-------
m The source of the PCBs in The Woodlands is not known. It is
of interest that the peak of PCBs in all components of the eco-
system appeared during a period- of intense ct.t and fill opera-
tions as well as utility installations. An abandoned land fill
with disposed capacitors or other electronic materials could have
affected the increase in concentration as could the use of road
oil contaminated with PCBs. Neither of these possibilities were
verified by observation.
Chlorinated Hydrocarbon Pesticides—
In the first year of investigation trace levels of DDE were
observed in samples of crayfish, mosquito and bluegills obtained
from The Woodlands aquatic ecosystem. After completion of The
Woodlands Golf course, the surface waters and soils in ponds and
adjacent ditches were examined for halogenatei compounds.
In the spring of 1975 Mirex, a chlorinated camphene, was
detected in water, soil and some aquatic organisms (Figure 11).
The highest values were found in mosquitofish and the lowest in
water samples with the residues in soil being intermediate be-
tween the fish and water, if there was any biological amplifi-
cation _ in the fish from this aquatic system it was limited to
approximately*a four-fold increase (Figure 11).
Water from the Conference Center Lakes (A and B) was used
for irrigation of portions of the golf course, since these man-
made impoundments were the potential recipients of both irriga-
tion and stormwater runoff, the lake water sediments and the
mosquito fish were examined for Mirex. The results are shown in
Figure 12. The highest level of pesticide was observed in raid-
to late summer (1975) and thereafter concentration in all these
components of the pond ecosystem diminished rapidly with the soil
residues showing the slowest rate of decline. Mirex concentra-
tion in Lake Harrison was at least one order of magnitude lower
than concentrations observed on The Woodlands Golf Course. The
lower concentration due to lake water dilutior. by groundwater
pumpage, runoff and treated wastewater.
During August of 1974 chlordane was also detected in the
golf course study. Residues were found in soil, water and aqua-
tic organisms (Figure 13). The highest levels of chlordane were
found in the crayfish from golf course ponds. Somewhat less was
observed in the same organisms from ditches adjacent to the
course. Similar, although not as pronounced, results were ob-
served in the concentration of chlordane in the waters from the
same areas (Figure 13).
Bacteriological Enumerations
Surface Water Characteristics—
Bacterial counts during dry weather were determined in
48
-------
60
50
40
MiREX
(PPB)
30
20
10-
0
o
ASONDJ FMAMJJASOND
1974 1975
Figure 11.
Temporal distribution of Mirex in
The Woodlands gold course.
•-• mosquitofish (Gambusia sp.);
O-O soil; o-o water.
49
-------
40 -
30
Ml REX
(PPB)
10
n n
ASONDJFMAMJJ A~S 0 N D
1974 1975
Figure 12.
Temporal distribution of Mirex in the
Conference Center Lakes (A&B).
d-D soils; •-• mosquitofish (Gambusia sp.)/
o-o water.
50
-------
40 -
30 h
CHLORDANE
(PPB)
20
10 -
Figure 13 .
ASONDJ FMAMJ JASOND
1974 1975
Temporal distribution of chlordane in The Woodlands
gold course A-A crayfish (Carribarus sp.)from pond^
A-A crayfish (Cambarus sp.) from ditch? o-o pond
water; .-. ditch water; mosouitofish (Gambusia sp )
from pond; a-O soil from pond. — sp';
51
-------
Panther Branch. Indicator bacteria numbers varied widely as
indicated below:
Total Coliform (TC)
Fecal Coliform (FC)
Fecal Streptococci (FS)
pseudomonas aeruginosa (PS)
Staphvlococcus sp. (ST)
Salmonella sp_. (SA)
Range (number/I00 ml)
200-10,000,000
10-214,000
10-1,580
10-53,600
1-23,400
1-5,800
The maximum values are comparatively high for a rural, forested
area receiving no direct sewage discharges.
Storm events monitored for water quality were also monitored
for microbiological content. The range and mean bacterial counts
observed during each storm event at The Woodlands are presented
in Table 12. All bacterial counts are reported as logarithms to
base 10. Substantial numbers of bacteria were observed in The
Woodlands storm runoff, including runoff from the undeveloped
watershed, site P-10. pathogenic species were identified at all
Woodlands monitoring sites in relatively high numbers. For ex-
ample, Salmonella sp_.were identified at a mean count of 77/100 ml
for storm event #10 at site P-10. In the majority of events the
maximum bacterial count occurred before or coincidental to the
hydrograph crest.
A summary of the three storm events monitored at Westbury
Square is presented in Table 11 (the storm event of 11/75 was
not monitored for water quality).
TABLE 1.1. SUMMARY OF WESTBURY RUNOFF BACTERIAL QUALITY
Storm
5/75
6/75
(n=12
TC
FC
FS
PS
ST
SA
TC
FC
FS
PS
ST
SA
log No./lOO ml
Min. Max.
8.70
4.60
4.61
4.40
4.30
Mean
7.48
4.34
4.12
3.88
3.91
(continued)
52
-------
TABLE 11 ("continued-!
log No./lOO ml
Storm
11/75
(n=5)
1
TC
FC
FS
PS
ST
SA
Number of
Min.
6.26
3.96
4.39
4.31
TNTC
2.00
Samples
Max.
6.67
4.50
4.68
TNTC2
TNTC
4.17
Mean
6.49
4.28
4.51
4.54
3.19
2
Too Numerous to Count
Stormwater bacterial counts at Westbury were aen«*r*Ti
greater than or equal to counts observed at^hJ WoolKSs
sample variations were common for P
-------
H
CO
tN
H
^i
o s
o
,-j
s
*0
tfS
SI
M
B
8
S
TJ
fl
CO
o|
0
• 4
a 5
m
o
5? c
ai
,.
4-
1
c
c
•r
n
a
co VD oo r~ in "=r
ro oo ^ oo ^j* CT»
in ro ro CN ro rH
CN OO -<3< O O CN
O CN CN O OO -^
vo *3< "3* co co CN
CO O O O O C>J
^< vo o o cTi in
•<3« CN CN CN CN rH
a K e s R a
r~
in o
\ ro
CN 1
rH ft
CN 00 rH
a\ vo vo i i
• ••III
T* ro ro
r~ CN o
ro ro ro I 1 1
• • • I I I
in in m
en r- «* i i
i i i
ro CN CN
o y co co EH rt
^
r>
oo o
•H ro
> a,
ro o in in r~ oo
rH CN VO rH r- 00
in ro CN CN CN rH
CN v£> vo t — ro CN
ro o *3* t — r~- in
VD ^ ro CN ro CM
O O O O O
o o o o o ro
-3« CN CN rH rH rH
V V
CJ CJ CO CO B f<
V.
m o
\ rH
CN 1
in CTI ro
rH 00 f- 1 1 1
• ••III
r- ro ro
ko ro ro CN
r- iH ro "* 1 I
• • • • 1 I
o o in oo
o CN en r- i i
• • • • i i
VD CN rH CN
o y co co EH <:
^
v^
CN 0
CN ro
\ 1
r- o r~ o o
rH o ro oo ^r
in ro ro H ro
rH CN ro o VD
VD ro o o H
in ro ^ CN «*
oo oo in oo en
vo in oo "* en
«* CN CN rH CN
y p co co EH
EH fa fa ft CO
m
\ o
•<3< CO
ro ft
oo ro ro vo r-
cn VD r- ** r~-
in ro ro ro ro
o in in in CN
en vo rH •ctf in
vo ^j* in ^i4 m
00 O O O O
r~ o o ro o
«* CN CN CN CN
E^ % S S &
f"**
CO
-------
•8
n
f4
-iH
-p
C
o
u
CN
i-H
11 1
§*
rH
6 >
2 fl
O
M 2
£
C
•i—
c
e
C
£
C
to
S
o s
o
rH
§3
m ^
t7> CJ
Q "jd
M S
3
«J
H
fi
I
B
CO
CTl
in
in
0
ro
•3*
%
in
£^
CN
rH
ro
CO
in
in
ro
o
•
vo
vo
ro
in
&J
I-O
[""••I
\
ro
ro
Cn
CN
3
ro
O
O
rH
V
K
o
H
ft
co
ro
ro
o
en
*
ro
O
vo
rH
fej
«
n
M
t~-
00
CM
ro
ro
o
0
rH
V
ft
rH
CN
•«*
00
ro
•
**
vo
ro
ft
in CN ro
in in in
rH CN rH
en o o
VO ^ rH
CN ro CN
000
o o ^
rH rH rH
V V
ft % $
CTi CN O
ro rH ^
ro «* rH
\/
ro in o
OO ^J* ^
ft ft ft
ro 'tf rH
V
r^ r^ o
VD VD ^
CN ro rH
V
ft & $
OO
rH
VO
in
t^
vo
0
vo
^3*
^
in
v^
CN
rH
ro
CN
in
•<*
0
r-
•*
in
00
ro
#
in
r-
^
ro
CN
ro
vo
ro
0
o
CN
V
K
CQ
(U
1
in
00
o
ro
•
rH
0
0
•
rH
V
&
<
"1
l-H
ro ro ^ in
in r- o CN
ro (N rf rH
ro co ro rr
en m CN CN
ro ro in CN
o o in o
0000*3*
CN CN CN rH
V V . V
CO CO EH <3
pn pt tO CO
VO CN ** O
rH in rH •*
• • • •
rO rH CN rH
V
CN ro •<* o
ro CN CN rH
V
O 0 0 0
vo o o «*
CN rH rH rH
V V
£0 CO EH rf
rH ^*
*^f* rH
en in
^1* rH
in CN
0 ' 0
O 0
^t* i~H
V
B K
in
\ ***
CN . vo
vo ro
00 O
r~ o
y &
in
r-
CN O
•H ro
ro" fl
CN
rH
ro
ro
in
ro
00
00
rH
r°
b
o>
CN
CN
cn
e
ro
o
0
*
__|
r^
S
r-
m
en
rH
o
o
r-|
V
s
CO
in
rH
in
CO
•
H
O
O
1
I— 1
V
ft
o r~~
m »*
rH rH
in o
on **
rH CN
O 0
o ^
rH rH
V
EH
-------
•a
0)
•H
4J
cs
o
o
r~*
pa
£
«
<
B
tii
o S
o
> .
