are sig-
nificant in stream analysis. Spatial variations are caused by
floating material such as oil, sawdust, and floating biota, or by
concentration gradients caused by incomplete mixing (e.g., par-
ticle distribution in laminar flow or benthic evolution of
anaerobic metabolic products). Consequently, point sampling may
not yield concentrations representative of average stream qual-
ity. Multi-point, cross-sectional sampling is desirable, al-
though a single sampling point may prove adequate. Results
from a cross-sectional sampling study are discussed later in
this section. Single sampling points were used in this study
because of the relative costs and benefits derived from multi-
point sampling.
Uncertainty caused by temporal variations can be reduced by
increasing sampling frequency. In stormwater runoff, temporal
changes in constituents correspond approximately to changes in
discharge. Sampling frequency in this study, a function of
stream stage height, yielded sufficient resolution to detail
pollutograph peaks. Based on project experience, empirical
guidelines have been established for choosing a sampling fre-
quency in any watershed. Starting with a selection of typical
storm hyetographs and hydrographs, time to cresting, maximum dis-
charge and total accumulated rainfall were studied. Over a
three year period, sixty percent of the storm events at one test
site had double peaks at 4 to 6 hours and at 20 to 26 hours.
Eighty-five percent of the storms showed maximum discharge
linearly correlated with total rainfall. Time to cresting ap-
pears to be primarily a function of watershed hydrology while
maximum discharge can be estimated from accumulated rainfall.
The average sampling frequency for a resolution of .25 ft (76
cm) rise in stream height would be given by:
38
-------
. _ estimated maximum stage height (ft)
Samples/hour - 0>25 (ft x time to max. stage height (hr)
In this project, a minimum sampling rate of one sample per
2 hours and a maximum of one sample per 15 minutes were set for
the hydrograph rise. On the falling limb of the hydrograph, a
frequency half that of the rising limb gives sufficient resolu-
tion. This was reduced to one quarter the initial frequency in
the last third of the sampling period. At P-10, the U.S.G.S.
sampler was set to sample at one hour intervals or every quarter
foot rise in stream level depending on the expected maximum dis-
charge. Sampling frequency was halved during the falling limb of
the hydrograph. Tine interval sampling was used when low runoff
levels were expected. Samplers at P-10 and Lake A were set to
30 minute sampling intervals for the stage rise and every hour or
every other hour for the fall. Hunting Bayou, Westbury, and Lake
B were sampled at 15 minute intervals on the rise and half hour
to hour intervals as streamflow decreased.
Sample processing—
Samples were in the field for a maximum period of 8 hours
and were not refrigerated. Temperature within the samplers re-
mained near ambient at all times.
Sample transportation required approximately one hour to the
laboratory where each sample was processed according to the flow
chart indicated in Figure 14. Non-filterable solids were deter-
mined by filtration through 0.45y Millipore HATF membrane fil-
ters, preserved with 40 mg of mercuric chloride mg/1 of solu-
tion and refrigerated at 39° C (4°C). Preserved, filtered and un-
filtered samples were stored in large screw cap culture tubes
(Kimax, 195 x 25 mm). Samples for bacteriological analyses were
refrigerated for at least 6-10 hours before analysis.
Sample
50 ml««—4—«••••10 Oral Measure Bact'erio-
| I Conductivity logical
Preserved and Filtered I Aliquot
refrigerated I Measure pH
Measure
Filter Filtrate Turbidity
dried and preserved and
weighed refrigerated
Figure 14. Flow chart for storm samples.
39
-------
Sampling procedures—
At Westbury and Hunting Bayou, and when only one site was
sampled at The Woodlands, manual sampling was employed. A field
crew consisting of 2 research assistants arrived on site before
rain started. An initial grab sample was taken at onset of rain
and subsequent samples taken according to the schedule outlined
above. Sampling from a bridge consisted of lowering a bucket
from the downstream rail. A water quality sample, and BOD bottle
for Winkler DO determination, were dipped from the bucket. The
cost advantages per site of automatic sampling over manual samp-
ling by technicians can be seen in Table 7. The cost advantage
is greater as storm frequency and project period increase. Uni-
formity is also better with automatic sampling.
For multi-site sampling at The Woodlands, manual collection
was not feasible. However, if the sampler servicing crew found
a sampler inoperable, one member would remain to collect samples
manually while the second member reported sampler failure. Re-
pair was then made in the field or manual sampling was continued
at that site as necessary.
Sampling crews were responsible for collecting filled bot-
tles from the automatic samplers, replacing with clean, empty
bottles and resetting the samplers to initial conditions. Bat-
teries were replaced after 3 sets of 24 bottles had been col-
lected. Crews were instructed to record data such as time of
bottle change, stage height, bottle numbers, etc.
TABLE 7. ANNUAL STORM SAMPLING COSTS IN DOLLARS
PER SITE FOR TWELVE STORMS. ASSUME A
TYPICAL STORM IS SAMPLED FOR 48 HOURS.
Automated Manual
Sampler ^
2
Extra Battery
Battery Charger^
2
Chain, Lock, Etc.
Equipment Maintenance (labor)
Other Materials (maintenance)
Sampling '
Overhead (60% of salaries)
1400
30
25
50
250
150
600
510
3015
150
2880
1728
4758
1 2
Assume technician at $5.00/hour Capital expenditures
Does not include costs and time of transportation to site
40
-------
LOGISTICS OF STORMWATER SAMPLING, MONITORING AND ANALYSIS
Preparing for a Storm
Laboratory—
Completion of all analyses and data recording for the last
storm event was mandatory before preparation for a new storm.
Thus, inquiries could be made regarding ambiguous situations be-
fore memories lapsed, and confusion between storms was avoided.
All equipment was organized and supplies restocked. Three sets
of 500 ml sample bottles (24 per set) for each Rice sampler and
320 screwcap culture tubes were cleaned and boxed. Millipore HATF
membrance filters were dried at 140° F (60° C), weighed and stored
in numbered plastic culture dishes. Reagents for the Auto-
analyzer, TKN, COD, and TP analyses, and buffers were restocked
as needed. Stock standards were replaced every 4 months and
tested against the new solutions as a check for continuity in
accuracy. Working standards were prepared from the stock stand-
ards on the day used.
Field Equipment—
Batteries for the samplers were charged to capacity and
specific gravity was tested. Dry cell batteries were replaced
in all field monitors. Field checks were made of all samplers
and monitors to ensure proper operation. Float switches were
manually submerged to test automatic sampler functioning. The
pH monitor was calibrated against pH 7 and pH 10 buffers, the
DO monitors were calibrated against air saturated water, and
the turbidity monitor against a Formazin standard. Float switch-
ers were set to trip with a 1.2 in (3 cm) rise in stream level.
Storm Watch—
Three methods were employed to monitor storm activity at The
Woodlands, located 40 mi (64 km) from Rice University.
Weather data, including hourly updated radar reports, were
used to plot movements of frontal systems and to estimate times
for arrival of storms. Telephone inquiries to The Woodlands
Security Office, open 24 hours a day, gave onsite information
about weather conditions. Finally, research personnel travelled
to the site when rain was imminent and reported the arrival of
sufficient rain for sampling. They also constituted the first
sampling team shift. Westbury and Hunting Bayou were close
enough (10 minutes by car) that local conditions dictated initia-
tion of sampling.
Transportat ion
Although the automatic samplers could generally be left un-
attended for 12 to 24 hours, samples analyzed for bacterial para-
meters and certain chemical species undergo degradation in this
41
-------
period. Consequently, crews were sent every 6 hours to pick up
samples. Two hours for round trip and four hours for sample
collection were adequate. Shifts were started when a storm was
imminent.
Manual sampling provided more difficult logistic problems.
Two vehicles were necessary with overlapping shifts. Crews were
assigned 8 hour shifts and left every 6 hours. This allowed for
a 2 hour overlap for roundtrip travel and 6 hours for field duty.
The first shift departed in the field equipment truck and the
second vehicle was used as transportation to and from this
truck, eliminating transfer of equipment. Transportation costs
for a roundtrip from Rice University to The Woodlands was ap-
proximately $15.00 (assuming $0.15/mileor $0.09/km) if all
four stations were visited.
Termination of Sampling
Sampling was terminated when discharge reached 0.1 of the
maximum value or when the slope of the hydrograph was zero. This
generally occurred from 24 to 72 hours after onset of the storm.
Thus, it was possible that up to 12 field crews and 12 lab crews
might be needed during a storm. This was accomplished by crews
which were available for multiple shifts.
Laboratory Procedure
Sample processing—
Laboratory crews, consisting of two people each, were sche-
duled to begin work at the estimated arrival time of the field
crews. Samples were processed immediately to avoid degradation
and to eliminate confusion from a backlog of samples. When the
samples arrived, a staff member, always one of the lab crew, made
decisions on which samples to process based on the hydrograph.
Frequently, frequency of sampling was greater than necessary to
adequately describe the hydrograph. An aliquot of each chosen
sample was transferred to a sterile sample bottle for bacterio-
logical examination. Each sample was shaken thoroughly to dis-
lodge settled sediments before any transfer or measurement.
Each step was laboriously described since even rudimentary pre-
cautions, obvious to staff members, were often overlooked by
assistants.
A portion of each sample was then filtered through pre-
weighed membrane filters, the transfer and filtering apparatus
being "poisoned" with a small volume of sample or filtered sam-
ple before filtering the bulk of the aliquot. Mercuric chloride
(2 g/100 ml demineralized water) was added, one drop per 25 ml
sample, to give a final concentration of 40 mg/1. Preserved
samples were stored in screwcap culture tubes at 39.2°F (4°C)
until analysis. Conductivity, pH and turbidity were measured on the
remaining sample. For algal bioassays, remaining sample volumes
42
-------
were composited into a series of larger sequential samples. For
a typical storm, processing required approximately four hours
per shift.
Analysis—
Research assistants, working regular 4 hour shifts, assisted
with chemical analyses. Two parameters were determined on the
Autoanalyzer each day with a maximum of 160 storm samples.
Rather than monitor the Autoanalyzer continually for adjustment
of baseline and gain, the parameters were set to approximate
values and "blank-standard-blank" series run every ten samples.
These served as references for calculating concentrations later.
Changeover from one parameter module to another took approximate-
ly one hour including equilibration. Analyses for an entire
storm were completed within 7 days.
Data Reduction and Recording—
Each person performing an analysis was responsible for cal-
culating concentrations from the raw data. Values were recorded
in laboratory notebooks and composite data sheets by the data
compiler. Data sheets were spot-checked by the staff and placed
in a folder with all primary data sources such as field sheets,
"autoanalyzer" charts and total organic carbon (TOG) recorder
charts. Data was then punched on IBM computer cards and a final
data printout with dates, site description, times, discharge, and
all parameters was obtained. Table 8 presents definitions used
in analysis of hydrologic data. A large part of the hydrological
data bank was supplied by the USGS.
OTHER SAMPLING
Dry Weather Flow
Data was collected during dry weather periods for background
levels of water quantity and quality at The Woodlands. Sampling
at eleven dry weather sites was conducted only if five days had
elapsed since any significant stormwater runoff. Preservation,
processing and analysis was identical to that for stormwater
runoff samples. Sampling trips were conducted approximately
once a month over a two year period in order to collect enough
data for statistical analysis. One man-day (8 hours) was suffi-
cient time to visit all sites, collect samples, and conduct field
measurements. A second man-day (8 hours) was necessary for
stream gaging with a pygmy meter to construct or check rating
curves. In either case, about one-third of the time was spent in
travel. Samples returned to the lab were filtered, preserved and
analyzed for pH, conductivity and turbidity, involving about 2
man-hours of work. Chemical analyses were completed within a
week using methods described in the following section.
43
-------
TABLE 8. HYDROLOGICAL DEFINITIONS AND CALCULATIONS
Storm Event:
Total Rainfall
Duration of
Rainfall;
Rainfall Intensity
Type of Storm
Event:
Antecedent
Rainfall
Condition;
Total Discharge
or Flow;
Runoff:
Peak Flow;
Base Flow or
Groundwater Flow;
Time of Passage;
Time of
Concentration:
Discrete period of rainfall producing
runoff monitored during the study.
Calculated via Thiessen method using total
precipitation data recorded over 2 separate
drainage areas, P-10 and P-30. The Thiessen co-
efficients are calculated using these rain
gauges: Porous Pavement, Egypt, Confluence
and W. G. Jones.
P-30 Drainage, Area (ac) P-10 Drainage Area (ac)
Porous P. - .125 Confluence- .258
Confluence- .328 Jones - .100
Jones - .075 Egypt - .642
Egypt - .472
Shortest time period during which 85% of
total precipitation occurred.
Total rainfall divided by duration.
Described type of rainfall - e.g. one
period of rainfall, 2 periods of rainfall,
etc.
Total precipitation in 1 week period prior
to storm event. Calculate same way as
Total Rainfall (see above).
Volume of storm event hydrograph.
Overland flow volume. Equal to total dis-
charge minus base flow.
Maximum stream discharge.
Calculated by multiplying the instantaneous
discharge prior to the storm event by the
time of passage.