1
o .
tr> c!
Si
1
Station I
rH C
Ei "Q
o S
o
rH
* •
& I
si
^
1
'•r
£
f-
|
|
u
VD ro o rH crj in
o o in VD CM ro
in ro ro 01 ro
Ol O VD VD VO O
vo vo o ro ro o
in ro "S1 ro ro oj
o o o o o rH
o o o o ro m
•<3< Ol 01 Ol 01 rH
v v
HKBRBtf
in *3
oo oo r^- r** vo CNJ
^* CO ^* VO C"*4 C^*
to ro ro oj ro r-H
0 0 01 rH in 04
VD ^ 'd' ro ro 01
t-~ O O O O rH
r~ o o o ro m
ro 01 01 01 01 H
p O en CO EH rf
EH pL4 P4 CM to c/j
in
r^
' \ 0
i c: 7
! > d
oo en 01 01 g o
o o r~ VD ro o
^D ro in «* "* r~i
•^i o cr> o o o
•<* o in o r- o
r> r) CO CO &•< ^
m
^ o
^ 2
crv CM
oo VD in in oo cr>
rH ro o ro co oo
• •••••
in ro ro 01 01 rH
CTi O "* VD in VD
r-- 'S1 04 oj -31 ro
vo ^ ** ro ro CM
Ol O O O O rH
ro o o o ro in
^r 01 01 01 01 rH
v v v
& & g S g 3
in
r-
"\ o
\ I
•* CM
56
O rH CT O O
ro *3* < — VD VD O
531 Ol CN rH rH r—1
CO CO f' O CO O
in r-- v£i cr> 01 o
IT) Ol Ol rH Ol rH
88-8888
tf "' v1 v
o o' 01 co fc fjS
u tj_j Jj,| p^ Ua uJ
m
S s
of cu
in ' r~- c"> rH r-- co
01 ro r-i co rH CM
in ro n 01 ro
o ch o r~ ro rH
co ro "V in oo in
m ^f T ro ro 01
O O CD O CO rH
O O O O ^J* in
•vf Ol CM Ol Ol rH
V
g B 12 B S-fii
in w
r~-
r^ w*
\ 5
•^ M
•H
fl
O
-------
ts
Q)
CJ
•rH
a
o
o
CN
m
PQ
^1
0
H
g >
MH
O
S3
c
•f-
"5
c.
H*
6
•r-
-i-
4j
to
5 (Q
o s
o
rH
,O X
5
?l
i-4 g
§
1
g
s
CO
i — 1 CO VO CT\ OO tf
in OO OO CN CN rH
CN rH O"i f» OO C —
O CO OO O^ n* rH
VO «vf -^ OO OO CO CO EH rt* CN CN
CN oo r- vo
• • • •
^* ^1* CO r-H
^ p*- to r**
*>3* ^3* LO ^^
•^j rH r-- o
OO OO CN rH
V
CO CO EH -f£
&4 Oj CO CO
•a
-------
rH C
6 ro
o s
0
rH
S
UH
O
Sc
i
M
B
8
1
5
CO
*I
o s
o
s d
52 j9
UH
0
Si
1 |
« T
8 ^
3 -1
H 4.
W 1
; i i i : '<
a s £ 3 o §
in co co co co rH
V
o oo oo in vo o
ro r~ rH rH CO O
*3< CM CO CO CM rH
B a a s K a
VD
\ o
r^ co
CO Ol
1 i ! i ! i
CM O (T> ^ rH O
CT> oo f~ o ^ r~
in •«* co co "sf CM
O «* CO O O O .
ro o CM co r~ o
in co co CM co CM
fi jjtj fe ^ 55 en
i m CQ
i r~-
' ^ S
1 CO 1-4
1
i i i i : :
,H ,-H CO in 0 0
rH cri oo en in o
in CO CO rH •si' CM
VD 00 00 0 Op 0
ro r- r~ o CM o
Ttf CM CM rH CM CM
o o cn cn EH rt
P K En fl< &3 03
VD CQ
s "
CO I-H
:!!:::
T* in ** o oo o
o ~l ^o
•«^ ro co CM co CM
o ^ o o in o
CM VD F- O <3" O
• •••*•
rjt fM CM i-H CM CM
O O W CO EH <
f) ftl |i| P-i UJ CO
vo
\ o
r- rH
ro- 04
0)
3
fl
•rl
rj
o
w
58
-------
!i
.0
•H
SI
° .
!
s
g
s
CO
rH. M
R (O
O JaJ
o
rH
O
si
Q) &
1 S
t) 'r-j
C! "g
O fl
CN 8
i
g &
M
i i i i i i
i i i i i
H oo rH o in o
m r~» rH r~- «* o
vo ro ro CM •TP CM
H o m o in o
oo en ** o "3" o
"* CM CM CM ro CM
V V
O O CO CO EH rt
u) pq
\ m
in ^
i i t i i i
i i i i i i
U3 vo ro rH oo o
<7I 00 O rH rf O
*>f CO "* CM »* CM
V
oo o vo o in o
o CM m o a\ o
•^ ro ro rH ro CM
V
^""
VD
\ O
§ 2
i i i i i i
ii i i i i
P** G^ G\ O C^ O
CTi ( — O CTi OO O
CO CM CO rH CM rH
V
VO OO CTi O O O
r- »* vo o o o
ro rH CM rH CM rH
V
RK**H*
VD <|
in ,x
( I I I i I
I I i I I I
^o o vo o oo o
CM in r~ ro CM o
• •••..
"* CM CM CM CM rH
CM O O O OO O
r~~ CT> ro o «3< o
rO rH CM CM rH rH
V V
U CJ CO CO EH f£
vo
\ o
S 2
-
i ! i ! ! !
r- o cr» o cr> o
t-» o m o in ro
m ^ ^j* ro ^3* CM
o o o o o o
o o o o o o
rO CM CM CM CM CM
V V V V
P fa fa CU W CO
\ 0
in ro
59
-------
TABLE 13. SUMMARY OF PANTHER BEANCH AQUATIC ALGAE
Division
Number of Species
Identified in Panther Branch
Chlorophyta
Cyanophyta
Chrysophyta
Euglenophyta
Pyrrophyta
92 a
15
16
1
1
43
8
12
15
1
79 Total number
a number of taxonomic species/division
Standing crops of algae at site P-10 and P-30 are presented
in Figures 14 and 15, respectively. On a seasonal basis, algal
cell numbers fluctuated more at site p-30 than at p-10. The
algal standing crops at both sites were dominated primarily by
members of the euglenoids (Euglenophyta), and green algae
(Chlorophyta) were minor components of the total algal popula-
tions. Numbers of blue-green algae (Cyanophyta) were compara-
tively low at both sites, even though they were more numerous at
site P-30. A survey of various collection sites along Panther
Branch (May, 1974) also indicated that standing crops in this
stream were dominated by euglenoids and/or diatoms (Chrysophyta),
while numbers of green algae and blue-green algae were compara-
tively low. Dominance of this type is indicative of slightly,
acid streams with high organic carbon content.
Soil Algae—
Fifty-two genera of algae were identified in soils collected
from disturbed and undisturbed sites in The Woodlands (see Table
14). Undisturbed soils had larger algal species diversities and
fewer algal numbers than disturbed soils. Green algae (Chloro-
phyta) genera were most numerous in soils from the forest, while
disturbance of soils favored the development of a more diverse
blue-green (Cyanophyta) flora and inhibited green algae diver-
sity.
The change in algae characteristics is probably a result of
increases in soil pH and nutrient content which accompanied soil
disturbance and fertilization. Undisturbed soils had the lowest
pH (6.1), compared to high pH values of 6.7 to 7.8 in disturbed
soils. The results confirm floristic surveys in other regions
of the United States which show that alkaline soils favor the
development of more luxuriant blue-green algal flora than do
acid soils (41, 42, 43). Higher algal numbers in disturbed soils
corresponded to the greater concentrations of nitrogen and phos-
60
-------
CHLOROPHYTA
CYANOPHYTA
CHRYSOPHYTA
EUGLENOPHYTA
TOTAL
NDJFMAMJJASO
MONTHS
Figure 14. ^Seasonal algal standing crops at P-10
in Panther Branch.
61
-------
I I I I I I I I I I I I
Chiorophyta
—— Cyanophyta
—x— Chrysophyta
Euglenophyta
NOJFMAMJJASO
TIME (mo)
Figure 15. Seasonal algal standing crops at
P-30 in panther Branch.
62
-------
phorus found in these soils. The increase in soil nutrient con-
centration is due primarily to fertilization.
TABLE 14. SUMMARY OF EDAPHIC ALGAE IN THE WOODLAEIDS
Undisturbed
Disturbed Soils
Division Forest Soil Golf Course Lawn
Chlorophyta
Cyanophyta
Chrysophyta
Euglenophyta
Total
27a
3853b
586
{' 11
1297
3
46
5736
7
39629
4
1446
4
13084
1
16
54159
4
17400
4
7195
3
12790
— —
11
37385
a numbers of
determined)
b , -
genera identified
(species
were not
Since terrestrial and aquatic ecosystems are connected
hydrologically, they cannot be considered as totally disjuncted
units. Surface drainage serves as a major component in this
hydrologic linkage and is thus an important ecological parameter.
Land usage in a watershed can determine the quality, as well as
the quantity, of surface runoff, thus influencing aquatic
habitats in the watershed. In addition to providing potential
nutrients for algal growth, surface drainage probably transports
algal cells to aquatic habitats from surrounding soils. Since
land use affects edaphic algal populations, it must also influ-
ence the diversity of algae which could be transported by sur-
face runoff. For example, a more diverse assemblage of blue-
green algae could potentially enter lakes in the study area by
surface drainage from disturbed, rather than undisturbed,•
soils.
Stormwater Comparison with Sewage
Table 15 compares stormwater runoff and sanitary wastewater qual-
ity. Stormwater constituents are lower than raw sewage with the
exception of suspended solids in runoff from Hunting Bayou and The
63
-------
fa
0
gj
§
CO
Q
W
EH
M
s
EH
p
§
X
O
rfj
s
fa
CO
Q
H
EH
gj
EH
J2j
P
fa
I
P**j
P
01
PS
EH
Si
§
O
EH^
%%
>H !3
EH H
H
, 1 rft
HH **••
gs
Ot ^
OS
fa Kl
EH J
& 5.
*
m
rH
ij
CQ
,-jj
EH
H
(d 'd
o c
•H (d
e g
0) CD
j; p
o
a
rH CD
(d tn
-P ^i
O &
EH 0
CO
(d o
4J ^f*
o a
EH CO
O
ft
^
^-4
(d
*d c!
i fli
n™i vy
0) Cr>
•n O
K SH
jj
•t**
rH -H
(d 3
I 1
•H
o
"
CD
rH >d CO
-P CO -H
O P|"~l
EH W O
3 CO
CO
H
fa
0
1
W
<
&
o
EH
W
co "*
VO in rH •<* rH in
1 I I I I I
ro o o\ co r*1* o\
^t1 ^J* ^ ro r**» ro
o^ o ro ^}* oo ^J1
o ro in rH CNJ H
O O 0 O H rH
1 1 t 1 1 1
co co i — i in i — i co
o H H o •<* r-
o o o o o o
rH rH 'J rH <* O\
VD ^ H CO CT> rH
H H ^ H CO CM
1 1 1 1 1 1
O VD O> t^ VD CO
rH o r- H m •*
CJ O rH O rH rH
H r- in t~-
r— CM co "5f o o
VO CO CM CM CM r-
1 1 1 1 1 1
in cri ro •<* I-H -sj1
• O CO CM t-« CM
VD rH CM
1
rH
CO
^
CO *-+ CD
P 0 }H
^-* (*) 3 id
o M i OS
rH CO ft ^1 D1
1 rg <_- (d CO
ft £ PQ
- G CT> M
CO CO g -H ,Q O
SnMOcXrXCco !5
O O (d (d 3 CO W
pti fa i-3 i-3 W |2 CO
o o m
in '* H
CO
O O rH
H rH
o o o
d o
CO fl C!
in o (d
•P u >
C 0 'd
p co <
CO
•P
fl
CO
CO
H
n
o
to
I>1
m ly
o c
CO
JH -rl
CO O
fit 1
It 1
c! fa
SH d
O CJi
II t „ 1
to
to d)
s p
•rl 1
•p
(d -P
»H d
•p rd
fl rH
CO ft
O
c a
0 0
O -rl
•p
G fd
fd g
(D (d
S o
I s
t^! M
tJi CO
•rl -P
f &
•* rn
of flo\
ored.
oodlandi
0) -P !3
c! C CD
id o ,£
& E EH
•H CM
64
-------
Woodlands developing areas. Advanced wastewater treatment yields
lower GODandTSS values than stormwater, however nutrients remain
higher. AS a result, sophisticated nutrient removal processes
would be necessary to achieve low concentrations observed in
stormwater runoff.
Ten samples of untreated and chlorinated secondary waste-
water effluent were obtained from two large, well-operated
sewage plants in Houston for bacteriological examination. The
data obtained is compared to the bacteriological quality of
stormwater runoff in Table 16. Bacterial numbers were lowest in
chlorinated effluent, higher in runoff and highest in untreated
wastewater. Urban runoff contained higher numbers than forest
runoff.
OBSERVED TEMPORAL AND SPATIAL VARIATIONS IN STORMWATER RUNOFF
QUALITY
Pollutograph Analysis
A pollutograph is defined as a plot of pollutant concentra-
tion versus time. In this sub-section it is plotted for storm-
f!6w quality. Temporal changes of water quality during runoff
events are important to the understanding of the impact of these
non-point sources on stream quality. The time-concentration rela-
tionship _ is also critical in consideration of stormwater treatment
alternatives. Pollutographs observed during the study exhibited
the five generalized patterns shown in Figure 16. These concentra-
tion patterns were common to all watersheds, although levels of a
particular parameter 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 16.2. This dilution pattern applies to
other streamwater constituents, including7 some found in waste-
water effluents, which are concentrated in dry-weather flow. The
concentrations of SOC, soluble COD and total COD often increased
as runoff progressed, with highest concentrations observed at the
end of the runoff (Figure 16.4). Streamflow contributions from
interstitial and bank storage flow is greatest late in runoff
and could account for the pattern if enriched by contact with
soils serving as an organic carbon source. DO concentrations
in stormwater increased proportional to flow and assumed a
hydrograph-shaped pollutograph (Figure 16.5). Increased re-
aeration at greater streamflows accounts for this phenomena.
Several parameters observed at site p-10 remained at a constant
level throughout the hydrograph, including pH, NHo, NO, and
soluble COD (Figure 16.3). This pattern was not commonly ob-
served at the other watersheds where land use is diversified.
65
-------
CD
O
pa
o
oi
o
OS
w
E-
<;
3=
w
H
en
a
2
Pi r-H
w s
H
< O
S: O
OS
O
H
en
erf
m
CO
s
O
CO W
t-t t-t
erf co
< >>
cu j
S <
o ss
CJ <
VD
r-l
CO
E-
H 01
IS 4-»
Hunici
Was tew
Ji
4J
n
*
c
I
"1
P3
4J
C
3
S
M
P.
go
ra c
•00) ,
S3
0 O
fra*
c
I'St
l< 4J 0-0
Untreate
X 3
4J
• C
en a
ss.
x§
ss
WJ
* C
D» 3
S3
• C
II
4J
X §
S3
!"£
II
10 r-
•r co •
co r- «>
r-C iH 'r-(
XXX
IN O CO
0 O O O O C
X X X X X fc
ft *»
m in » CM w
CO •» •* ^ •» "O
X X X • X X tn
t- «> f> en m -u
tn in tn in ^H o
rt M M H H
CD
o
:!!!!!
« I 1 ! 1 i
n
n
H
» » » „ n
IM
X X X X X n
Ol
vo in \o n to
in ft ft f
000 0
* x « S « S
O^*O CD
n (M n H
*^a **o ^a *"o o
» r- H ^ «
o o o
H rt ^ ^ o
X X X w jo co
•-C ^ M
in ft r*t
f-l N H
L'
0) • «)
u i i nJ >H
E E 4J 0 Oln) pH
C M W-H E • -HI V
o o u o oJ >itn c
,H (M pH **-» ^ O T3 «1 j:]D Q
ni*ri n -H flj o 3 oJO E
4Ji-l U>H O O OH fljIU i~4*
es as £s ssl sis isa
OJ
c
0)
sewage treat]
I
I
i
•o
01
u
1
n
at
n
1. Averaged
n n n a
: 0) t" D^ DI
(0 0) (U 0> 0)
(N PI ^ tn
66
-------
O
s
cc
*-
Ixl
O
z
O
O
UJ
UJ
ac
4
UJ
o
tr
FIRST FLUSH
DILUTION
STABLE
INCREASING
HYDROGRAPH
TJME INTO STORM EVENT
Figure 16. Generalized pollutographs observed
for storntwater parameters.
67
-------
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, ortho P,
TKN, TP, NH3, NO2 and N03. The Hunting Bayou watershed has a frac-
tion of impervious surface area and includes a diverse land use
mix including industrial and commercial activities. Turbidity and
TSS parameters exhibited the highest peak values over baseline,
while peak values for other constituents were,less pronounced.
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, NC>3, NH3, soluble COD
and SOC. Correlation coefficients for a majority of the parameters
were greater than r = 0.8 (r = correlation coeff.), however, three
cases showed fair to poor correlation; P-30, NO3 (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 Westbury, P-30 and P-10.
TSS loads are highest in P-30 as a result of construction
activities in the watershed and urban runoff solids loading is
greater than forest runoff (Figure 19).
Runoff loads for nonspecific parameters ftota] COD,.soluble COD
and SOC are shown in Figures 19-20) are higher in forested water-
sheds 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, insoluble 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 pollu-
tant. Table 17 shows these rankings for mean regression values
at one inch (2.54 cm) of runoff for all parameters and sites.