Time from first rise in stream to 0.1 of
the peak flow. If the hydrograph does not
return to this level due to successive
rainfall a hydrograph separation technique
is used to extend the recession limb to 0.1
of the peak. When runoff is minimal and
base flow greater than 0.1 of peak flow,
then the inflection point on the hydrograph
tail is assumed the end of passage.
Time from first rise in stream to peak
flow.
1 ac = .405 ha
44
-------
Groundwater
Water samples were collected from seventeen wells at nine
sites in The Woodlands. The wells consisted of PVC tubing ap-
proximately two inches in diameter driven into the soil at
depths of two to five feet. Measurements were made of water
level and well depth. Samples were tested in the laboratory for
pH and conductivity, preserved with mercuric chloride (40 mg/1),
filtered and refrigerated at 39.26 F (4° C). All analyses were
performed as described below.
ANALYTICAL METHODS
Analytical Techniques
Parameters measured are listed in Table 9* All methods
were approved by EPA or from EPA "Methods for Chemical Analysis
of Water and Wastes" (17).
Conductivity (Industrial Instruments, Inc., #RC 16 B2 Con-
ductivity Bridge), turbidity (Hach, #1860A), and pH (Chemtrix,
#40) were measured concurrently with filtration. Soluble NH.,,
NOj, N02/ PC>4 and COD were measured in filtered samples using a
Technicon "Autoanalyzer II" system. Total COD, SOC, TKN and TP
were measured using EPA methods (17). NH3 resulting from the
Kjeldahl digestion and 0-PO4 resulting from the TP digestion
were measured by the "Autoanalyzer II." All analyses were com-
pleted within one week following a typical storm.
Problems
The EPA Methods (17) recommend filtration of samples through
glass fiber filters to determine non-filterable solids. Filtra-
tion of Woodlands samples through Gelman type A filters left a
turbid filtrate, attesting to the fine particulate and/or col-
loidal nature of the samples. Gelman type E and Whatman type
GFC left similar turbidity in the filtrates. Various membrane
filters with 0.45npore size were tried, and all gave satis-
factorily clear filtrates. However, various degrees of leaching
were noted at levels which would interfere with analyses (Table
10). Lack of adhesion of residue to Nucleopore filters made
weighing difficult. Consequently, the choice was made to use
Millipore HATF filters.
Automated methods (Technicon) for TP, TKN and COD proved
unsatisfactory in many respects. When samples with high TSS
from stormwater runoff were analyzed on the Technicon Auto-
analyzer II with continuous digestor for TP and TKN at the
0-4 mg/1 level, results were inconsistent. Solids coated
45
-------
TABLE 9. WATER QUALITY PARAMETERS FOR STORMWATER
RUNOFF AND LOW FLOW
Always Measured
Turbidity Total Kjeldahl nitrogen (TKN)
Conductivity Total phosphorous (TP)
pH Ammonia (NH3)
Non-filterable solids Nitrate (N03 )
Soluble organic carbon(SOC)Nitrite (NO2)
Total chemical oxygen ortho-phosphate (ortho-P)
demand (COD)
Soluble COD
Sometimes Measured
Biochemical Oxygen Demand (BOD)
Volatile non-filterable
solids
Total solids
Temperature
Dissolved oxygen
TABLE 10. LEACHING IN MEMBRANE FILTERS
Filter
Millipore HAWP
Millipore HATF
Millipore EHWP
Nucleopore
Gelman Glass Fiber
Vol Deionized Water
Filtered (ml)
50
200
50
100
50
Wt Loss
0.5
0.0
1.0
0.0
0.4
(rag)
46
-------
the inside of the digester helix reducing resolution and in-
creasing holdup. The COD test could only be effectively run on
filtered samples. particulate matter caused negative inter-
ference with this sensitive colorimetric method. A second
problem was the inability to match response with the manual COD
method on semi-refractory compounds. COD's of 30-90% of the
manual method were measured by the automated method. Even an
increase in digestor temperature from 293° F (145° C) to 329° F
(165° C) did not improve correlation significantly. The problem
stems from the short digestion period (22 minutes) for the auto-
mated method compared to the 2 hour reflux period for the stan-
dard manual method. Nonetheless, this method was useful in mea-
suring relative changes in COD.
SOURCES OF ERROR
Confidence in the data was strengthened by a series of tests
leading to estimates of precision and accuracy. Experimental
procedures were evaluated for sources of error and four stages
were judged most critical. These were sampling, preservation,
processing, and analysis. Results of these tests for four para-
meters are summarized in Table n, and a discussion follows.
Analytical
Standards, with concentrations representative of natural
water samples, were analyzed in replicates of five for NO3, NH2/
phosphate and COD. Results are reported in Table 11. Analy-
tical methods resulted in almost insignificant errors.
Preservation
Preservation methods were evaluated on high-flow samples
from Panther Branch. One set of samples was filtered in the
field, preserved with mercuric chloride and iced down for trans-
port to the laboratory. A second set of samples was transported
to the laboratory with no special treatment, representing normal
stormwater sample handling. The samples were filtered, pre-
served with mercuric chloride and refrigerated along with the
other samples. The following day both sets of samples were
analyzed for PO4, COD, NH3/ and NO^. A week later the same pre-
served samples were analyzed again (see Table 11). P04 levels
were lower than normal. Consequently, the data is not helpful
in evaluating preservation techniques but indicates a. problem
in preservation of low level P04 samples. COD, 1-103 an<^ ^H2 data
confirmed adequate preservation techniques.
Sampling
Five samples were simultaneously withdrawn from a 9 square
inch (58.1cm2) cross-section in Panther Branch using a multiport
peristaltic pump sampler. The five samples were returned to the
47
-------
TABLE 11. PRECISION, ACCURACY, AND PRESERVATION
COD
Analytical "\
(Single Sample) 1.2%(0.4) 0.6%(0.8) 1.2%(50.) 1.0%(0.1)
Cross-Sectional
(12 Sample Grid 3x4) 10%(.05) 11%(.23) 20%(59.) 44%(.25)
Overall
(Single Sample) 9%(.06) 12%(.22) 1%(54.) 10%(.21)
Field Filtered and
JTJ.COCJ.VCU, JLUCU dlH-l
Run Immediately
Lab Filtered and
Preserved and Run
Immediately
Lab Filtered and
Preserved and Run
One Week Later
.0532
.044
.049
.140
.155
.150
30.1
31.6
30.6
0
0
0
.0043
.OOO3
.0123
Percent refers to a/v x 100. Number in parenthesis is p.
IJL refers to the mean and a refers to the standard deviation
2
Single numbers refer to concentrations in mg/1.
Levels of o-PO4 were lower than normal.
48
-------
laboratory, and filtered and preserved by normal procedures.
Streamflows were characteristic of low-flow conditions. Samples
were refrigerated and analyzed within the week. Results are re-
ported in Table 11 as relative standard deviations (%) and con-
centrations analyzed. There are significant differences due to
sampling.
Cross-sectional Uniformity
To determine variations in parameters due to incomplete mix-
ing in Panther Branch, the multiport peristaltic pump sampler was
set up to take either vertical or horizontal samples over a
cross-section of the stream. Samples were returned to the lab-
oratory, filtered and preserved by normal procedures. Results
are reported in Table 11 as relative standard deviation (%) and
concentrations analyzed. A cross-sectional sampling resulted in
relatively little variation for NC>3 and NEU compared to the
single sample variability. However, this was not the case for
COD and ortho P where cross-sectional sampling led to signifi-
cantly greater relative standard deviations.
49
-------
SECTION 7
RESULTS AND DISCUSSION
DATA SUMMARY
Dry Weather Monitoring
Surface water quality within The Woodlands was determined
during dry weather periods to establish a baseline water qual-
ity. The baseline will be useful for determining (a) the effect
of urbanization on the water resources of The Woodlands over the
next 20 years, and (b) immediate comparisons between stormwater
runoff quality and baseflow water quality. Dry weather sampling
sites are indicated in Figure 12. Dry weather, or low flow, re-
fers to time periods when the stream stage is essentially con-
stant. The time period following a storm event required to
establish low-flow conditions depends on factors such as antece-
dent moisture conditions, time since last storm, rainfall dura-
tion and intensity, and groundwater elevation. In the Panther
Branch watershed, low-flow conditions were normally established
4-8 days after a storm event.
Low-flow water quality data for Panther Branch, Bear Branch
and Spring Creek are presented in Table 12. The headwaters in
the stream system are deficient in inorganic nutrients but signi-
ficant contributions in developing areas increase concentrations
below P-10. The primary nutrient input is from the golf course
immediately upstream of P-30. Organic concentrations are high
(50 mg/1), consisting of relatively non-biodegradable (BOD <
2 mg/1) leachate from decaying vegetation in the forest. COD
dilution occurred downstream and the lowest concentrations were
observed in Spring Creek. TSS changed drastically
as the stream passed through developing areas where construction
activity and borrow pits were located in the floodplain. Low-
flow TSS as high as 1600 mg/1 were observed at P-30.
Storm Events
Data characterizing hydrological, physical, and chemical
aspects of 43 distinct runoff events resulted from 17 selected
rainfall periods with streamflow being sampled simultaneously at
one to four of the monitoring stations established by Rice Uni-
versity (refer to Site Description, Section 4). The number of
50
-------
CO
2
O
H
EH
t*
s
EH
2
w
u
0
u
w
o
2
H
g
^
•l.
U)
rtX
K
W
EH
•T<
g)
$
ft
><
EH
H
t T
M
3
fd
M
<
3»
^
O
&4
1
is
1 — 1
&4
o
OH
f*£
a
D
CO
CM
,_)
w
9
<
rH
C7>,ii
C It
•H a)
ctu
CO
tr>
c
0 -H QJ
1 CVH
ft E co
flj
CO
o
« rH
M I O
CQ ft n
l
Jj C* fti
A\ /)\
Vi/ U/
£ CD 'd
-P £ C
C 4J CO
03 G)
CU *Q
b»
52
0-H QJ
H rH -P
IO. *i-J
M^ ~
ft E co
CO
C U o
fij C H
P-i *ti 1
M ft
** m o
fO H
O O O O
in cn VD n
^0 ^J* ^D ^3
H O H CNJ
O O rH O
m TJ«
o in oo
H 0 O H
O 0 H O
CM
rr en CM ro
vo CM ro en
o o en o
o o o o
r>- in vo
VO CM in rH
O O CO rH
o o o o
o
ft
fl-t O b^i M
EH ft EH 2
in en
rH in
O CO CO
0 O CO
In en
o in
o o r»
o o o
00
vo en
O CM
o o o
o o en
CO
T rr
in **
O CM
o o vo
O O CM
vo vo
o CM in
0 0 •»
0 0 VO
H
CM ro CO
o o a
22^
en
0 t^
rH
CN
H fM
SJ- g
,_)
o
VO CD
<* in
r-
rr vo
rH ^*
O
CM 0
H n
>>
4J
>l •<+
4J U ^
•H -H jj
T3 m u
•H -H 3
J3 U t3
M 0 C
3 a o
EH CO O
CM
en
H
in
in
rH
O
en
H
en
en
H
o
CM
U
O
CO
«*
vo
CM
o
rH
in
o
CM
in
o
o
in
m
in
Q
O
U
rH
ra
•p
o
EH
CO ^
en rH
H
"31 H
r- CM
CO
O 00
O rH
^3*
in
vo r^
0 H
**
r- o
O CM
Q
O
O
0)
H
A
9
rH Q
0 0
CO CQ
r^- t^- CM o
vo vo en VD
rH H
rH
en r- vo
n CM VD H
vo "3* oo in
rH
VD 00 O O
VD -q1 O CM
CM
oo ro ^3*
CM CM en o
vo in r» co
rH
vo
*^* lO ^5 ^4*
vo LO en iH
rH
0)
Cn
M
rt
• A
d, o
e w
a o
CO -P.
4J U
C 3
c ^
m p?
M O
CO 0
(tf -
-------
runoff events monitored at each sampling site were as follows:
Sampling Site Number of Runoff Events
P-30 12
P-10 8
Lake A 8
Lake B 8
Hunting Bayou 5
Westbury Square 2
Hydrological Observations—
A summary of the hydrological data is presented in Table
13. Hydrological parameters are specifically defined in Table
8 (Section 6). Note that Total Streamflow is the sum of Base-
flow and Runoff. The number of storm events monitored was
limited by two drought periods, each of six month duration dur-
ing the project.
Rainwater Quality—
Previous investigations indicate air pollution may contri-
bute to surface water pollution through rainfall and/or dry fall-
out, even to the extent of pollutants traveling via air from in-
dustrial and agricultural regions to be deposited in undeveloped
areas (18, 19). Samples collected in the Houston area and at
The Woodlands assessed relative contributions of rainwater
quality to stream pollution. Results presented in Table 14 in-
dicate a substantial nutrient and COD content in rainwater at
both sites. A difference exists between Houston and The Wood-
lands rainwater in regard to NH3 and NO3 content for the
storms sampled. The rainwater data are compared
to stormwater data in Table 15. A study of air quality at The
Woodlands indicated high levels of hydrocarbons, 7.6 ppm non-
methane hydrocarbons, whose source was attributable to vegeta-
tive emissions. These ambient air hydrocarbons may contribute
to the soluble COD in rainwater. The study also found an ab-
sence of NOX at The Woodlands in contrast to serious NOX air
pollution problems in the Houston urban area (20).