Confidence intervals (95%) are included in Table 17 to indicate
significant differences in pollutant loads at one inch (2.54 cm)
runoff. Confidence limits for Westbury show a particularly large
spread due to the small number (2) of storms monitored for that
watershed. Significant differences for the total COD, soluble COD
and SOC cases are indicated, with some overlap in the TSS case.
The patterns of nutrient response for the urban developing and
forested watersheds is distinctive, with the urban response pro-
ducing loads up to an order of magnitude larger. Hunting Bayou
ranks as the producer of the largest pollutant loads.
68
-------
1
o
k
LU
O
0.
<
i
o.
O -P
m a
-------
CE
IU
o
ov/san
u.
LL
O
£L
<
Q-
I
Q.
O
Ul
X
O
CO
o
(O
o
•3-
O
w
o
o
o
ro
W
m
ro
O
O -P
(U
to -H
(i O
•H -H
W M-)
Cl (1)
O O
-H O
-P
OJ -H
M^
11 r [
m
-------
I
00
UJ
X
o
to
o
a.
O
P
o
o
H
(0
-p
o
4->
d 'd
(0 0)
•H
tots
co 3
E-i -P
03
5-1 •—
O tn 4-1
""* (1) 0)
W rC -H
& 03 O
•H S-i -^
& cu m
tD 4->
4J fd
m rd Q^
O S-l
3 m o
MHO
m
O 03
en
rH
Q)
tr>
•H
e
o
in
CM
II
£
•H
ov/sai
71
-------
!£
(C
:D
(0
to
Id
*n5>
o
cr
i
a.
c
a.
O
u
o
CO
Q
O
U
H
•§
H
O
W
O 4J
W -H
O< O
•H-H
w m
c
-------
TABLE 17. POLLUTANT LOAD RANKING OF THE FOUR STUDY
AREA WATERSHEDS
RANK
2 3
ss
ECOD
3COD
soc
NO3
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
1-2.87
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 Level. All confidence
gression values at 1 inch (2,. 54 cm) of
in Ib/ac.
P30 = Woodlands P30 Watershed
P10 = Woodlands P10 Watershed
HB = Hunting Bayou Watershed
WB- = Westbury Watershed
Note: inch x 2.54 = cm, Ib/ac x .184 = kg/ha
levels are for mean re-
runoff. All loads are
73
-------
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-
flow and concentration.
A comparison of annual loads for TSS, TP, NC>3 and total COD
is shown in Table 18 for the P-10 forested site and P-30 urbanizing
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 natural background or point source
origin. Consequently, non-point loads can be quantitatively deter-
mined as a function of land use patterns as more storm data becomes
available from other urbanizing watersheds.
The annual 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—
Storm event #10—Stormwater quality monitored at site P-10
represents runoff from a forested, undeveloped watershed and ac-
cordingly 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 runoff from 5,500 acreas (2250 ha), in addi-
tion to the area monitored by the P-10 site. The additonal area
includes construction activity of The Woodlands Developed Corpora-
tion. 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 of drizzle. The area rain
gages measured 2.00, 2.65, 3.42 and 3.97 in. (5.08, 6.73, 8.64
and 10.1 cm) of rainfall, upper to lower watershed gages, respec-
tively, with the Thiessen adjusted rainfall calculated to be
2.43 in. (6.17 cm) on the P-10 watershed, and 2.76 in. (7.01 cm)
on the P-30 watershed. Average rainfall intensity was 0.76 in/hr
(1.83 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' 74
-------
VD
CO
in
(0
oo
II to
\
rQ' pQ
H <-{
rH H
75
-------
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.
Site
P-10
P-30
Streamflow Volume
acre ft. ha-m
1614
2829
(199)
(349)
•Runoff
in. (cm)
1 . 2
1.57
(3.05)
(4)
Peak Plow
cfs (m3/s)
1170
1100
(33.1)
(31.1)
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 asso-
ciated 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 in-
creased although no significant change has been observed for
oxygen demand. Results for storm event #10 are summarized in
Table 19.
76
-------
656 TONS
84 TONS
l i
a
o
o
O
ixi
S
-------
0.4.
_j 0.3.
•^
s 0.2.
O.I .
0.0
675 LBS
264 LBS
z
t-
4.0.
3.0.
2.0.
J.OJ
0.0
til
O
£E
<
o
en
5
en
8
UJ
I
I
120.
2 90 J
x
CO
u.
o
60.
3O.
1065 LBS
6OO LBS
2829 AC-FT
1614 AC-FT
15 3O 45 6O 75 9O
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, TKN and TP for the storm event of
April 8, 1975.
1 Ib. = .4536 kg
1 ac-ft = .123 ha-m
1 cfs = .028 m /sec
78
-------
CONSTRUCTION AREAS
PROPERTY BOUNDARIES
1 mi = 1.6 km
1 ft = .305 km
FEET
Figure 23,
The Woodlands construction activity in
relation to the P-10 and P-30 sampling
sites.
79
-------
TABLE 19. COMPARISON OF STORMWATER QUALITY AT P-10, P-30
AND DEVELOPING AREAS DURING STORM #10
Forest
with
Development,
P-30
The Woodlands
Forest Development
P-10 (P-30) - (P-10)
Drainage Area (acres) 21,606
(ha) 8,750
Streamflow Volume (ac-ft) 2,829
(ha-m) 349
16,050
6,SCO
1,614
199
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
22.0
58.7
43.0
0.021
0.125
0.244
0.017
0.272
1.41
347
18.2
44.5
33.2
FC/FS bacteriological ratios and land use--An important com-
parison employed in examining data is the relationship between
FC and FS concentrations. A FC/FS ratio > 4 suggests the present
of human wastes. Between 2 and 4, the FC/FS ra.tio may suggest
human wastes mixed with other source materials, and 1 < FC/FS <
2 value represents an area of uncertainty. If values fall be-
tween 0.7 and 1.0, a predominance of livestock or warm-blooded
animal waste may be suggested. Following the latter range, FC/
FS values less than or equal to 0.7 strongly suggest a predomi-
nance of warm-blooded animal waste other than human wastes.
Mean FC/FS ratios for storm events, low flow and sewage de-
terminations are plotted in Figure 24. Note that FC and FS
bacterial numbers in The Woodlands stormwaters are greater than
low-flow waters, however the ratio remains the same. Urban run-
off at Westbury also exhibited a similar ratio although the
bacterial numbers were much greater. High ratios for the sewage
determinations confirm human waste contamination. Lake B runoff
contained the lowest ratios indicating stormwater impoundment
may have beneficial effects or that organisms attached to sus-
pended sediment. TC, FC, FS and ST bacterial species were all
observed to settle in quiescent water. Additional data have been
reported by Olivieri, et al., (44).
80
-------
81
-------
A least significant difference statistical analysis was ap-
plied to FC/FS ratios from all stations resulting in the follow-
ing sequence:
Location FG/FS (Geo. Mean)
Storms, Lake A
Storms, Lake B
Low Flow, Lake A
Chlorinated Secondary Sewage
Low Flow, P-10
Storms, P-30
Low Flow, Lake B
Low Flow, P-30
Storms, Westbury
Low Flow, Woodlands
Storms, P-10
Raw Sewage
Secondary Treated Sewage
0.17
0.53
0.58
0.73
0.92
0.97
0.99
1.26
1.47
1.68
2.11
2.42
13.30
Those groups which demonstrated no significant dif-
ferences between subsets are indicated by the verti-
cal lines on the side.
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-
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 (Ib) 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 (45). Another water quality modeling approach is dis-
cussed below.
pollutant Load Modeling for Multiple Events
The load-runoff relationships presented above (Figures 17-
20) provide the'foundation for an uncomplicated, yet satisfac-
torvT 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.
82
-------
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 of
an.
kech watershed. Initially three parameters 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 elope can be roughly determined from the cumulative rela-
tionship produced from field data. The initial slope value depends
primarily on initial conditions, and the range variable is deter-
mined 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 calibration. During a storm
event the value of the slope decays exponentially by the same
means employed in both the SWMM (46) and the Storage, Treament, and
Overflow Model '(STORM, (47).
libs pollutant
washed off in
any time interval
Lbs
remaining on
the ground
or:
-dP
dt
= kP
(1)
which when integrated takes the form:
P0-P = PQ (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. (!•.27 cm)
of runoff uniformly delivered in 1 hour washes away 90% of the
pollutants (22). As a. result the equation can be written:
(4)
83
-------
The equation used to decay the load-runoff line slopes is:
-4.6 rt
PDS = 1-e
(PDS),
(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, 1 and 10 monitored
during the study. Storm events 8 and 9 were considered too small
for use in the simulation. Predicted solids loads and the ob-
served streamflow hydrographs are presented in Figure 25. 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 26, the simulated curves compare satisfactorily to the
observed mass flow curves. Table 20 gives comparisons of simu-
lated to observed values for total pounds TSS, and peak magnitudes
for each of the three storms.
t
Unit Loadoqraph 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 (48)
Qn = kr(n)
(I) ^
exp(-tA)
(6)
where S = total storage (one inch (2. 54 cm)of runoff); k = constant;
n = outflow from n^*1 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 reservoirs can be extended for mass
flow curves in order to develop a corresponding unit pollutoqraph
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
84
-------
S/9
sanos aaaN3dsns ivioi
SdO
Sn03NVlNVlSNI
85
-------
to
0) 0)
en
bO
o
Tf I
CJ
10 CJ
OI X S/9
sanos
TYIOI
10 ea —
?OI X SJO
SnO3NVlNVlSNI
10 O
f) 4J
O CO
fi
CO
•a 8
fd m
s-i
en to
o a)
M >
VO
OJ
<0
%
O
-------
g
CQ
04 Pi
CQ H
CQ H!
O CM
H
Q 05
S E^
CQ
EH vo
^*
CQ
< EH
.«
o 0
CQ
(tf
CO
eu
H
«
£4
^
1
^|
4->
rt
3
S
CQ
S CD
i-i
•P 3
CQ a
«5j<
in
o
o
in
CM
ro
0
o
vo
CM
T^ |
CM
in
iH
0
0
CM
m
ro
r-l
0
O
r-4
*3f
O
CM
CM
0
0
crv
CO
f^
CM
0
O
CT>
ro
in
. •
CO
r-4
in
f**.
r««
0
CO
00
vo
o
VO
CM
Crt
0
0
CM
CM
VO
CM
O
in
ro
CM
in
r-i
VO
0^
CM
O
^j1
ro
^
o
vo
o
&
ro
o
in
r^
CO
t^
o
cr>
ro
o
in
GO
in
ro
CM
r-l
ro
rH
ro
in
00
0
VO
ro
in
II
£1
H
H
W
•4J
O
c
CO
0) CM
d o
•H rH
05 ^
0) C
e o
•H -H
-P -P
(d
0) W
m o
cu tn
tl *"H
•H
10 C
•H G
-H
M CD
87
-------
~dt~
(7)
where MI = total mass; go = mass inflow; g1 = mass outflow;
ki = 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, .de-fined as the ::nass flow in kg/sec
plotted vs time, becomes
M
gn =
(8)
The similarities of equation 8 and equation 6 are obvious, where
gn is mass flow (kg/sec), a equals (1 + kk1)/k, cind M is total
mass.
Hydrograph and mass flow simulations for Storm 10 on the
P-10 site are shown in Figure 27. 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 can be ovtained 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 (1 inch, 2.54 cm, or
greater) on Hunting Bayou of 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 USB or physiographic
factors in the watershed. In general, the time of peak of the unit
loadograph is related to n and k by the equation
t =
(9)
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 tp
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 lagr
ging and superposition of the unit graphs. A unit pollutograph
(concentration vs time) is found by dividing the ordinates of the
88
-------
Ul
o
a:
< •
o
CD CM
5 g
°
01
2
12
9.
3.
HYDROGRAPH
tt - 6
K = .13
o
(O
(O
2.4
1.8.
1.2.
0.6.
0.0.
20 28 36 44 52 60 68
HOURS INTO STORM
OBSERVED FITTED
Figure 27.
Fitted curves for storm runoff and pollutant
mass flows observed at P-10 on 4/8/75.
1 cfs = .028 m3/sec
89
-------
unit loadograph by corresponding hydrograph flows. Because of
the linear load-runoff relationships which have been developed
tne linear assumption of unit response is further justified '
load°9raph approach suffers the same limitations as
hYfr°g^a?h1 method with re^rd to assumptions of uniform
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.
Storm Water Management Model
SWMM is composed of five integral computation blocks as
shown in Figure 28. The Executive Block controls all activity
within the model because all input-output functions for other
blocks are programmed into the Executive Block. The Runoff
Block computes quantity and quality of runoff for a given storm
and _ stores results in the form of hydrographs and pollutographs
i*-J?«?t?1t° th| ?ain sewer system. The Transport Block sets up
initial flow and infiltration conditions and performs flow
quantity and quality routing to produce combined flow hydrographs
and pollutographs for the total drainage basin and at selected
^Jfm;diate PO1^3' Quantitv a^d quality of flow are stored and
SSS* ? Si6*?"1?3 criteria in the Storage Block. Diapers ioT
effects of the discharge in receiving waters are computed in the
JTXr-S* ^M^ B^°Ck- A m°re detailed description is available
in the User Manual - Volume in (46).
4.1, ~In ^eneral only one or two computational blocks, as well as
the Executive Block, are used in a run. However, all blocks may
be run together. The use of independent computation blocks
allows for examination of intermediate results. The necessity
for at least 35 OK bytes of core storage in SWMM leads to hiqh
run costs and limits the number of options to be analyzed.
The SWMM release of February 1975, referred to in this re-
port as the original SWMM version, was extensively modified The
capabilities of the modified version have been expanded to model
runoff and water quality from natural drainage areas. Study
Model Application —
Specific data required as input to the original SWMM are de-
scribed in Table 21. A quantified description of the watershed
provides a computational basis for the model and includes the
rainfall hyetograph for the storm to be modeled, a physical de-
scription of each subcatchment to be modeled including the drain-
age area, percent of impervious cover, ground slope, Manning's
roughness factors, estimated retention storage for both the per-
vious and impervious surfaces, and the coefficients to define
90
-------
g
o
ai
B
X
-Cp
u «
K O
o3
t-eo
._. r-^ §s
"« " VV^0"
vt c
1 s
lo£
(T2O
is
00
CNJ
0)
I
xo
So
91
-------
TABLE 21. MODELING REQUIREMENTS BY
Item 3.
Item 4.
Item 1. Define the Study Area
Land use, topography,
tract data, aerial photos, area boundaries.
Item 2. Define the System
tions of inlet structures.
Define System Specialties
Flow diversions, regulators, storage basins,
Define System Maintenance
Street sweeping (description and frequency). Catch-
basin cleaning. Trouble spots (flooding).
Item 5. Define the Receiving Waters
General description (estuary, river, or lak «>
sured data (flow, tides, topography, water quality).
Item 6. Define the Base Flow (DWF)
Measured directly or through sewerage facility °Pfr-
a?ing data Hourly variation and weekday vs. weekend.
^characteristics (composited BOD and TSSr results >.
Industrial flows (locations, average quantities,
quality) .
Item 7. Define the Storm Flow
nailv rainfall totals over an extended period (6 months
SrionqSr) encompassing the study events. Continuous
'
92
-------
Horton's soil infiltration equation. Input data also define
hydraulics for the storm sewer system in each subcatchment and
the main sewers or open channels in terms of gutter length,
slope, bottom width, and roughness coefficient, cross-sectional
area, side slopes, channel slope, and roughness factor. For
water quality modeling, a code defining the specific land use in
each subcatchment as well as the street-cleaning frequency, the
number of dry days prior to the storm event, the number of catch-
basins per unit area and the quality of their contents must also
be specified.
Horton's infiltration equation is used to calculate the in-
filtration rate of rainfall into the soil as a function of time
by Horton's relationship (49).
Manning's roughness coefficients were necessary for each
drainage element to describe the hydraulics of the drainage sys-
tem. Gutters and open channels were assigned an initial value
of 0.10, while a value of 0.03 was used for sewers. These
values were adjusted during model calibration to accommodate
higher peak flows. Combined sewers are not used in any of the
study areas and therefore all initial flows were zero.
A 10 minute time interval was the limit for modeling ac-
curacy at minimum cost, and all SWMM runs were made with regard
to this condition.
Hunting Bayou Modeling—
Input data for the Hunting Bayou drainage system were ob-
tained from existing engineering maps and site inspection. The
subcatchments and drainage network used as input data are shown
in Figure 29. The total drainage area of 1976.8 acres (499.86 ha)
is divided into 24 subcatchments ranging in area from 25 acres
(10.11 ha) to 138 acres (55.85 ha). Each subcatchment was
assigned a land use class for modeling water quality in SWMM.
The drainage system includes 23 gutters and pipes in the Runoff
Block and 44 manholes and conduits in the Transport Block.
Infiltration rates were originally estimated and then calibrated
by consecutive modeling runs.
Five storms were modeled initially. The rainfall data for
these storms were obtained from reports published by the U.S.
Geological Survey. All five storms occurred during 1968 and
1970, prior to the initiation of this project. Consequently,
only water quantity was modeled and no water quality data were
available. Comparisons of observed and computed hydrographs for
these five events, presented in'Table 22, indicate reasonable agree-
ment. The average absolute error in runoff volume was 26% (see
Table 22) of the observed value, while the average error in peak
flow prediction was 20% of the observed. The temporal agreement of
the hydrographs was very good. For instance, the times of peak
flow agreed within ten minutes in four of the five instances and
93
-------
LEGEND.
STUOV AREA BOUNDARY
SUBCATCHMENT DIVIDE
SEWER PIPE
OPEN DITCH
MANHOLE
MANHOLE NUMBER
1 ft = .305 m
Figure 29
•«*>*« *• ----
Subcatohments and drainage network in Hunting Bayou,
94
-------
average error was twenty-two minutes. However, computed values
tended to predict faster returns to low-flow conditions than were
actually observed.
TABLE 22. COMPARISON OP SWMM PREDICTED RESULTS
WITH OBSERVED FLOW MEASUREMENTS FOR
HUNTING BAYOU STORM EVENTS
Date of Storm
09/08/68
09/17/68
11/05/68
10/22/70
11/09/70
Total Runoff
(ft3 x 106)
Measured
4.84
9.06
4.40
12.69
1.52
Predicted
4.50
4.98
5.07
12.84
2.46
Peak Flow Rate
(cfs)
Measured
325
330
282
665
160
Predicted
303
302
337
549
205
NOTE: 1 cfs = .028 m3/sec
Three more recent Hunting Bayou storms on 3/26/74, 4/11/74
and 5/8/75 were sampled and the water quality and flow prediction
capability of the SWMM was tested. The 5/8/75 storm event will
be presented here as an example. Figure 30 shows the observed
and SWMM predicted hydrographs for this event.