At The Woodlands, rainwater nutrient concentrations were
greater than runoff water, while the opposite relationship pre-
vailed in the urban watershed. Experiments were conducted to
determine the capacity of soils for stormwater nutrient removal.
Four samples of soil from various locations in The Woodlands (see
below) were dried and weighed. The samples were extracted with
demineralized water until no further NHo was measured in the ex-
tract, and then equilibrated with 30 ml portions of 1 mg N/l
(ammonium sulfate). After centrifugation, the supernatant was
analyzed for NHo and then discarded. This was repeated until no
further adsorption was measured. Findings indicate low levels
of NH3 are definitely adsorbed by soils, with the greatest
52
-------
1^1
0)
•H
O
4-1 MH
O 1-1
C 0)
3 0
OS U
- 4->
M-l M-l
k. "u 1
(V c K
Py ao
g a *
1 s
W u
>H ro o
o ?S
§ BJ
<-J H in
g n
r'1 MO
K * ">
CH
!g 03
« &
W 4) 3
W O
2 S£
£ '«
O E 4J
EH n) M-I
rft <" '
uj M )H
ro
• rH ~
2 *°
•H O .-1
f-t IH
-
CO "-IB
'tj (00)
rf; t-i js
^ -S8
fO -H
K
(U
-p
•H
to
0)
4-)
10
a
3£
O
00
in
vo
H
o
ro
in
0
H
ro
rH
fS
O
«>
fN
iH
n
CN
o
o
*f
o
^r
m
m
OJ
rsi
o
CN
0
f>
1
CM
(n
T3
c
ro
T3
0
O
1/18/7^
r-H
10
0
i
rfl
«
tn
C
4-1
C
D
X
3/20/7-i
cs
m
0
n
in
CM
H
O
O
0
(N
•^1
in
r^
0
g
o
S~i
a
m
Oi
c
4-1
c
3
a
3/26/7^
ro
ro r~ OO CMrH COOlCTi CO^CfHiH VDT'*'* r-) ^^
r~ co m r^r^ r-Hroro rorororo inincMCN CNI ^
ro (N rH r- rHCMCM OOOO CNCM
coco-^r^ VDVO '"'o
0 rH O r-- VDCMCO^ t>^j*
rH^r^- m^rincorHin
in
r> CM ro co r^ ^o
^j1 ro *a* oo co H oorof-- oo^fM COH
in in co COCN H o Hr^CMrH voororo roo-
cMroiM HI- ts^HcriaiCN
CM CM CO H CO ^D
rH CM rH
o r^ o o o CM co
• • • •• • ro ooor^ oooo mo
rH CO roCO O CM^CNCM OO^rO CM^O
ro rorsi ^in rH or-rHcM r-~ro
rH H rH (H
H rH
0 „
<^**^D in t^rH OTO rH
vrjWCDO^rHrH r>CM CM^t in
CM rH
in ^ co
o CM ^o inrH r^ o ro^ooo •fl'ooo ^DO
•<)• 0 rH 'TCM rH 0 rHOOO OrHOO •To
a
tn
co LI in
rH rHfOCOCMCOI^ 0> K£)
•^ co o oo roxjrH ooroo- ooro
rH CM rH
O
2
in in ^o CMCM ro^o^o incMrHrH ^oror^r^ rHin
ro ^r ** min ror-jfN r^'-ococo r^^cMCM cor^
O O ro rHrH OOO OOOO CMCNroro Oo
0 O OO O, o, cu 0,0, euxx pjftXi; ftonxx >i
(0 fOfO fOfO fOfOtO
03 U) 01 CO (0 tOr^r? 0) r?>4 W J rj PQ
rcJT3'T3'OrT3 T3 T3 >i
CTCCCCC c C tni4
C(0(0(OrOfO fO (0 C3
4JfdTJTDrO'O 'd 13 -P4J
cooooo o o cm
3OOOOO O O D<1)
03SSSSS S 3 KS
i>r~r~r^r~ i- r~ r~
\ \ \ x. \ x. x. x,
rHoicoinT ro co co
rHCMCNOO rH O O
X.X.X.X.X, X, X, X.
^•^•OCMro ro •<* in
rH rH
•*rm^or-co CM o H
rH iH
o
1>O
S 6
rt LO LO 00
LO "^ O OJ
. «* to O
(N • . .
II II II II
eo.pt/)
o
53
-------
Q)
(continu
•
r^
H
C •
•rl C
10 -H
K -
C C
0) O
TI E
0)
O X
0) 0)
4-1 0)
4-1
c
o
-H
o
MH MH
O M-'
C 0)
3 O
« U
- 4J
M-l MH
•vtH 1
3 U
04 id
b
U
.* 3
10 O
0) rH
Base Flow
acre-f t.
en
u
11 S
in o
id H
m b
Stream-
icre-fta
rH *
10 S
4J 0
0 rH
EH MH
rH
rH OS
id 0)
MH J3
•H C
0)
4J
-H
W
0)
4-)
10
Q
*
r-
ro O»
0s! rH
H
U3 r-H
CN ro
m I-H
0 O
rH O
ro in
O CN
O"\ i — 1
n rf
• 0
CN
CN
^r o
rH
P" fN
LI rH
*H rH.
rH
m co
OO CN
rH rH
Hunting Bayou
Westbury
06/30/75
CM
rH
in ro co co
HCN
ft 00 ro CO
^
in rH r-
rH O O ^T
^r co in
en o o H
OO
rn m i oo
• • I
0 O 1 00
rH
^ O
. .00
0
T (N
rH O
0 0 0 O
0)
in co 10
tN rH Si rH
co a\ o in
0-1 0 -H rH
Q
o
•z
jnjnHH
rH iH
Woodlands P-30
Woodlands P-10
Lake A
Lake B
09/05/75
rH
in in co co
CO CO O O
rH rH rH rH
n in o
IN rH •*
H
ro o co
r*- t> oo
H in H
rH
10 (N rH
O 0 O 1
0001
id
in c\ 10 4J
m o co id
r^ o oo
«H in rH O
co co ro ro
CN rs ro ro
Woodlands P-30
Woodlands P-10
Lake A
Lake B
10/25/75
•*T
iH
SS33
rHrH rHrH
CO rH CO CO
ro
Cft rH O
O rH O in
\o ro in co
(N O 10
iH H
in o co
rH
ro CO
• .00
rH | O
0)
CO CM O CO
rH rH Q
o
ro
Woodlands P-30
Woodlands P-10
Lake A
Lake B
03/07/76
in
rH
in ro r*- r~ oo ro in in
rH-HCNCN CNrHfOrO
O^CTVOOOO inincooo
-HOVOVP OOOO
rH rH rH rH
ev)Lnov£> en i— i o o
^D^^fO VDOrHlO
(N rH rO rH
i — n
OrHr-r^ inminrH
in CM ^p H
in TT ^o ro
rO CN V£J rH rH
CTi ro
ro CN
• • O O • • O O
^ CN
00 rH
^ CO rH rH
oo m o o o o
corn mS^ro
& co \o CN m ^r in
o\ m «* iH
\£>
rH rH
OO p.
rH H
•
rH
0)
01
id
id
MH
MH
o
c
id
o
MH
0
4J
componen
0)
C
• H
U §
-P 3
10 U
rH
samflow is calcu
2 of rainfall as
a Total str
Percentag<
1)
o x B B
• ^f CO O
CM ...
II II II II
Cl O 4-1 W
•H Cfl M-H m
o
i— 1 rH rH
rH
1
II
CO U
1 0)
•H\
ro
CO II
^J1 • OJ
LJ jj v
MH
C 0 UH
•rl id O
0)
4J
54
-------
CO
H
CO
EH
H
D
a
w
H
CO
^
E-"
-p
q
0
3
.p
•'•H
-P
cn
G
O
U
Q
o
u
0
i-H
rj
j
o
CO
ft
o
X!
-p
M
0
00
O
•y
Hr]
2
0
-P
•H
10
G
co
I X
G
cn
I X
G
cn
i ^
C
CO
I X
rH
cn O
VD CN
r-
r-~ o
• *
ro r^
rH
rH • rf
CO
•^r .in
c — 1 [ZJ i — 1
CN cn
in rH
^J* ^D
rH m
0 0
CN • cn
rH CO CO
O • O
S
rH 0
rH CN
rH
vo !"•»
in rH
• •
0 0
m
0
CN rH
m f)
• .
o ft o
in cn
00 rH
CN cn
H 0
f~1 ^J* CO
in
o
H CN
m CN
• •
o p.. o
CO
cn
CN T3
O fd
-P H
cn T3
3 O
0 0
ffi IS
0
U
G ^3
fd G
. . o rd
cn m -H
0 r- m rH
rH \. -H fd
ftcN G U
g rH Cn -H
rd \. -H Cn
cn ex) cn O
rH
KH - IH o
O m O -H
t*~ • PQ
5H \VO rH
0 m r^ 0 0
0 CN \ > X!
g \o 0 -P
p r~ CN H
G \ 0
-CN tn -P
II in -o
r~ m • cn
G \r- C
rH \ II O
TH cn -H
G \.rH ft -P
O r^- \ fd
•H rH D
-P G rH • -H
fd O cn rH
•H G G ft
> >i O O ft
ID 4J -H <
^ w -S Id xt
t3 M C g -P
M 0) fd -H -H
rd > rH X £
t3 -H 13 O
G G O n tn
rd ID o ft 0
-P !S ft -H
cn 0 rd -P
00 cn
II -H XI X! -H
Pn EH U -P
cn rH rrj
-p -P 0 -P
- fd rrj S -CO
*
o
[ —
cn
rH
«.
.
r^
.
J2
».
cn
m
4-1
•rl
rH
o
t3
o
o
0
rH
tn
G
W
^
.
O
G
H
<
rH
rH
rd
ffi
1
0
O
•rH
rH -P -P
\TJ T! Cn G •• G
Cn 0 0 C rd X 0
g -P -P -H U M &H
G 0 0 3 MH ,G
•rH rH rH -H U
rH rH -P G CO •
CQ O O w tn co
0 O O 0 -H T) 0
d -P cn G u
r-H co tn rrj G
rd 0 0 = -P 0
> rH rH -P O G -H
ft ft= GOO
G g g 4-> CO
rd rd fd -P u cn
0 tn cn G C x!
g 0 • -H -P
G G T) CO g H
II -H -H 3 0 fd
rtj nj -M • P^ (U
| X OH Pi CO 2 — W
f . • .
rH CN m -vT
55
-------
TABLE 15. COMPARISON OF RAINWATER AND RUNOFF QUALITY
IN HOUSTON AND AT THE WOODLANDS
(ALL UNITS IN mg/1)
Constituent
soluble COD
NH3
N03
ortho P
Houston
Rain
14
.31
.52
.01
Runoff1
20
2.1
.38
.57
The Woodlands
Rain
15
.22
.31
.04
Runoff
44
.08
.05
.005
Avg. of stormwater data collected at Hunting Bayou
Avg. of stormwater data collected at P-10
adsorption occurring in undisturbed forest soil:
Soil Sample
Golf Course
Roadside
Swale
Woods
Total mg N/q Adsorbed"
0.027
0.022
0.017
0.043
mg N/g dry soil adsorbed in equilibrium with
1 mg N/l solution of ammonium sulfate.
The data suggest that NH3 in rainwater is ad-
sorbed to a large extent in undisturbed forest soils where it is
metabolized by plants or nitrified and possibly denitrified by
soil microorganisms. In an urban area receiving the same am-
monia rainwater load, there is more impervious area resulting in
higher ammonia in the runoff. In addition, soils in an urban
area have a lower
utilization rate.
NH^ adsorption capacity and lower
Of greatest significance is the effect of nutrient wash off
on lake eutrophication at The Woodlands. Ward and King (21)
have shown nitrogen to be the limiting nutrient for algal growth
in The Woodlands lakes during wet weather. As development con-
tinues, increasing nitrogen input from stormwater runoff can be
expected and lake enrichment may result.
56
-------
Groundwater Quality—
Surface water quality at The Woodlands is influenced by in-
filtration from the perched water table. To determine ground-
water quality, samples were collected from 17 wells located with-
in the watershed 3 days after a 2.8 in. (7.1 cm) rainfall. The
wells.consist of 2 in. (5 cm) I.D. PVC pipe driven 2 to 5 ft (0.6
to 1.5 m) into the soil. Results presented in Table 16 indi-
cate The Woodlands groundwater contains greater amounts of
nutrients but less organic material than runoff. Low pH values
are consistent with the naturally acid soil due to the presence
of humic and tannic components, high clay and low carbonate con-
tent.
TABLE 16. COMPARISON OF GROUNDWATER QUALITY AFTER
RAINFALL TO RUNOFF WATER QUALITY, THE WOODLANDS
Constituent
TKN
NH3
Total P
0-PO4
pH
Specific
Conductance
Total COD
SOC
2
Groundwater
2.1
1.4
0.09
0.02
5.1
200
22
10
Runoff3
1.37
0.08
0.06
0.003
5.9
110
59
22
Data in mg/1 except specific conductance as micromhos/cm and
pH in pH units.