Original SWMM predictive capabilities are based upon
dust and dirt accumulation data acquired in Chicago, and the
extrapolation of these data to natural drainage areas is a
limitation of the model, which resulted in poor water quality
predictions for this watershed. Consequently, a simplified
approach to water quality prediction in SWMM was developed
which does not consider pollutant buildup or input data on
dry days, street cleaning frequency, land use, or curb length.
Instead, pollutant availability loading rate at the beginning
of the storm is input. This information was produced by The
Woodlands project stormwater monitoring program. The user
determines effects of dry days, street cleaning frequency
and land use external to the model.
The new SWMM version was also run for the 5/8/75 storm event,
and the resulting predicted TSS pollutograph is shown in Figure 31
(left-hand figure). The loading rates used for this pollutograph
prediction are as follows:
95
-------
^_^
UJ
cc wj
in AT\
UJ ^4J
a
£0
!
!
«J
:
^
i
'
i
!
i
}
A
/
/
f j
i
./
r
r\i
i
i
i
j
i
i
i
i
i
^ i
N
.
'
i
\
\
\
N
^
i|
j
,
I
,
>
"S
b
^,
1
;
!
i
i
!
x 1
!
^j
1
•
!
!
•^. '
!
. 1
i
STREAM
FLOW
DISCHARGE
STATION
Ji-20
21 22 23 24 1 23456 7 8 9 10 M 12.
TIME (hrs.)
STORM OF 5/08/75
ui
o
O
o
OBSEUVIJ3
^OMPUTEl)
SUSPENDED
SOLilDS BY
ORIGINAL SWWM
VERSION
200 22 23 24 I
23 4567
TIME (hrs.)
8 9 D 1
STATION
M-20
Figure 30.
Predicted hydrographs for T55S concentratioinis
at Hunting Bayou ((5/8/75) ,
1 cfs = ,028 m3/sec
96
-------
[0 en i
£ CD -P
o -P M-I
H (d 0)
m M H
« £ is o
en
D§
g m fl
£
•d w Q
s w xi
(d fti CO
w
&
ft
in
vQ
LO
t-~
II
o\4J 's
-P co C J3
^"^ 0 ^r
H in O ;°
a^s-
0)
CD (0 CD
> « -P
Q) iJi ft
X! -H O
O 4-> O
0)
4J O
4-1 CD
CD -H
U CO (Q- (fl
•H EH CD
CD
CD
O CD
4-1 ^-1
fO
CD
M
3
tn
(l/fiui)
97
-------
Pollutograph
Prediction
TSS Loading Rate According to Land Use
m (Ib/acre)
Residential Construction Undeveloped
1-85 2.31 0.23
A removal coefficient of 21.2 min"
4.6 in the SWMM.
was used, as compared to
=T^T A_tr^l and error procedure was followed to determine the
above loading rates and removal coefficient combination that
would reproduce the observed pollutograph. Loading rate and
removal coefficients derived are valid only forttS s£rm Ssed
nd aPPlic*tion of results to othe? stSSL Is
Prevailing antecedent conditions and raXfall-
™ •s are similar for both storms and if study areas
are identical or at least homogeneous. ^uay areas
Although simulated pollutographs were curve-fit to
observed pollutographs, actual andcomputed masl ^ans^
*X2 not correspond. As shown in Figure 31 (right-hand fiou
SSFVi11^ P°°r reSUltS WSre outlined usingthe product o?'the
thJf Sn^^ograph -and pollutograph. it wa2 determined that
this condition resulted from a failure to properly predict the
timing of the peak of concentration. P-^UJ-C-C -cne
4-a«4- 5° improve the modeling of total mass loading of a pollu-
tant the new version was used to model pollutant mass flow rates
Again the loading rate and decay factor were adjusted to repro-
duce mass flow rates, using the following values:
TSS Loading Rate According
to Land Use (Ib/acre)
Mass Flow
Prediction
Residential
4.0
Construction
5.0
Undeveloped
0.5
A removal coefficient of 35 min-1 was used for this case.
Results, shown in Figure 31 (right-hand figure), indicate mass
flow rates can be accurately predicted using this method (dashed
line). The method was applied to other stormwater quality para-
meters, COD, N03, TP, with similar results.
Panther Branch Modeling—
The data input to the SWMM for the Panther Branch drainage
system was developed from existing engineering maps and numerous
site inspections of the watershed. The total drainage area for
Panther Branch of 21607 acres (874.40 ha) was divided into
57 subcatchments ranging from 21 acres (815 ha) to 1366 acres
(552.80 ha). The input parameter called "width ©f sub-
catchment" is defined as the width over which overland flow oc-
98
-------
curs. Values for this parameter were f-irst estimated by the
method described in the SWMM User's Manual (46). These values
were subsequently reduced by approximately 40% to achieve cali-
bration. The Panther Branch drainage system is made up of 57
"gutters" and 61 transport elements of varying characteristics.
A major drawback of SWMM at The Woodlands is that the area
below P-10 was in a transient state due to development. Con-
tinually changing land use affects the quality of runoff and
consequently P-10 is regarded as a control point. The area above
this gage is in a relatively stable condition and will give a
more accurate measurement of the pollutant loading due to the
land use rather than from a construction area. Stormwater runoff
from a construction area can vary in quality from storm to storm
depending on the stage of construction, and modeling proved diffi-
cult. Consequently, it is presumed that several construction
areas where the natural ground had been disturbed and stripped
of the protective vegetative cover contributed more TSS than SWMM
could predict from available input data.
Five storm events on Panther Branch have been modeled.
Similar to the data for Hunting Bayou, all subcatchment, gutter
and transport element data for Panther Branch were identical for
all runs. Infiltration rates were determined similar to
Hunting Bayou. The original SWMM was used to model both water
quantity and quality for two storm events on 10/28/74 and 12/5/74
and only quantity of flow for the remaining events (1/18/74,
4/8/75 and 11/24/75). Computed flow peaks and volumes agreed well
with the observed flows; the average absolute error in the volume
of runoff between observed and computed hydrographs was good
except for the storm events of 10/20/74 and 11/25/74 when the flow-
peaks between observed and computed hydrographs were approximately
three hours apart. Water quality modeling at P-30 was not
acceptable using the original SWMM. The 10/28/74 storm event at
P-30 was a multi-peaked hydrograph, and TSS modeling by the ori-
ginal SWMM was not accurate. The maximum observed TSS concentra-
tion of 1000 mg/1 during the second peak in streamflow was com-
puted to be 273 mg/1, a much lower concentration. Runoff quality
modeling of the 12/5/74 storm event was similarly too low in
concentration and total load.
The 12/5/74 storm event was modeled for both the upstream
gage, P-10, and the downstream gage, P-30. Since the entire
drainage area had the same land use before development began, most
differences between the upstream gage and the downstream gage can
be attributed to the changing land use in the developing area.
Using the original SWMM version, the computed peak concentration
of 142 mg/1 TSS at Station P-10 was in good temporal agreement with
the observed value of 130 mg/1. The falling limb of the observed
pollutograph occurred more rapidly than the simulation pollutograph,
resulting in'a difference of 11,330 kg (25,000 Ib) or a 40% error
99
-------
in computed total load. Observed TSS production at Station P-10
was about one-third that at Station P-30.
The modified water quality version of SWMM was also used to
model the 12/5/74 storm on Panther Branch for TSS, COD, N03 and
TP parameters. The pollutographs and mass flow curves were sepa-
rately computed using loading rates obtained by trial and error.
Simulation results are summarized in Table 23. Optimized polluto-
graph and mass flow curves corresponded well with the observed
data. However, mass flow rates calculated from the optimized pol-
lutographs were not as accurate as computed optimized mass flow
rates. TSS predictions at Station P-30 were compatible to observed
data with slight differences for occurrence of peak flow rates.
Modeling of TP at both P-10 and P-30 and NO3 at P-30 not entirely
satisfactory.
Swale 8 Modeling, Existing and Future Development—
Existing drainage and planning maps were used to develop the
input data for Swale 8, the watershed above Lake Harrison. Site
inspections to determine drainage area boundaries and extent of
construction were conducted on a periodic basis because this
watershed was in a transitional stage. During the project, the
channel was enlarged and construction of Lake C was underway
whiles Lakes A and B had already been filled.
The total drainage area for Swale 8 of 459.3 acres (185.87
hectares) was divided into 10 subcatchments ranging from 23 acres
(9.30 ha) to 66 acres (26.71 ha). Land uses, for the upstream
subcatchments were classified as open space, whereas the last
three downstream subcatchments were designated as multi-family,
residential and commercial. Seventeen drainage system elements
were used to model the entire area. Of these, two elements were
storage units, Lakes A and B, and all six channels were trape-
zoidal in shape as a result of channel enlargement.
Swale 8 storm even on 4/8/75 was modeled because the only
other observed storm event, 3/13/75, had a peak inflow into Lake
B of 0.06 m3/sec (2.0 cfs) from 0.81 in. (2.06 cm) of rainfall.
The transitional phase of development in Swale 8 gave rise
to several problems in modeling runoff. The most severe problem
is the lack of lake volume data. The topographic maps prior to
lake construction show the natural ground contours, but the
reservoir areas were used as borrow pits for fill material for
the dams as well as other construction at The Woodlands. Conse-
quently, the original storage capacity of the reservoirs was not
known and no subsequent reservoir surveys have been conducted;
therefore, the elevation-area-capacity data for these lakes were
100
-------
CO
&
w
ro
ro
o
H
X
•* Sf
CM CM
8
CM
CO
rt
0
K
y
W
vo
CO
«n <-4
OA
CN
«s
5
• • •
O\ i-l
CO r-l
«
ro
o
w
o
• •
g
CGCN
•a
en CM
° ' **
5
in
in
«s a
CO s^»
VO CM
o in
«d"
CO
CN CN
8
fa
8°
H
1-3
H
Q
fe
H
B
(4
H
EH
CM
H
J
1
« w
•y -««
ro
O
0) 03
w Q ^ w
CM r-l
en co
o
cr»
CM
NO
Is*
CM
0 0
vo
1 II
CO Q CM 'rj-
CO O O O
En CD 52; a,
V0
vo
i u
CO Q CM T
co o o o
t-l u ^ a,
O
(U
n
tn
in
G
•H
H
H
13
o
101
-------
only approximate. Also, groundwater was being pumped in Lake A
and pumpage rate was not recorded.
The outflow structure for Lake A is controlled by different
outlets at different water surface elevations. The outflow rat-
ing curve (discharge as a function of water surface elevation)
is composed of three segments, one controlled by the low-flow
orifice, the second controlled by weir flow through the flood
discharge outlet which in turn is limited at extreme flows by
the capacity of the outfall conduit and resulting in the third
segment of the rating curve. The SWMM is not capable of model-
ing this complex outflow scheme.
Under the conditions described above, the modeling of runoff
storage in the lakes proved to be difficult. Several attempts to
model the outflow from Lake A for the storm of 4/8/75 were un-
successful. The extent of assumed data was too large in magni-
tude to approximate the correct operation of Lakes A and B. As
a result, additional modeling of the watershed was conducted
only on that drainage area of Swale 8 upstream from Lake Harrison
at point £>-10 (Lake B gaging station) . The results of this
modeling effort are discussed in the following paragraphs.
Due to various external influences, urban development at The
Woodlands did not proceed as rapidly as expected. Site develop-
ment plans were available for Phase I, and in early 1976 a major
portion of the Swale 8 watershed was being planned for develop-
ment. Using these plans for the watershed, three development
scenarios were evaluated for modeling: (1) existing conditions,
(2) immediately developing conditions, and (3) future but not
ultimate conditions.
i
Water quality predictions by the modified version of SWMM
were attempted. Changes in land use and increase in impervious-
ness were computed from plat maps provided by The Woodlands De-
velopment Corporation and input to the SWMM. As described
earlier, the modified quality prediction version required the
input of loading rates for each pollutant. The initial loading
rates used were derived from the p-30 watershed modeling experi-
ence. Based on previously described experience with pollutograph
differences resulting from computed hydrographs, it was decided
that only mass, flow rates would be modeled. Runoff from the
storm on 4/8/75 was chosen for modeling; however, due to its
multi-peak complexity, only the first hydrograph peak was model-
ed. Results from the first computer run indicated that the load-
ing rates determined from the results of modeling at Station P-30
were too low. Observed peak mass flowof.TSS was three times the
peak mass flow computed from loading rates derived at Station
P-30. These differences are a result of the extreme effects of
lake and golf course construction, as well as channel improvement
concentrated in the Swale 8 watershed. Also the freshly sodded,
102
-------
EH
W
W
(0
H O 0 H
IT) en po CM
OCM
co co
OH
O CT» CTi (31
cr^ n
CM in
in co
vo
CO CM
VO CO
iH CM
CM
in CM VD H
i—1 VD iH CM
e
CO
J2;
8
i
en co
en ro
o H
• • t i
o in
00 CO
o
iH
CQ
o
£
H
8
§
ft
M
*""l!
p
• «
en oo
00 VD
CM
H
&
(0 CU
M CQ
CO P 4J O
CO O -H ,£)
EH U J3 ft
CQ
-------
developed areas severely eroded during the intense rainfall of
tnis period. Loading rates were revised for both construction
and developed areas. For Swale 8 the TSS loading rates from de-
veloped areas were 82% of the rate from construction areas. In
contrast, the same ratio at Station P-30 was 78%.
The modified version of SWMM was run for two future water-
shed development conditions described earlier, using the rainfall
on 4/8/75 to provide a basis for comparison between existing and
future conditions. AS anticipated, the larger proportion of area
under construction changes the pollutant loads: considerably; the
changes range from an increase of 77% for TSS to a decrease of 8%
for nitrates. '
After the construction phase of development has been com-
pleted, peak pollutant loads do not decrease as may be expected
but_the total mass of pollutant do decrease. These dramatic '
environmental effects of construction activities are listed in
"?abf^ 24C One reason for the increase in peak mass flow rates
is the change in the runoff hydrograph. After construction the
hydrograph peak is increased by approximately 40%. Another
reason is the increase in the input loading rates for developed
areas, which results in a doubling of peak mass flow rates for
the parameters N03, COD and TP. The 20% increase in the TSS
mass flow rate is a result of hydrograph modification due to
urbanization.
In summary, the modified water quality modeling version
greatly improved the capabilities of the SWMM. Water quality
modeling results are much more dependable, and observed events
can be adequately simulated. Each of the stores used to test
the new SWMM version was selected to present a range of flow
water quality and land use data; thus, the modal was tested over
a range of different conditions.
STORMWATER ALTERATIONS AT THE WOODLANDS
Water Quality Needs
Irrigation—
t 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 Th<> Woodlands and is
excellent to good for most plants" (50). The presence of nutri-
ents in stormwaters slated for irrigation is not high and in this
case considered an asset rather than a pollutant. TSS concentra-
tion or large particulates could cause mechanical problems such
104
-------
as pump damage or clogging of sprinkler heads, but careful
placement of the intake structure will avoid these difficul-
ties. TSS concentration in Lake Harrison during low-flow
conditions is about 100 mg/1 and average particulate size is
estimated between 25 and 250 mg/1, levels acceptable for
pumping requirements. The velocity in the distribution system
will keep the solids in suspension. Particulate size criteria
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 life-
forms. 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 controlled by a regular
lake maintenance program and algal blooms can be prevented by
reducing the lake detention times and nutrient levels (40).
Recreation—
In a discussion of recreational water uses, two divisions must
be considered: contact and noncontact. The water quality require-
ment for contact recreation, which involves substantial risk of
ingestion, is more stringent than that of noncontact (51). 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
arid 2000/100 ml maximum were suggested for official noncontact
waters (52). Fishing water criteria invoke an additional require-
ment that harvested species be fit for human consumption. Edible
fish species should be free of toxic chemicals and pathogenic bac-
teria 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 results.
Coliforms in the digestive systems of fish caught at Woodlands are
in higher concentrations than from other lakes but presumably do
not reach the edible portions of the fish (52).
Water Supply Uses—
Lake Harrison ranks as a poor raw drinking water source
because it would require a high level of treatment before use (50).
With groundwater, a less expensive and more reliable source is
easily obtained. Its use for this function is to be restricted to
emergencies.
105
-------
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 (..26 m3/sec>. Eor 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 15 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.
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
hour pause by less intense rainfall totaling 3.97 in. (10 cm).
The hyetograph is shown in Figure 32. 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,
656.2 ft (200 m) upstream of Lake B. Lake C was constructed to
serve as a wet weather pond and golf course water hazard.
Unfortunately, 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 33. Characteristic of runoff response in a small water-
shed, the multi-peaked inflow hydrograph was a product of the
sporadic hyetograph (Figure 32). Intense stormwater flow deepen-
ed the inflow channel by 6 in. (15 cm) and obliterated bales of
106
-------
001 08 09
SJO '39MVHOS10
. 8
8
(O
to
o:
§
O
§
o
o
s
o
ffl •
4J
AJ
J (1)
-------
hay placed in the channel to act as flow c°^rol deviceJ- Storm-
water flow crested shortly before noon on April 8 at a recora
discharge of 123 cfs (3.48 m3/s). Bank storage and ponding
helped to prolong minimum flow in the channel for two days, con-
tributing to the total inflow runoff volume of 93 acre ft (114 m^).
TnelakegsyStem effectively damped inflow fluctuations The
hydrograph peak traveled through the lakes in a half hour.
«._*>_ _. —,—.*..•»—»—»v* *i ^r*\n o i TnS3.ri GOIlC"J.l L.J- d. L. A.\jj.+*y » x^-*- m _
ticie comparison ua- luccai* ^ *->,« outflow was nutrient enriched as
a'rSultNol ont o^a^omSinStion of two sources:
(1) unmeasured runoff from the fertilized --._- ---_
1 to the lakes and/or direct precipitation on the
lakes,
The ouality of water held in the lakes prior to
SS storm event. (Water impounded in the lake
prior to the storm event approximated the runoff
volume.)
ai£fe^=
proximated maximum concentrations.
sedimentation is a
of 2660 mg/1 at
a
(117 t)
the voume of the
assume<1.
mass
215 S
ing Panther Branch by the lake system.