2
Mean concentrations of samples collected within 3 days of
2.8 in (7.1 cm) rainfall on April 14, 1975.
Mean concentrations for runoff samples collected at P-10
during April 8, 1975 storm event.
Stormwater Runoff Quality—
Discrete water samples were collected for the purpose of
defining temporal stortnwater quality during a storm. Over 850
stormwater samples were collected and over 12,000 separate water
quality analyses were performed. A summary of the water quality
data is presented in Table 17, including mass load and flow
weighted mean concentrations for each parameter. Mean concen-
trations are calculated from discrete sample results weighted
57
-------
according to instantaneous stream discharge at the time of col-
lection. Mass load represents total amount of constituent pass-
ing the monitoring site during the runoff event. Mass load is
incalculable where no discharge occurred, as with lake storage
of stormwater.
All water quality samples collected are representative of
the total streamflow volume (including baseflow) which is re-
ported in Table 17. In the majority of events, overland runoff
is approximately equal to total streamflow and samples can be
considered representative of runoff. For storm events producing
low runoff, water quality is influenced by baseflow.
Comparison of Stormwater Runoff and Wastewater Quality
A water quality comparison of stormwater runoff to untreated
and treated municipal wastewater is presented in Table 18.
Stormwater runoff quality is better than untreated wastewater
except for excessive solids concentrations. Treatment of waste-
water reduces the oxygen demand below levels of stormwater run-
off but sophisticated removal processes are required to reduce
nutrient levels in wastewater below those in stormwater runoff.
OBSERVED TEMPORAL AND SPATIAL VARIATIONS IN STORMWATER RUNOFF
QUALITY
Pollutoqraph Analysis
A pollutograph is defined as a plot of pollutant concentra-
tion versus time during a storm event. Temporal changes of
water quality during runoff events are important to the under-
standing of the impact of these non-point sources on stream
quality. The time-concentration relationship is also critical
in consideration of stormwater treatment alternatives. Polluto-
graphs observed during the study exhibited the five generalized
patterns shown in Figure 15. These concentration patterns were
common to all watersheds, although levels of a particular para-
meter were site dependent.
Specific conductance of groundwater which feeds streamflow
is high due to dissolved minerals and as a result stormwater in-
flow decreases stream conductance similar to the second polluto-
graph shown in Figure 15. This dilution pattern applies to other
streamwater constituents, including some found in wastewater ef-
fluents, which are concentrated in dry-weather flow. The con-
centrations of SOC, soluble COD and total COD often increased as
runoff progressed, with highest concentrations observed at the
end of the runoff. Streamflow contributions from interstitial
and bank storage flow is greatest late in runoff and could ac-
count for the pattern if enriched by contact with soils serving
as an organic carbon source. DO concentrations in stormwater
increased proportionally to flow and assumed a hydrograph-shaped
58
-------
TABLE 17. RUNOFF WATER QUALITY SUMMARY-
MASS FLOW AND WEIGHTED AVERAGE
Storm
I
1
2
3
4
5
6
7
8
9
10
11
Date
01/18/74
03/20/74
03/26/74
04/11/74
04/22/74
10/23/74
12/05/74
03/04/75
03/13/75
04/08/75
05/08/75
Site
The Woodlands P-30
Hunting Bayou
Hunting Bayou
Hunting Dayou
The Woodlands P-30
The Woodlands P-30
The Woodlands P-30
The Woodlands P-10
The Woodlands P-30
Lake A
Lake B
The Woodlands P-30
The Woodlands P-10
Lake A
Lake B
The Woodlands P-30
The Woodlands P-10
Lake A
Lake B
Hunting Bayou
Westbury
Streamf low
acre— feet
2334
2.51
42.1
12.4
5.80
937.
1267
833.
3.32
Stored
.135
119.
79.1
2.38
1.77
2829
1614
93.4
93.2
28.9
7.16
x cone
Ibs mg/1
7.73 .001
2.96 .437
45.1 .395
16.9 .503
0.70 .044
43.0 .017
33.7 .010
14.1 .006
0.23 .025
NA .008
.012 .033
3.15 .010
1.15 .005
.043 .006
.102 .021
85.2 .011
15.0 .003
3.75 .015
1.25 .005
40.4 .516
13.8 .710
TP
x cone
Ibs mg/1
No Data
No Data
46.4 .407
30.3 .899
7.53 .478
256. .101
328. .096
171. .076
1.14 .127
NA .097
.077 .210
52.3 .162
19.7 .092
.893 .137
2.54 .530
675. .088
264. .060
26.2 .103
27.9 .110
84.2 1.08
14.2 .730
NEL
XTTOnc
Ibs mg/1
215. .034
16.5 2.44
28?. 2.51
24.6 .732
5.27 .334
236. .093
355. .103
199. .088
.880 .200
NA .062
.238 .650
22.2 .069
19.2 .089
.760 .117
1.06 .221
1144 .149
339. .077
39.6 .156
28.1 .110
94.4 1.20
17.3 .890
NA - Not applicable - a mass loading was not calculable.
Stored - Ho discharge of stonrwater from the lake system.
1 ac =
1 ft = .305 m
1 Ib = .4536 kg
(continued,
59
-------
TABLE 17. (continued)
N02
x cone
Ibs wg/J
5.30 .001
.353 .052
7.42 .065
2.18 .065
.170 .011
12.5 .005
IS. 5 .005
.280 .000
.055 .006
NA .008
.016 .044
1.08 .003
.750 .003
.065 .010
.113 .024
73.1 .009
17.1 .004
8.24 .032
2.18 .009
4.52 .058
.754 .039
NO,
x cone
Ibs nvj/1
191. .029
2.96 .438
58.1 .509
11.4 .338
3.71 .235
94.2 .037
79.9 .023
27.6 .012
.960 .107
NA .189
.131 .357
35.9 .111
22.1 .103
1.32 .203
1.74 .362
1181 .154
284. .065
70.4 .280
37.6 .150
40.1 .511
7.20 .371
TKN
Ibs x cone
xlO mg/1
No Data
No Data
40.1 3.52
5.25 1.56
2.11 1.34
No Data
311. .905
187. .826
1.50 1.66
NA .829
.066 1.79
35.1 1.09
34.5 1.61
.401 .620
1.99 4.14
1065 1.39
600. 1.37
33.3 1.31
47.1 1.86
25.5 3.25
4.25 2.19
TSS
Ibs x cone
xlO3 mg/1
1492 236.
.480 70.9
22.5 197.
4.10 122.
14.8 939.
775. 305.
296. 86.2
60.5 26.8
.426 47.0
NA 166.
.104 283.
28.8 90.0
14.4 67.0
.991 152.
13.8 2877
1312 171.
168. 38.5
61.9 245.
322. 1273
16.2 207.
.474 24.4
soc
Ibs x cone
xlO2 mg/1
No Data
2.59 38.3
18.5 16.3
18.6 55.3
2.74 17.4
455. 17.9
No Data
No Data
.802 9.00
NA 14.0
.027 7.35
69.3 21.6
48.2 22.5
.900 13.8
.705 14.7
1563 20.4
962. 22.0
34.4 13.6
40.8 16.2
13.7 17.5
3.13 16.1
Total_COD
Ibs, x cone
xlO"1 mg/1
No Data
5.18 76.6
114. 99.8
34.3 102.
14.7 93.1
No Data
1962 57.1
1400 62.0
3.02 33.4
NA 47.0
.181 49.2
192. 59.3
136. 63.4
2.47 38.0
5.90 123.
4037 52.6
2572 58.8
106. 41.8
161. 63.7
140. 179.
10.4 53.8
(co
Soluble COD
Ibs, x cone
XlO* mg/1
No Data
No Data
No Data
10.8 32.0
8.53 54.2
859. 33.8
1405 40.9
988. 43.7
2.88 32.0
NA 32.5
.120 32.6
143. 44.2
111. 51.9
1.87 28.8
1.37 28.5
2981 38.8
1888 43.1
66.8 26.4
82.0 32.0
27.5 35.1
6.09 31.4
ntinued)
1 ac = .405 ha
1 ft = .305 m
1 Ib = .4536 kg
60
-------
TABLE 17. (continued)
Storm
I
12
13
14
15
16
17
Date
06/30/75
09/05/75
10/25/75
03/07/76
03/08/76
04/04/76
Site
Bunting Bayou
Hestbury
The Woodlands P-30
The Woodlands P-10
Lake A
Lake B
The Woodlands P-30
The Woodlands P-10
Lake A
Lake B
The Woodlands P-30
The Woodlands P-10
Lake A
Lake B
The Woodlands P-30
The Woodlands P-10
Lake A
Lake B
The Woodlands P-30
The Woodlands P-10
Lake A
Lake B
Strear.f low
acre— feet
116.
11.2
9.82
.918
Stored
1.51
117.
57.1
18.9
No Data
14.9
11.2
Stored
.384
99.3
58.3
6.78
2.71
45.9
2.83
14.5
5.10
ortho P
x cone
Ibs mg/1
212. .677
17.1 .560
1.67 .063
.051 .021
NA .011
.148 .035
12.0 .038
.342 .002
.068 .001
NA .027
1.18 .029
.039 .003
NA .005
.124 .052
7.92 .029
.217 .001
.131 .007
.088 .012
5.38 .043
.027 .003
.148 .004
.077 .006
TP
,. x cone
Ibs mg/1
403. 1.28
34.6 1.14
4.75 .180
.082 .033
NA .050
.608 .148
46.7 .147
7.32 .047
1.35 .026
NA .081
No Data
No Data
No Data
No Data
No Data
No Data
No Data
No Data
18.7 .149
.468 .061
2.4S .063
1.07 .078
NH3
x cone
Ibs mg/1
660. 2.10
4.40 .145
5.36 .201
.075 .030
NA .069
.567 .138
26.1 .082
8.34 .054
2.27 .044
NA . 108
2.58 .064
1.67 .055
NA .063
.089 .037
46.8 .174
3.61 .023
.723 .039
.247 .034
48.9 .390
1.21 .158
4.79 .121
1.66 .120
NA - Not applicable - a n>aas loading was not calculable.
Stored - No Discharge of stortnwater from the lake system.
1 ac = .405 ha
1 ft = .305 m
1 Ib = .4536 kg
(continued)
61
-------
TABLE 17. (continued)
NO
x cone
Ibs mg/1
14.0 .044
.803 .026
.480 .018
.007 .002
NA .009
.047 .012
1.56 .005
.690 .005
.143 .003
NA . 010
.593 .015
.321 .010
NA .011
.035 .014
3.70 .014
1.66 .011
.225 .012
.328 .045
2.13 .017
.052 .007
.205 .005
.163 .012
N03
x cone
Ibs mg/1
117. .373
11.8 .388
8.13 .305
.075 .030
NA . 009
.955 .233
47.0 .147
4.67 .030
1.19 .023
NA .113
10.2 .252
1.43 .047
NA .076
.393 .164
48.0 .178
3.69 .023
1.64 .089
1.31 .141
52.9 .420
1.04 .135
4.53 .112
4.92 .355
TKN
Ibs x cone
xlO mg/1
124. 3.94
4.51 1.48
1.16 .435
.024 .097
NA . 172
.158 .410
9.69 .305
4.04 .261
.942 .184
NA .280
No Data
No Data
No Data
No Data
No Data
No Data
No Data
No Data
25.5 2.04
.516 .672
2.90 .735
1.29 .929
TSS
Ibs x cone
xlO3 mg/1
51.0 182.
2.13 69.8
2.67 100.
.016 6.54
NA 24.0
6.70 1633
53.8 169.
1.18 7.60
1.85 36.2
NA 421.
9.55 237.
6.47 212.
NA 184.
1.77 738.
78.1 290.
20.6 130.
3.27 177.
3.07 419.
No Data
No Data
No Data
No Data
SOC
Ibs. x cone
x!0z mg/1
71.6 22.8
5.09 16.7
No Data
No Data
No Data
No Data
66.0 20.7
37.8 24.4
2.38 4.66
NA 20.6
5.55 13.7
6.72 22.0
NA 6.00
.322 13.4
42.4 16.0
41.4 26.0
I. 11 6.05
.962 13.1
18.5 14.8
1.27 16.5
3.98 10.1
1.94 13.3
Total_COD
Ibs x cone
xlO2 mg/1
252. 80.4
11.8 38.9
No Data
No Data
No Data
No Data
135. 42.4
70.0 45.1
10.9 21.2
NA 43.9
No Data
No Data
No Data
No Data
No Data
No Data
No Data
No Data
61.7 49.3
3.08 40.0
7.88 20.0
6.00 43.3
Soluble COD
Ibs x cone
xlO2 mg/1
61.2 19.5
4.36 14.3
9.40 35.2
.852 34.5
NA 19.5
.878 21.4
77.0 24.2
58.9 38.0
7.21 14.1
NA 24.2
12.2 30.2
14.2 46.6
NA 15.4
.631 26.3
88.3 32.8
86.6 54.7
2.73 14.8
1.78 24.3
43.3 34.6
2.97 38.7
9.20 23.3
5.15 37.2
1 ac = .405 ha
1 ft = .305 m
1 Ib = .4536 kg
(continued)
62
-------
TABLE 17. (continued)
Ston
1
1
2
3
4
5
6
7
8
9
10
11
12
1?
14
IS
16
17
m
Date
01/18/74
03/20/74
03/26/74
04/11/74
04/22/74
10/28/74
12/05/74
03/04/75
09/13/75
04/08/75
05/08/75
06/30/75
09/05/75
10/25/75
03/07/76
03/08/76
04/04/76
Site
The Woodlands P-30
Hunting Bayou
Hunting Bayou
Hunting Bayou
The Woodlands P-30
The Woodlands P-30
The Woodlands P-30
P-10
The Woodlands P-30
Lake A
Lake B
The Woodlands P-30
P-10
Lake A
Lake 3
The Woodlands P-30
P-10
Lake A
Lake B
Hunting Bayou
Westbury Square
Hunting Bayou
Westbury Square
The Woodlands P-30
P-10
Lake A
Lake B
The Woodlands P-30
P-10
Lake A
Lake B
The Woodlands P-30
P-10
Lake A
Lake B
The Woodlands P-30
P-10
Lake A
Lake B
The Woodlands P-30
P-10
Lake A
Lake B
All measurements in mg/1 except:
Turbidity
150
-
135
100
305
250
65
20
35
121
184
50
30
105
1200
130
28
165
330
110
18
140
14
100
7
19
830
120
12
24
325
74
12
-
137
138
5
-
120
70
7
53
110
Turbidity in
pH
6.5
-
-
6.5
6.5
5.7
6.0
5.2
7.5
7.5
8.1
6.6
6.1
7.5
7.8
6.8
5.9
7.54
6.9
7.6
7.3
7.2
7.3
7.05
6.6
8.01
7.6
7.4
5.8
8.1
7.4
.
-
-
_
-
-
-
.
7.3
6.9
7.9
7.2
.JTU;
Temp.
14.7
20
15
22
22
-
-
-
14
16
-
13.5
13.3
17
11.5
-
19
20
20
24
-
27
-
-
-
-
_
18.5
18
22
19
-
-
-
-
-
-
-
-
-
-
-
-
pH In pll
Spec.
Cond.
_
-
-
545
200
110
58
56
375
115
415
280
290
160
185
103
110
140
80
460
175
360
125
208
280
336
135
275
247
536
167
281
410
-
155
-
3G5
410
115
313
284
363
125
uni ts ;
Total
D O Solids BOD
9
2.5
4.6
2.25
3.5
-
-
-
8.3
7.8
7.6
6.4
7.2
7.7
7.4
7.8
6
7.65
7.1
3.2
3.34
3.3
4.6
-
-
-
_
7.5
6.6
8.2
7.6
-
-
-
-
-
-
-
-
-
-
-
-
-
17
-
22
-
450 6.1
385
110
-
_ _
-
- -
-
_
_ -
-
-
_
.
-
- -
-
- -
-
-
_
_ _
-
-
— —
-
_
-
- -
_
-
-
-
_ _
-
_
-
-
Temperature in
Cnntiqrarto and Specific Conductance in micromhos/cm.
63
-------
W
0
15
W
10
P
W
EH
3
2
EH
Q
§
O
rfj
3
w
to
H
^
3
EH
|
x
fa
fa
0
y[
i
H
EH
jg
§
O
EH^
,_!
O &>
pf i S
EH H
H
l|
o£
W i4
EH »4
S %~'
•
00
I—I
w
,4
CO
EH
rH
<0
U
•H
E
CD
.C
CJ
rH
flj
•P
O
rH
fd
-P
O
EH
x:
us
TJ
rH
0)
•1—
«
H
(tf
O
^~*
tJ
CD
-H-g
(0 "
ii (i)
o a
E-i to
3
CO
r
'O
c
(0
E
Q)
P
C
CU
tr
>
X
O
0)
M
0
A
a
CQ
o
ft
c
/11
cu
o
JJ
•H
CO
'O
rH
0
CO
H
i
2
D
«
«
W
EH
^
•>"
S
o
EH
CO
CO
vo
I
CO
•^*
o\
o
o
1
CO
o
o
rH
VO
rH
1
O
H
0
r**
i
m
vo
*—
o
rH
1
ft
*— •
4J
W
J^
O
fa
rH
in
1
o
^*
o
CO
o
1
CO
rH
0
rH
•"S-
H
VO
O
O
rH
m
1
CTV
0
rH
1
•H
>
P 0
fO
t 1 1
CD ft
Tj *-^
c
D -P
C
JJ 0)
to E
cu &
5-1 O
O
fa
CO
CN
i-H
1
a\
T
CO
in
o
l
rH
rH
0
•*
rH
1
0\
rH
00
r-
oo
CM
H
1
i-l
'tf
0
^,
O
CO
1
vo
m
H
r-
o
i
rH
3
O
>^
rtJ
CQ
tn
•iH
4J
p;
3
a
^*
in
o o in
I in 'fl1 rH
CO
ON
CO
«d<
rH
rH
O O i-l
1 f— 1 rH
CO
t^
O
ON
rH
CM
O O O
| TJ- CO H
CO
rH
O
o o m
1 O CM
CM
CM
4J CM
C -P
CU C
e cy
cu -P E
s-t 1 Cl) 5-J T3
M -P nj l)
d w
CO H G CU t3
S COl D CO <
CQ
-P
C
CU
cu
E
M
O
CO
m
o
cu
3
C
^
O
M-l
CQ
C
O
•rt
• i
•P
ft)
CU
u
c
0
0
c
(0
0)
cu
4J
•g.
•iH
cu
o
i-H
M-l •
m cu
O M
O
CU -P
G d
CH E
i-H
^1
O
c
cu
•iH
••H
tl 1
Mi
rj
c
CO
cu
p
-p
m
rH
c
o
• H
•P
a
nj
cu
cu
I
CO
£
(fl
rH
o
o
CU
EH
CM
64
-------
Z
g
Z
UJ
O
Z
O
o
LU
_J
HI
(T
Id
O
CC
<
X
o
tn
Q
FIRST FLUSH
DILUTION
STABLE
INCREASING
HYDROGRAPH
T!ME INTO STORM EVENT
Figure 15. Generalized pollutographs observed
for stormwater parameters.
65
-------
pollutograph (Figure 15). Increased reaeration at greater
streamflows accounts for this phenomena. Several parameters
observed at site P-10 remained at a constant level throughout
the hydrograph, including pH, NH3, NO.,, NO2 and soluble COD.
This pattern was not commonly observed at the other watersheds
where land use is diversified.
First Flush—
The "first flush" pollutograph pattern is characterized by
an abrupt rise in concentration early in the runoff event. At
Hunting Bayou, the "first flush" was observed for the greatest
number of runoff constituents including TSS, turbidity, O-PO4,
TKN, TP, i
-------
ISCHARGE
0
CO
o
UJ
Z
K
Z
CO
Z
60.
45.
30.
15.
0.
75-
60.
co45-
u.
°30.
15.
0.
75-
60.
45.
30.
15-
0.
5
4.
3.
2.
I .
0
P-IO
STORM 9
5,
4
3.
2.
I .
0
WESTBURY
STORM 12
HUNTING /
BAYOU f
STORM 11
345678
HOURS INTO STORM
10
Figure 16. Comparison of TSS pollutographs at
P-10, Westbury and Hunting Bayou during similar
storm flow (hydrograph is unbroken line).
1 cfs = .028 m3/sec
67
-------
|H
pTj ^^
EH r-l
pa i
'T1
co pa
t—J t™*
J H
fa co
EH EH
CO
pa
^f
E DATJ
STORM
*.,
h*"
H co
EH
§ K
< EH
S PS
O O
U Pm
CT^
i—)
ca
CQ
s
^
in
f- <>0
\ CN
00 CN
\
f PS
I pa
i EH ffl
1 |zi S m
Pa ID r>
> a \ oo
1 Pa co t~-
P i-l i-H
' K § M
EH Pa
CO EH
P
r> o
\ on
in .-H
\
i ^
i — i
j
*
[
*~.
-^
1
ci
0
•H *• — ^
JJ r]
X) CQ
-Ur-l
•P CM-l
U (U
Q) UM-J
4-1 CO
m o
ta u n
o
= CO -P
£ co m
CO EH O
P -H
^-1 y ng
h fd C
Q) -r-l
•P ft *— '
CO
(H
-rH
h
CN
CN
n
^
r-l
rH
CO
M-l
o
^
CQ
•P
fd
0)
cn
fd
•s
CO
•H
P
CO
j3
O
(U
fd
-P
a
.p
CQ
a
H
g
fd
(1)
^4
-P
CO
CN
CO
in
•
o
' Hours
CO
3
r-l
fe
O
4J
C
0)
^
pa
m
o
c
5
o
-P
c
•H
i)
in
0
00
CN
"co"
U
—
Q)
JJ
PS
CD
CJl
^
fd
rC
U
CO
•r-l
P
S
0 X
i— i fd
m
FrT
II i
4-1
O
a
PS
0
JJ
-H CQ
s_i
0) 3
g O
•H E
r i %_^
in
r^.
*
rH
LO
•
CN
in
CT>
iH
CQ
\
rd
d)
D^
rd
•s
CQ
•r-l
P
J^
rd
d)
ft
c
0)
5
•p
a)
rQ
(U
g
EH
-a
0)
CQ
PI
rd
i— i
pa
CJ
0)
CO
\
e
00
CN
O
•
II
CQ
U
-H.
i-l
68
-------
Effect of Land Use on Stormwater Runoff Quality
Pollutant Load-Runoff Relationships—
Total pollutant loads were plotted against total runoff of
each storm event and regression lines were fitted to correspond
to Hunting Bayou, Westbury, P-10 and P-30 watersheds. Fitted
lines and associated correlation coefficients are shown in
Figures 17-20 for the constituents TSS, total COD, TKN, TP, NOs,
NHs, soluble COD and SOC. Correlation coefficients (r) for a
majority of the parameters were greater than 0.8, however, three
cases showed poor correlation; P-30, N03 (r = 0.775), Westbury,
NH3 (r = 0.397), and Hunting Bayou, SOC (r = 0.161).
N03, NH3, TKN and TP relationships, shown in Figures 17-18,
indicated that the urban watersheds produce nutrient loads
greater than the forested watersheds. In all cases Hunting
Bayou nutrient loadings are highest, followed in order by West-
bury, P-30 and P-10.
TSS loads are highest at P-30 as a result of construction
activities in the watershed. Urban runoff TSS loading is
greater than forest runoff loading (Figure 19).
Runoff loads for nonspecific parameters (total COD, soluble
COD and SOC), shown in Figures 19-20 are higher in forested
watersheds than urban watersheds, with the exception of high
total COD loads from Hunting Bayou. The data suggest organic
material in runoff decreases with urbanization, however, in-
soluble pollutants, sediments and oils will increase.
A ranking of the four watersheds, on a Ib/ac/in of runoff
basis, illustrates the relative conditions for each site and
pollutant. Table 2o shows these rankings for mean regression
values at one inch of runoff for all parameters and sites. Con-
fidence intervals (95%) are included in Table 20 to indicate
significant differences in pollutant loads at 1 inch runoff.
Confidence limits for Westbury show a particularly large spread
due to the small number (2) of storms monitored for that water-
shed. Signifcant differences for the total COD, soluble COD and
SOC cases are indicated, with some overlap in the TSS case. The
pattern of nutrient response for the urban developing and for-
ested watersheds is distinctive, with the urban response pro-
ducing loads up to an order of magnitude larger. Hunting Bayou
ranks first as the producer of the largest pollutant loads.
The load-runoff relations developed from several storm
events can be extended in a useful way to estimate total annual
loads for selected pollutants. A measured or predicted annual
streamflow hydrograph is required along with average low-flow
concentration values. During storm events, the load-runoff re-
lation is used to predict the mass flow, while during intermit-
tent low flows, mass flow is estimated by the product of stream-
69
-------
O
o
g
H
*
X
a
w
i
8
00 &
CO N-
? i'
to
d
-------
o:
m
to
UJ
OV/S81
in
O
Z
LL.
o
(0
LU
X
o
2
I-
X
•
?"
a
g"
o.
o
a>
£
o"
a
M
O
IM
CO
a
-1-1
£
CO
d
O
-r-(
•p
4-1
m
o
d
3
i
to
o
CJ
o
o
o
tn ro S
V^ f-f fl
^i* to ^*
in o m
OJ
II II
1 3^-S
00
r-l
-------
OV/SST
m
en
LU
o
X
•
o
ro
g
i
CL
O
>»
0>
w
to
C
O
•rl
4J
(1)
Q
o
O
fi
tn ro e
X £ O
^t* in -vf
in o m
^t
M O II ll II
I 4J
T3 rQ O C
rO >rt iH n) -H
O C
J (tf rH r-t r-l
CTl
i-l
0)
M
3
tn
OV/S81
72
-------
r to
£
£C
3
m
f-
v>
LU
$
«
o
I
•
i
Q.
O
§
U
O
•H
^
CO
c
O
-r-l
4J
a
m
m
o
c
3
M
i
T3 '
(0 U
O O
J CO
(1)
M
3
cn
tn
-------
TABLE 20. POLLUTANT LOAD RANKING OF THE FOUR STUDY
AREA WATERSHEDS
RANK
2 3
ss
rcoo
SCOD
soc
N03
NH3
TKN
TP
Area
Upper CL*
1" Value
Lower CL
Area
Upper Cl
1" Value
Lower CL
Area
Upper CL
1" Value
Lower CL
Area
Upper CL
1" Value
Lower CL
Area
Upper CL
1" Value
Lower CL
Area
Upper CL
1" Value
Lower CL
Area
Upper CL
1" Value
Lower CL
Area
Upper CL
1" Value
Lower CL
P30
61.59
43.48
25.36
HB
25.62
18.88
12.14
P10
10.01
9.86
9.70
P10
5.045
5.00
4.95
WB
0.12
0.088
0.057
HB
0.58
0.48
0.38
HB
1.049
0.95
0.85
HB
0.40
0.28
0.16
HB
42.30
37.73
33.15
P10
14.13
13.50
12.87
P30
9.27
8.76
8.25
P30
4.78
4.54
4.30
HB
0.10
0.087
0.072
WB
1.28
0.069
-1.14
WB
1.50
0.40
-0.76
WB
0.90
0.24
-0.43
WB
88.02
13.66
-60.70
P30
12.72
12.09
11.46
HB
6.64
4.44
2.24
WB
4.48
3.75
2.66
P30
0.038
0.020
0.0031
P30
0.037
0.031
0 .025
P30
0.36
0.30
0 .24
P30
0.028
0.021
0 .014
P10
10.16
8.19
6.21
WB
33.13
9.48
-14.17
WB
31.56
4.06
-23.45
HB
2.64
0.70
-1.25
P10
0 .020
0.012
0 .0051
P10
0 .020
0 .018
0 .016
P10
0 .38
0 .28
0 .18
P10
0.017
0 .014
0 .011
* 95% Confidence
gression values
Level. All
at 1 inch
confidence
of runoff.
levels are
All loads
for mean re-
are in Ibs/acre.
P30
P10
HB
WB
Woodlands P30 Watershed
Woodlands P10 Watershed
Hunting Bayou Watershed
Westbury Watershed
Note: 1 in = 2.54 cm
1 Ib/ac = 1.12 kg/ha
74
-------
flow and concentration.
A comparison of annual loads for TSS, TP, N03 and total COD
is shown in Table 21 for the P-10 forested site and P-30 urban-
izing site at The Woodlands. The developed load-runoff relations
were used to calculate the storm generated mass flows. The
urbanizing watershed appears to be contributing greater loads of
TSS on an annual basis compared to the forested site.
The procedure allows direct comparison of storm generated
pollutant loads from non-point sources with the low-flow contri-
butions, which are primarily of point source origin. Consequent-
ly, non-point loads can be quantitatively determined as a func-
tion of land use patterns as more storm data becomes available
from other urbanizing watersheds.
The annual TSS load calculation can be used in conjunction
with the U.S. Geological Survey grab sample method to calculate
annual sediment loads. Relative accuracy of the two techniques
remains undetermined.
Effects of Land Development on Runoff Quality—
Stormwater quality monitored at site P-10 represents runoff
from a forested, undeveloped watershed and accordingly serves as
a baseline for assessing changes due to urbanization. The P-30
sampling site located 6.8 miles (11 km) below P-10 monitors run-
off from an additional 5,500 acres (2250 ha) which includes con-
struction activity of The Woodlands Development Corporation. A
comparison of these two sites during storm event #10 illustrates
the effects of construction activity on runoff quality.
Heavy rainfall over the Panther Branch watershed on April 8,
1975 produced large amounts of runoff sampled at P-10 and P-30.
Precipitation associated with the storm event began shortly after
midnight on April 8, 1975 and continued till noon the same day.
The storm featured 3 periods of intense rainfall at 4:30, 8:30
and 10:00 A.M. with interposing pauses or drizzle. The area
rain gages measured 2.00, 2.65, 3.42 and 3.97 inches (5.OS, 6.73,
8.69 and 10.08 cm) of rainfall, upper to lower watershed gages
respectively, with the Theissen adjusted rainfall calculated to
be 2.43 in (6.17 cm) on the P-10 watershed and 2.76 in (2.01 cm)
on the P-30 watershed. Average rainfall intensity was 0.76 in/hr
(1.93 cm/hr) and antecedent soil moisture conditions were dry
with no rain recorded 7 days prior to the storm and 2.5 in (6.35
cm) the preceding month. Watershed runoff began after midnight
April 8 and ended three days later. As shown below, the volume
of runoff observed at site P-30 was greater than P-10 but peak
discharge was essentially the same.
75
-------
CN
r-3
m
13
0)
C
_Q
g
o
u
w
0)
,c
to
n
HI
•P
n)
13
o
0*1
i
TD
C
n)
o
,— j
1
ft
U
rCi
to
0)
4J
id
*
o
rH
1
CU
to
• e
0 V4
a o
Cfl
«t
o
rH
Q
8*
rH
CN
0
rH
CO
O X
to
rQ
rH
CM
0
rH
D-l
H X
to
a
in
o
i—4
CrO
c2
L^ [Q
rQ
rH
O
rH
§*
to
rQ
CN
O
rH
CO
§
to
rQ
CM
0
rH
PH
H X
to
£
in
o
•-*
W X
k"1 to
£
•"*
f
c
rH
CO
CO
rH
CN
01
rH
CM
CO
CN
CO
rH
VO
in
CO
CO
VD
01
O
0
co
o
CO
00
't
o
t
r-
-
4J
O
<4*
m
CO
CN
rH
CN
01
in
rH
VO
&
in
rH
VO
in
CO
^
01
CO
CO
CO
rH
r-
[^
*>r
CO
ID
r-
CN
vo
>
o
a
^i *t
rH rH
rH rH
CO CN
01 O
r~ ^>
00 CN
CN >*
o o
CN m
rH
^i VO
00 CN
vo in
CN
VD 00
•* CO
in •*
CN ^
«rjl rH
^ CN
VD CN
rH CO
<* rH
rH CN
t~- rH
CO rH
0 C
r- Tj1
r^ CN
Tj" VO
in v£>
P^ rH
CN CN
^t 't
^t rH
a id
< s
CN O
rH 01
ID CO
CN r-
VD CO
o o
0 O
•* CO
rH rH
vD vD
O rt
rH O
oo in
U3 CN
in co
rH CO
CN CN
CO O
01 ^1*
in rH
CN 0
01 r-
in rH
O 0
Ql ^
C rH
3 3
O O
CN CN
CN O
CN
rH CN
CN O
O O
r- CO
r- o
O 0
r>.
\D CN
CN O
O O
O CO
in rH
rH O
CO vO
VD O
0 O
0 0
in in
VD O
0 0
O 0
CO -H
o o
o o
in
r^
—
•
< w
in
CN
CO
t-
"*
(^
CO
in
in
in
CO
r-
VD
o
•-
-------
Site
P-10
P-30
Streamflow Volume
acre-ft. (ha-m)
1614
2829
(199)
(349)
1
1
Runoff
in. (cm)
.2
.57
,(3
(3
.05)
.99)
Peak
cfs
1170
1100
Flow
(m3/s )
(33.
(31.
1)
1)
Runoff
Coeff .
50%
57%
The greater runoff volume for P-30, almost twice that of P-10,
was a result of three factors: (1) larger drainage area, (2)
heavier rainfall in the lower basin, and (3) impervious areas in
The Woodlands development.
Figures 21 and 22 compare P-10 and P-30 pollutographs for
TSS,total COD, TP and TKN. Hydrographs are also presented
in the figures for flow rate and time references. Pollutograph
analysis should consider the following:
1. Those areas of The Woodlands developed or under
construction encompassed only 10% of the total
watershed. The majority of stormwater runoff
originated in undeveloped forest lands.
2. Developed or construction areas were located
adjacent to P-30, as shown in Figure 23. As
a result, runoff originating in these areas
was observed early in the storm event.
The pollutographs indicate high TSS loads at
P-30 (Figure 21), a result of sediments washed from easily eroded
construction sites. Although the developing area comprised 10%
of the watershed, it contributed as much as 80% of the TSS
load at P-30. Total COD (Figure 21) exhibited a "first flush"
at P-30, probably due to associated high TSS. The
major portion (70%) of the total COD was soluble. Significant
increases in TKN and TP at P-30 resulted from wash off of
ammoniated phosphate and urea based fertilizers applied to the
golf course in the developing area (Figure 22).
Development within the Panther Branch watershed has resulted
in stormwater runoff quality changes. TSS and
nutrients have increased although no significant change has been
observed for oxygen demand. Results for storm event #10 are
summarized in Table 22.
STORMWATER QUALITY MODELING
Several techniques are available for the prediction of water
quality responses in a watershed. The SWMM model has been
adapted for natural drainage conditions at The Woodlands in an
effort to simulate stormwater quality response. The model oper-
ates from a relation between runoff rate and pollutant load, but
the prediction of hydrographs has been more successful than pol-
7_7
-------
656 TONS
UJ
tr
x
o
en
Q g
S *
2 en
|S
I
en
z
84 TONS
I 1
20
120-
90-
60.
30.
0
Figure 21,
128 TONS
,-^fr^Vv •*
» ^ • •
2829 AC-FT
1614 AC-FT
15 30 45 60 75
HOURS
P-30
90
15 30 45 6O 75 90
HOURS
P-IO
Comparison of P-10 and P-30 temporal distribution
of streamflow, TSS and total COD for the storm
event of April 8,. 1975.1 1 cfs = .028
1 ton = 907.2 kg; 1 ac-ft = .124 ha-m.
78
-------
0.4.
_, 0.3.
S 0.2.
O.I .
0.0
675 LBS
264 LBS
I I I
* O
t- 2
4.0,
3.0,
2.0.
1.0.
QO
tu
o
(t
I
to
5
120
9 90
co v
D x
S co 60 J
Z f-
< o
I
30.
1065 LBS
600 LBS
2829 AC-FT
1614 AC-FT
15 30 45 60 75 90
HOURS
P-30
15 30 45 60 75 90
HOURS
P-IO
Figure 22. Comparison of P-10 and P-30 temporal distribution
of streamflow, TP and TKN for the storm event of
April 8, 1975. 1 lb = .454 kg; 1 cfs = .028 m3/sec;
1 ac-ft = .124 ha-m.
79
-------
CONSTRUCTION AREAS
PROPERTY BOUNDARIES
WOODLANDS, TEXAS
Figure 23.
The Woodlands construction activity in
relation to the P-10 and p-30 sampling
sites.
1 mi = 1.6 km
1 ft = .305 m
80
-------
TABLE 22. COMPARISON OF STORMWATER QUALITY AT P-10,
P-30 AND DEVELOPING AREAS DURING STORM #10
)rainage Area (acres)
(ha)
Streamflow volume (ac-ft)
(ha-m)
Forest
with
Development
P-30
21,606
8,750
2,829
349
Forest
P-10
16,050
6,500
1,614
199
The Woodlands
Development
(P-30) - (P-10)
5,556
2,250
1,215
150
Average Concentration of Water Quality Parameters (mg/1)
Ortho-phosphate
Total phosphorous
Ammonia
Nitrite
Nitrate
Total Kjeldahl Nitrogen
Total Suspended Solids
Soluble Organic Carbon
Total COD
Soluble COD
0.011
0.088
0.15
0.009
0.154
1.39
171
20.4
52.6
38.8
0.003
0.06
0.08
0.004
0.065
1.37
38.5
22
58.7
43
0.021
0.125
0.244
0.017
0.272
1.41
347
18.2
44.5
33.2
lutant response. The water quality procedure in the model is not
designed to simulate the response from natural drainage, and has
been updated for The Woodlands. New relationships between cumu-
lative load (Ibs) and cumulative runoff volume (ft^) have been
incorporated into SWMM for various parameters at The Woodlands.
In this way, concentrations can be predicted as a function of
runoff (22).
Pollutant Load Modeling for Multiple Events
The load-runoff relationships presented previously(Figures 17-
20) provide the foundation for an uncomplicated, yet satisfac-
tory, model for runoff pollutant load simulation of multiple or
individual storm events. Given time increment values for runoff,
the model consults time-varying load-runoff relationships to
calculate mass flows during storm events.
Variation in the average pollutant concentration over time
is approximated by variation of the load-runoff line slopes
(Figures 17-20). These slopes represent the ratio of mass/time
and volume/time, or mass/volume which is an average pollutant
quality concentration for each watershed. Initially three para-
meters are defined for each load-runoff relationship: the
average slope, the initial slope, and a factor which sets the
range within which the slope can vary. The average slope can be
81
-------
roughly determined from the cumulative relationship produced
from field data. The initial slope value depends primarily on
initial conditions, and the range variable is determined by the
spread in observed pollutant concentrations.
During dry periods the slopes are incrementally increased
up to but not above the pre-defined maximum. This corresponds
to the buildup of available pollutants on a watershed between
storm events. An increment chosen to increase the slopes is
required as input and is obtained primarily by trial and error.
During a storm event the value of the slope decays exponentially
by the same means employed in both the Storm Water Management
Model (23) and the "Storm" model (24).
Lbs pollutant , Lbs
washed off in *^X. remaining on
any time interval the ground
or:
(1)
dt
which when integrated takes the form:
P0-P = P0 (l-e-kt) (2)
Where P -P = Ibs washed away in time, t, and k is assumed to
vary in direct proportion to the rate of runoff, r:
k = br (3)
b can be evaluated given the assumption that 0.5 in (1.3 cm) of
runoff uniformly delivered in 1 hour washes away 90% of the pol-
lutants (22). As a result the equation can be written:
P0-P = P0 (l-e~4-6 rt) (4)
The equation used to decay the load-runoff line slopes is:
PDS = l-e~4'6 rt (PDS)Q (5)
Where PDS is the load-runoff curve slope at some point in time
during the storm event and (PDS) is the initial value.
A six month period of streamflow at site P-30 was chosen for
sequential simulation. This period dating from October 28, 1974
to April 8, 1975 includes storm events 5, 7 and 10 monitored
during the study. Storm events 8 and 9 were considered too small
82
-------
for use in the simulation. Predicted solids loads and the ob-
served strearaflow hydrographs are presented in Figure 24. Slope
parameters used were derived from the load-runoff relationship,
with the upper limit values found by trial and error.
Simulation results can be evaluated by comparing observed
and simulated mass flows for individual storm events. As shown
in Figure 25, the simulated curves compare satisfactorily to the
observed mass flow curves. Table 23 gives comparisons of simu-
lated to observed values for total pounds TSS, and peak magni-
tudes for each of the three storms.
Unit Loadograph for Single Event Simulation
The form of the stormwater mass flow curves, obtained from
the product of instantaneous concentration and discharge, re-
semble the general shape of the streamflow hydrograph and provide
a more useful measure of runoff loadings than the concentration
curves. For a given watershed, it is possible to generate a unit
hydrograph based on the incomplete gamma distribution. Using the
theory of linear reservoirs, the resulting equation for the unit
hydrograph becomes (25)
11"1 exp(-t/k) (6)
where S = total storage (one inch of runoff) ; k = constant; n =
outflow from n^h reservoir. The watershed is considered as n
serially arranged linear reservoirs, and it is possible to fit an
observed hydrograph by varying k and n.
The theory of linear reservoi'rs can be extended for mass
flow curves in order to develop a corresponding unit pollutograph
or unit loadograph for application to urban storm runoff. In
this way, mass flow curves can be generated in a similar fashion
to the hydrograph by varying appropriate constants. The water-
shed is considered as n serially arranged tanks with first order
decay, and the mass balance becomes
dM,
- g0 - gx - klM (7)
where Mi = total mass; go = mass inflow; g^ = mass outlfow; k^ =
linear decay coefficient. By solving the equation for n reser-
voirs in series, assuming outflow from one is inflow to the next,
and assuming M = kg (analogous to S = kQ) , the final resulting
equation for the unit loadograph becomes
83
-------
o
HI
O)
3
o
_j
o
CO
o
LJ
O
z
UJ
o
CO
O
to
I
Q_
01 WH01S
L wyois
9 WMOIS
C4
O
S/9
sanos
in
cvi
o:
a.
O)
00
CJ
o
o
X
CL
<
(T
O
O
o:
o
o
to
I
Q.
cr
m
LU
U.
O
LU
Q
>
O
o "b
CM
O
o
o
SJO
39dVHOSia
o
00
i
cu
0)
^
4J
in
CN
O
-U
00
04
d) O
-p 1-1
u
-.-I m
n3 O rj
*-t *O U3
a o \
C
rt
>-) M 11
^ Cn "
O m
CM
nvioi
84
-------
o
~t
o
f>
8
o
' CM
(0
M
*
tOI X S/9
SOHOS 030N3dSOS
a:
o
I
O
CO
CO
•H
CO
•
H3 CO
C -P
nJ C
(1)
•C >
<1) 0)
§B
co o u
,Q 4J OJ
O tO W
•a on
CO
oo
ON
o
M II
tn co
O (U co
^ > m
T3 n o
S, 3
ffi U M
m
-------
CO
M
O
W
vp
cN
rH
0)
'\
tn
CO
r<
«T
0)
&)
•
^
rj
T3
0
£i
Q)
CO
,0.
o
J
J_l
,£
T3
CU
4-)
fO
H
rj
g
rH
CO
in
0
M
M
W
ro
O
rH
X
CO
r£j
£
J3
0
Q)
M
0)
CO
Q
O
^
0)
4J
m
, — i
p
£
cc
e a
O E
4J £
•«t
in
o
0
m
CN
n
o
0
\Q
CN
rH
in
rH
O
0
CN
in
CO
H
O
0
r-
^
O
CN
CN
O
O
CTi
CO
t^.
CN
O
O
en
CO
in
•
CO
in
r~»
r-
o
00
00
\Q
r-
&
CN
Cn
O
O
CN
CN
\D
CN
cn
o
m
CO
CN
in
rH
VO
CN
o
^*
CO
r~«
o
\G
o
en
CO
0
in
c^
CO
[^
o
CTi
CO
o
m
r^
CO
in
CO
CN
rH
CO
H
CO
in
CO
o
H
oo
CN
a
-9
rd
e
•H
CO
m
o
C
•H
a
c
•H
0)
U
Q)
CO
\
&k
*3> 00
m CN
^ o
II II
,Q en
rH MH
U
rH
rH
t*
0)
4J
O
a
CD
CJ
c
•H
CO
0)
e
•H
4-1
C
•rH
'O
0)
SH
i-t
measi
CO
•rH
S-l
r^
86
-------
g^sfe-l*)"'1 «*<-»*> (8)
The similarities of equation 8 and equation 6 are obvious, where
gn is mass flow (kg/sec), a equals (1 + kk]_)/k, and M is total
mass.
Hydrograph and mass flow simulations for Storm #10 on the
P-10 site are shown in Figure 26. The timing of the hydrograph
peak and the total volume compare well, but the recession rate
is predicted lower than the observed. The simulation of TSS and
TP mass flow curves (g/sec) yielded similar results, with good
peak and total mass definition, but a predicted recession rate
lower than observed. This storm yielded 1.2 in (3.05 cm) of
runoff, and the unit response (in inches) can be obtained by
dividing all ordinates by 1.2.
The application of this approach is in a preliminary stage
due to lack of significant storm runoff data (at least 1 in or
2.54 cm) on Hunting Bayou or The Woodlands watersheds. As more
storm event data are collected from other watersheds in the area,
it will be possible to investigate relationships between the
gamma distribution shape parameters (n,k) and land use or physio-
graphic factors in the watershed. In general, the time of peak
of the unit loadograph is related to n and k by the equation
t = (n-l)k (9)
p 1+kk-L
Urbanizing watersheds should have lower values of n and k than
undeveloped watersheds of the same size. As n is increased, k
must be correspondingly decreased in order to yield the same t—
value for a given watershed.
The unit loadograph can be used in the same manner as the
unit hydrograph. Once the unit response has been determined for
a watershed, storms of varying intensity can be analyzed by lag-
ging and superposition of the unit graphs. A unit pollutograph
(concentration vs time) is found by dividing the ordinates of the
unit loadograph by corresponding hydrograph flows. Because of
the linear load-runoff relationships which have been developed,
the linear assumption of unit response is further justified.
The unit loadograph approach suffers the same limitations as
the unit hydrograph method with regard to assumptions of uniform
rainfall and initial conditions, but it does offer a relatively
simple and useful technique for analyzing stormwater pollution
response as a function of land use, watershed characteristics,
and hydrologic conditions.
87
-------
Ill
CO
tr
o
en CM
5 o
en x
UJ
en
CO
u_
o
12
9.
6.
3.
HYDROGRAPH
TOTAL SUSPENDED SOLIDS
N = 8
K = .085
28 36 44 52
HOURS INTO STORM
60 68
Figure 26 .
OBSERVED
FITTED
Fitted curves for storm runoff and pollutant
mass flows observed at P-10 on 4/8/75.
1 cfs = .028 m /sec
88
-------
STORMWATER TREATMENT AT THE WOODLANDS LAKES
Water Quality Needs
Irrigation—
Collected stormwaters are to be used for golf course irri-
gation at The Woodlands to supplement natural precipitation. The
critical water quality parameter for irrigation is salinity. Ex-
cessive salinity affects plants by increasing osmotic pressure in
the soil which limits uptake of water by plants. However, this
is not the case at The Woodlands and, therefore, salinity will
not be a problem. Electrical conductivity measurements in storm-
water runoff is less than 3000 micromhos at The Woodlands and is
"excellent to good for most plants" (26). The presence of nutri-
ents in stormwaters slated for irrigation is not high and in this
case considered an asset rather than a pollutant. High TSS
concentration or large particulates could cause mechanical
problems such as pump damage or clogging of sprinkler heads, but
careful placement of the intake structure will avoid these diffi-
culties. TSS concentration in Lake Harrison during
low-flow conditions is about 100 mg/1 and average particulate
size is estimated at 5 microns or less. Average storm event TSS
concentrations range between 25 and 250 mg/1, levels acceptable
for pumping requirements. The velocity in the distribution sys-
tem will keep the solids in suspension, particulate size cri-
teria will be set by the orifice size of the irrigation system.
Aesthetics—
The water quality level for aesthetics is presently met
without any stormwater treatment. The lake is devoid of floating
debris or objectionable odors and promises to support a wide
variety of lifeforms. Superficially, it resembles early stages
of other local man-made lakes. High nutrient levels may promote
macrophyte growth and algal blooms, but macrophytes can be con-
trolled by a regular lake maintenance program and algal blooms
can be prevented by reducing the lake detention times and nutri-
ent levels (21).
Recreation—
In a discussion of recreational water uses, two divisions
must be considered: contact and noncontact. The water quality
requirement for contact recreation, which involves substantial
risk of ingestion, is more stringent than that of noncontact
(27). Swimming is the primary example of contact recreation and
is prohibited in The Woodlands' lakes.
A water quality criteria designed strictly for boating would
be similar to that for aesthetics, with the added requirement for
fecal coliform levels. It is recommended that fecal coliform
levels of 2000/100 ml average and 4000/100 ml maximum be observed
for "unofficial recreation" waters. Levels of 1000/100 ml
average and 2000/100 ml maximum were suggested for official non-
89
-------
contact waters (28). Fishing water criteria invoke an additional
requirement that harvested species be fit for human consumption.
Edible fish species should be free of toxic chemicals and patho-
genic bacteria or viruses. The data to determine the fulfillment
of this requirement is not presently available. The consumption
of fish from such waters has been practiced without harmful re-
sults. Coliforms in the digestive system of fish caught at Wood-
lands are in higher concentrations than from other lakes but pre-
sumably do not reach the edible portions (28).
Water Supply Uses—
Lake Harrison ranks as a poor raw drinking water source be-
cause it would require a high level of treatment before use (26).
With groundwater, a less expensive and more reliable source is
easily obtained. Its use for this function is to be restricted
to emergencies.
THE LAKE SYSTEM
The man-made lakes at The Woodlands will serve as recrea-
tional centers, wildlife preserves and storage for stormwater
runoff. The lakes will contribute to the maintenance of a
perched water table necessary for plant life. The lake water
will also be used for irrigation of adjacent golf courses. In-
puts to the lake system are limited to stormwater runoff and
treated wastewater from The Woodlands Wastewater Reclamation
Plant with ultimate capacity of 6 mgd (.264 m3/sec). For a
year of average rainfall, treated wastewater will comprise 75% of
the flow through the lake system. Design effluent quality for
the plant is indicated in Table 20 and suggests that the lakes
will contain clear waters with somewhat elevated nutrient con-
centrations as compared to existing surface waters. Lake de-
tention time during dry weather will be approximately six days.
Consequently, treated wastewater will be the dominant influence
on lake water quality at The Woodlands.
The lakes will serve as stormwater storage reservoirs and, in
so doing, will remove significant amounts of pollutants, primari-
ly due to sedimentation.
Lake Harrison Sedimentation
Eight storm events were monitored at the lake system, ranging
from 0.26 in. (0.66 cm) to 3.97 in. (10 cm) of rainfall. In the
following paragraphs, the largest storm indicates the usefulness
of reservoirs for preventing release of construction site sedi-
ment washoff.
Storm Event #10—
Rainfall associated with the storm event began shortly
after midnight on April 8, 1975 and continued until noon the
same day. An early morning cloudburst was followed after a four
90
-------
hour pause by less intense rainfall totaling 3.97 in. (10 cm).
The hyetograph is shown in Figure 27. Runoff passing the Lake B
gaging station, the major inflow to Lake Harrison, originated in
a watershed undergoing intense development at this time. Much
of the drainage system itself was being constructed under speci-
fications of the "natural drainage" system including Lake C,
656 ft (200 m) upstream of Lake B. Lake C was constructed to
serve as a wet weather pond and golf course water hazard. Un-
fortunately, its low earthen spillway had yet to be sodded and
provided an erosion source within the drainage channel.
Lake Harrison inflow and outflow hydrographs are compared in
Figure 28. Characteristic of runoff response in a small water-
shed, the multi-peaked inflow hydrograph was a product of the
sporadic hyetograph (Figure 27). Intense stormwater flow deepen-
ed the inflow channel by 6 in. (15 cm) and obliterated bales of
hay placed in the channel to act as flow control devices. Storm-
water flow crested shortly before noon on April 8 at a record
discharge of 123 cfs (3.48 m^/s). Bank storage and ponding
helped prolong minimum flow in the channel for two days, contri-
buting to the total inflow runoff volume of 93 acre-ft (11.4
ha-in) . The lake system effectively damped inflow fluctuations.
The hydrograph peak traveled through the lakes in a half hour.
Stormwater quality—Table 24 compares flow weighted mean and
maximum water quality concentrations for runoff sampled at Lake
Harrison inflow and outflow. Since runoff volumes were roughly
equivalent, a comparison of relative loadings is redundant to
the comparison of mean concentrations. Greater values for o-PC>4
NH 3, NO, and NCU indicate the outflow was nutrient enriched as
a result of one or a combination of two sources:
(1) Unmeasured runoff from the fertilized area adjacent
to the lakes and/or direct precipitation on the
lakes,
(2) The quality of water held in the lakes prior to
the storm event. (Water impounded in the lake
prior to the storm event approximated the runoff
volume.)
Lake Harrison served as an equalization basin minimizing the
difference between maximum and average parameter concentrations.
A prominent flush corresponding to the first peak of the Lake B
hydrograph was evident for most parameters at the inflow. For
example, nitrate concentration at the flush, maximum value, was
an order of magnitude greater than average concentration. This
flush was not observed at Lake A and average concentrations ap-
proximated maximum concentrations.
Sediment removal—Superimposed on the lake hydrographs of
Figure 28 are the TSS pollutographs. The reduction
91
-------
Ofrl
8
o
CM
8
o
o
£! z
UJ
O
o
00
o
o
o
OZI
001 08 09
SdO '398VHOS10
PQ
X
Q) g
r£ M
-P O
-P
-P tfl
rd
in
nj
O CO
4J
0) ^-1
a
cu
CP c
O -H
$-1 tn C
>< id .
ffi tn --1
CN
-------
2800 r 140
CO
s
CO
800-
k 100
en
UJ
en
o
CO
5
400- 20
0
2800 r
800
400
140
- 100
u
- LJ
O
tr
o
CO
Q
20
/
SUSPENDED SOLIDS
DISCHARGE
OUTFLOW
04 8 12 16 20
HOURS INTO STORM
Figure 28. Reduction of TSS through "The
Woodlands lake system. (1 cfs = .028 m /sec)
93
-------
TABLE 24. SUMMARY OF WATER QUALITY PARAMETERS FOR SITES
LAKE A AND LAKE B DURING THE APRIL 8, 1975
STORM EVENT
OUTFLOW
Lake
Drainage Area (acres) 483
Runoff Volume (ac-ft) 93
Rainfall (inches) 3
Concentration of Water
Quality Parameters : *
Avg.
Ortho-phosphate 0.015
Total Phosphorous 0.10
Ammonia 0.16
Nitrite 0.032
Nitrate 0.28
Total Kjeldahl 1.3
Nitrogen
Total Suspended 245.
Solids
Soluble Org. Carbon 13.6
Total COD 41.8
Soluble COD 26.4
Specific Conduc- 130.
A
.4
.97
Max.
0.048
0.19
0.26
0.046
0.32
2.
356.
19.
45
31.
215.
INFLOW
Lake B
337
93.2
3.97
Avg . Max .
0.005 0.013
0.11 0.36
0.11 0.15
0.009 0.054
0.15 2.1
1.86 3.1
1273. 2660.
16.2 22.
63.7 87.
32. 45.
85. 304.
tance (micromhos)
Turbidity (JTU)
160.
210.
375.
900
* all concentrations in mg/1 except when indicated.
1 ac = .405 ha
1 ac-ft = .124 ha-m
1 in = 2.54 cm
94
-------
of solids by sedimentation is a significant lake function de-
sirable in stormwater management. The high TSS con-
centration of 2660 mg/1 at inflow was reduced to 356 mg/1 at out-
flow. Detention in Lake Harrison reduced the stormwater sediment
load from 160 tons (145 t) to 31 tons (28 t), an 80% reduction in
solids, storing 129 tons (117 t). This mass reduced the volume
of the 110 acre-ft (13.6 ha-m) lake by less than 0.1% if 80 lb/ftj
(1282 kg/m3) is assumed. Erosion from the Lake B watershed was
effectively prevented from entering Panther Branch by the lake
system.
Table 25 shows the reduction in stormwater sediment load by
Lake Harrison for all storms monitored. All but one storm event
recorded over 80% solids removal. Complete removal, 100%, is a
result of total stormwater storage by Lake Harrison and does not
preclude discharge at a later time.
TABLE 25. STORMWATER SEDIMENT REMOVAL AT LAKE HARRISON
Total Suspended Solids Load During Storm Event
Storm #
8
9
10
13
14
15 & 16
Ibs Input
(Lake B)
104
13800
322000
6700
115302
4840
Ibs Discharged
(Lake A)
Flow stored
within lake
991
61900
Flow stored
1850
3270
% Load
Reduction
100%
93%
80%
100%
84%
32%
. Ib = .454 kg
Estimated value (Lake B gage inoperative) calculated
using estimated 10.1 ac-ft (1.25 ha-m) inflow times
sample average concentration, 421 mg/1.
95
-------
REFERENCES
1. Field, Richard, and Pauline Weigel. Urban Runoff and Com-
bined Sewer Overflow. JWPCF, 6.: p. 1108, 1973.
2. Wolman, M. G., and A. P. Schick. Effects of Construction on
Fluvial Sediment, Urban and Suburban Areas of Maryland.
Water Resources Research, 3_: p. 451, 1967.
3. Guy, H. P., and G. E. Ferguson. Sediment in Small Reservoirs
Due to Urbanization. J. Hydraulic Div., ASCE, 88, HY2 : p.
27, 1962.
4. Changes in Quality of Water in the Passaic River at Little
Falls, New Jersey. U.S.G.S. Professional Paper 525-D. 1965.
5. Geldreich, E. E., L. C. Best, B. A. Keener, and O. J. Van
Donsel. The Bacteriological Aspects of Storm Water Pollu-
tion. JWPCF, 4.0; p. 1861, 1968.
6. Storm and Combined Sewer Overflows Section, Environmental
Impact of Highway Deicing, U.S. EPA Report No. 11040GKK06/71.
June 1971.
7. Bryan, E. H. Quality of Stormwater Drainage from Urban Land.
Present at 7th American Water Resources Conference. Washing-
ton, D. C. 1971.
8. Heaney, J. P., and R. H. Sullivan. Source Control of Urban
Water pollution. JWPCF, 43; p. 571, 1971.
9. Sartor, J.D. and G.B. Boyd. Water Pollution Aspects of
Street Surface Contaminants, U.S. EPA Report No. EPA-R2-72-
081, 1972.
10. Poche, R. M. A Baseline Census and Development of a Monitor-
ing System for Important Animal Species on the proposed Wood-
lands Site, Montgomery County, Texas. Consultant's Report to
The Woodlands Development Corp., April 1973.
11. Maestro, R. Criteria for Residential Wildlife Planning in
the New Community of The Woodlands, Montgomery County, Texas.
Consultant's Report to The Woodlands Development Corp., May
1973.
12. Haas, R. H., and M. C. McCaskill. Use of Large-Scale Aerial
Photography in Obtaining Vegetation Information for Urban
Planning. Consultant's Report to The Woodlands Development
96
-------
Corp., July 1972.
13. Kendrick, W. W., and D. Williams. Soil Survey of The Wood-
lands. Consultant's Report to The Woodlands Development
Corp., July 1973.
14. Water Resources Data for Texas. part 1 Surface Water Re-
cords, 1973 and 1974. U.S. Dept. of the Interior publica-
tion, prepared in cooperation with the State of Texas.
15. Winslow, D. E., J. A. Veltman, and W. H. Espey, Jr. Natural
Drainage Systems. Paper presented at the Spring Session of
the Texas Section of the ASCE, Beaumont, Texas, March
1974.
16. Shelley, P. E., and G. A. Kirkpatrick. An Assessment of
Automatic Sewer Flow Samplers. U.S. EPA Report No. EPA-R2-
73-261, June 1973.
17. Methods for Chemical Analysis of Water and Wastes. U.S. EPA
Report No. EPA-625-/6-74-003, 1974.
18. Dugan, G. L., and P. H. McGauhey. Protecting Our Lakes:
Wastewater Treatment is Not Enough. Paper presented at
46th Annual Conf. of W.P.C.F., Cleveland, Ohio, Sept. 30-
Oct. 5, 1973.
19. Schicht, R. J., and F. A. Huff. The Effects of Precipita-
tion Scavenging of Airborne and Surface Pollutants on Sur-
face and Groundwater Quality in Urban Areas, part II.
NSF #GK-38329, National Science Foundation, Engineering
Division, Washington, D. C., July 1975.
20. Whitehead, L. W. Some Microclimate and Air Quality Impli-
cations of Urbanization in a Southern Coastal Forest. Dis-
sertation, University of Texas School of Public Health,
Houston, Texas, 1976.
21. Ward, C. H., and J. King. Final Report. Maximum Utiliza-
tion of Water Resources in a Planned Community: Eutrophica-
tion Potential of Surface Waters in a Developing Community.
EPA #802433, U.S. Environmental Protection Agency, Oct.
1976.
22. Diniz, E.V., and W. G. Characklis. "Modeling Urban Runoff from
a Planned Community," EPA Conference on Environmental Modeling
and Simulation, U.S. EPA Report No. EPA-600/9-76-016, July, 1976,
97
-------
23. Environmental Protection Agency. Storm Water Management
Model, in four vols.: Vol. I: "Final Report," 11024DOC07/71,
July, 1971, 352 pp.; Vol. II: "Verification and Testing,"
11024DOC08/71, Aug. 1971, 172 pp.; Vol. Ill: "Users Manual,"
11024DOC09/71, Sept. 1971, 359 pp.; Vol. IV: "Program
Listing," 11024DOC10/71, Oct. 1971, 249 pp.
24. U.S. Corps of Engineers. Urban Runoff: Storage and Treat-
ment and Overflow Model "STORM." Hydrologic Engineering
Center Computer Program 723-58-L2520, U.S. Army, Davis,
California., May 1974.
25. Nash, J. G. The Form of the Instantaneous Unit Hydrograph.
Int. Assoc. of Sc. Hydrol., Pub. 45, Vol. 3_: 114-121, 1957.
26. McKee, J. E., and H. W. Wolf. Water Quality Criteria.
May 1963.
27. Davis, Ernst. Final Report. Maximum Utilization of Water
Resources in a Planned Community; Microbiological Quality
of Stormwater Runoff in The Woodlands, Texas. EPA #802433,
U.S. Environmental Protection Agency, Oct. 1976.
98
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-79-050b
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
MAXIMUM UTILIZATION OF WATER RESOURCES IN A
PLANNED COMMUNITY
Stormwater Runoff Quality: Data Collection,
Reduction and Analysis
5. REPORT DATE
July 1979 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
William G. Characklis, Frank J. Gaudet, Frank L. Roe,
and Philip B. Bedient
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Dept. of Environmental Science & Engineering
Rice University
P. O. Box 1892
Houston, Texas 77001
10. PROGRAM ELEMENT NO.
1BC822, SOS #2, Task 02
11. OOCXDBear/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
13. TYPE OF REPORT AND PERIOD COVERED
Final 7/73 - 12/76
14. SPONSORING AGENCY CODE
EPA/600/14
15. SUPPLEMENTARY NOTES one in a series of volumes of one report.
Project Officers: Richard Field and Anthony N. Tafuri, Storm and Combined Sewer Sectio
FTS 340-6674, (201) 321-6674.
16. ABSTRACT
An ecologically planned community (The Woodlands, Texas) has adopted a unique water
management plan designed to avoid adverse water quality and hydrological effects due to
urbanization while benefiting from the existing natural drainage. The initial years of
development were monitored by a comprehensive sampling and analytical program in an
effort to evaluate the innovative new water resources concept. Data on water quantity
and quality were collected during dry weather and during stormwater runoff. To supple-
ment the prime study site, stormwater samples were also collected at watersheds in the
Houston area. Parameters monitored during the reporting period were as follows: rain-
fall, streamflow, chemical oxygen demand (COD), soluble organic carbon (SOC), biochemi-
cal oxygen demand (BOD), ammonia (NHs), nitrate (NO3), nitrite (NO2), total Kjeldahl
nitrogen (TKN), orthophosphates (ortho-P), total phosphorus (TP), dissolved oxygen (DO)
pH, turbidity, total suspended solids (TSS), and specific conductance. Data were
analyzed for water quality relationships in an effort to predict pollutant loads
according to land use. Comparisons were made to wastewater and rainwater quality.
Significant relationships were observed between total volume of runoff and total load
of various pollutants. The load-runoff relations are a function of the type of land
use activity in the watershed and have been used to simulate stormwater quality
responses.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Water quality
Urbanization
Chemical analysis
Demonstration watersheds
Hydrologic data
Hydrologic models
Overland flow
Water sampling
13B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
113
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (Rev. 4-77)
99
i US GOVERNMENT PRINTING OFFICE 1979 -657-060/5446
------- |