Table 26 shows the reduction in stormwaterseaiment^oad by
^lete1^^!, 100,, is a
108
-------
2800 r '40
CO
Q
8
o
a
800
<
o
400
/ \
TOTAL
SUSPENDED SOLIDS
DISCHARGE
2800 r
,800
o
f-
400
100
(O
- UJ
o
CO
Q
OUTFLOW
J 4 8 12
HOURS INTO STORM
Figure 33. Reduction of TSS through The
Woodlands lake system.
1 cfs = .028 m3/sec
16
20
109
-------
TABLE 25. SUMMARY OF WATER QUALITY PARAMETERS FOR SITES
LAKE A AMD LAKE B DURING THE APRIL 8, 1975
STORM EVENT
Drainage Area (acres)
Runoff Volume (ac-ft)
Rainfall (inches)
OUTFLOW
Lake A
483
93.4
3.97
INFLOW
Lake B *
337
93.2
3.97
Concentration of Water
Quality Parameters:*
Ortho-P
TP
NH3
N02
NO3
TKN
TSS
SOC
Total COD
Soluble COD
Specific Conduc-
tance (micromhos)
Turbidity (JTU)
Avg.
0.015
0.10
0.16
0.032
0.28
1.3
245.
13.6
41.8
26.4
130.
Max.
0.048
0.19
0.26
0.046
0.32
2.
356.
19.
45
31.
215.
Avg.
0.005
0.11
0.11
0.009
0.15
1.86
1273.
16.2
63.7
32.
85.
Max.
0.013
0.36
0.15
0.054
2.1
3.1
2660.
22.
87.
45.
304.
160.
210,
375.
900,
1 ac = .405 ha
1 ac ft = .123 ha-m
1 in = 2.54 cm
*all concentrations in mg/1 except where indicated
110
-------
result of total stormwater storage by Lake Harrison and does not
preclude discharge at a later time.
TABLE 26. STORMWATER SEDIMENT REMOVAL AT LAKE HARRISON
TSS Load During Storm Event
Storm #
8
9
10
13
14
15 & 16
Ibs Input
(Lake B)
104
13800
322000
6700
115302
4840
Ib Discharged
(Lake A)
Flow stored
within lake
991
61900
Flow stored
1850
3270
% Load
Reduction
100%
93%
80%
100%
84%
32%
1 Ib = .453 6 kg
2
Estimated value (Lake B gage inoperative) calculated
using estimated 10.1 ac-ft (1.24 ha-m)inflow times
sample average concentration, 421 mg/1.
Disinfection—
The FC standard adopted by the State of Texas for contact
recreation is 200/100 ml (Iog10 = 2.3) (53). For noncontact
recreation the figures are one order of magnitude higher, i.e.,
2,000/100 ml (Iog10 = 3.3). Sixteen of 27 storms, considering
all stations, exceeded the noncontact recreation standard. All
the Westbury storm events exceeded the standard. Contact
recreation standards were met only for Lake A stormwater sampled
on 3/4/75 and 3/12/75.
Dry weather flow in Panther Branch met the contact recre-
ational standard at site P-10 but not at site p-30. The mean
FC value for p-10 low-flow data was 2.13 (log1Q basis) indicat-
ing that, on the average, the criteria for contact recreation is
satisfied. However, the value for stream water at P-30 was 2.38
and therefore unacceptable.
Ill
-------
Disinfection of stormwater is feasible, and it is generally
agreed that large dosages will be required to achieve adequate
reduction in indicator and pathogen densities. Chlorine or
chlorine dioxide has been reported to be the most effective dis-
infectant, and in many cases the least expensive (54, 55, 56).
Davis et al. (57, 58) discussed current disinfection research
and practices along with encountered problems which occur in
combined sewer disinfection and stormwater disinfection.
Samples were obtained from different locations during storm
events to determine disinfectant demand and effectiveness of
ozone and chlorine. Chlorine demand of Lake A stormwaters on
3/13/75 and 4/7/75 was 10 mg/1, and the ozone demand was in ex-
cess of 32 mg/1. These elevated demands were partially due to
suspended solids and oxidizable materials competing for the dis-
infectant. As a result, stormwater disinfection will be costly.
Chlorine and ozone toxicity—Water quality standards often
state that the final concentration of any waste in a receiving
water should not exceed 1/10 of the 96 hour LC5;) value (59). To
assume that 1/10 of even a true 96 hour LC^Q would not have
severe physiological effects and truly impair natural propagation
of all species is unrealistic. Tsai (60) reported that species
shifts occurred following the introduction of chlorine into a
Maryland river. Also, Arthur and Eaton (61) show that the re-
production of fathead minnows is drastically affected by exposure
to sublethal concentrations of chlorine. This evidence suggests
that sublethal concentration of chlorine is capable of producing
significant physiological impairment.
The approach taken herein was to establish a chlorine stand-
ard based on physiological responses of the test animal. The ef-
fect of chlorine and ozone exposure on the phys Lological function
of Ictalurus punctatus (the channel catfish) has been examined.
The physiological parameters evaluated were heart rate, blood
pressure, sodium uptake, ion excretion, and glomerular filtra-
tion. In addition, typical LC5Q bioassay tests were completed
for comparative purposes. Significant results are presented in
the following paragraphs.
Survival-mortality characteristics—The survival-mortality
characteristics of fingerling channel catfish to chlorine were,
examined and results are shown in Figure 34. A flow through
bioassay compared to a static bioassay results in a lower 96 hour
LC5Q. The LCcQ concentrations of 0.07 mg/1 for the flow through
bioassay and 0.45 mg/1 for the static bioassay were read directly
from the figure. A flow through or continuous ::low bioassay is
a more accurate measurement for 96 hour LC5Q of a highly reacting
toxicant like chlorine (62) . The chlorine concentrations in
Figure 34 are based on the measured amounts of total residual
chlorine added to the inflow and do not represent concentrations
during tests with fish. Thus the chlorine concentrations shown
112
-------
>
<
CO
(/»
<
o
CO
O
1MB*
A
O>
6
cn
•H C
C,
0)
>.
4- TJ
cn
T5
cn
cn
*!/ TO
O C. O
i"0 3 •*•"*
£j fft jQ
.
(U £4
C .C
GO
6
r-
•
o
03
6
10
0
*r
6
ro
6
CJ
Q
—
•v»
o>
E
UJ
z
oc
o
1
X
o
TJ Sj
p 0 3
ti «-H O
o x: --<
e o u-,
f-H O fl
(tf P
> c
> cn
3 ^-4 C
cn P fQ
OH O >
O TJ
oo cn
c c cn
O *r"l ^0
cn «-H o
• H C,-^l
t. a) jQ
(TJ CO
CU C 0
g -H -H
C •*-• P
O
-------
l v £ b? rePrese*tative at treatment outfalls
Samples taken from receiving waters should show chlorine concen-
X< s^lflcantly lower. Based on a 96 hour ie*ZfToT
and USing the "Aquatic Life Water Qualfgy criteria"
^^
noted that this concentration is below the liSit of Analysis?
The flow through bioassay was also used for ozone and the
survival-mortality characteristics of fingerling catfish are
, t.
reactive, it was decomposed in the presence of organics (fish).
In the presence of fish, ozone could not be detected in the
insensitivity of the analytical technique
^entrations were estimated from
by its odor (detectable odor
exceeds
chlorine^into the 10 liter recirculating system. Upon
??fe WaS a dr°P ±n bl°°d P^ssure, duS to ga
Following escapement from vagal inhibition, there
m;an.in=rease m blood pressure which in turn decreased
mmutes of exposure. This is thought to be due to an
e X vascular resistance, changes in heart rate
r be secondary compensations for the gill vascular
resistance chlorine ^xposure at levels approaching the 96 hour
LC50 (0.7 mg/1) were immediately detected as shown by a pro-
1111 ° in.blood Pressure. Continued exposure for 5 hours
increase in heart rate fr°ni 22 beats/minute to
At an exposure of 0.03 mg/1 chlorine there was
,.react:LOn shown b^ the bl°od pressure or heart rate
After two hours the fish appeared normal with no apparent SiJl
function. Thus, it is likely that with exposure to a chlorine
concentration of 0.007 mg/1 chlorine (1/10 of 96 hour LC I
fish will shown no .physiological response. 50
Tests to determine the effects of ozone on blood pressure
pulse pressure and heart rate indicate that ozone levels from
0.1 to 0.5 mg/1 for 10 hours had no apparent effect.
4-T, GJ1,1 S0d:i?m transPort~The influence of chlorine exposure on
the uptake of 22Na by gills of fish is shown in Figure 36. The
disappearance of 22Na from agueous phase (ordinateyaxis in Figure
114
-------
GO
C
P
t,
V JZ
GO GO
C 3
•H O
.C
M-i P
O
tn o
O rH
•H 4-1
P
tn 03
•H
p
o a>
f- 3
rfl 03
J2 O
u o.
X
>, QJ
P
•H 0)
fH C
m o
t. o
o
e o
i P
03 .C
>
03
tn tn
tn
03
3 03 -H
co o Xi
ro
0
fj
3
GC
•H
115
-------
36) is a measure of the removal of 22Na by the gills of the fish.
The addition of 0.1 mg/1 of chlorine at the start of the experi-
ment markedly reduced the ability of the gills to remove sodium
from aqueous solution. This reduction in uptake rate following
chlorine exposure suggests an impairment of normal physiological
function in the sodium transport system.
o p
The influence of ozone exposure on the uptake of Na by the
gills of fish is shown in Figure 37. Phase one of the experiment
utilized the fish as its own control. A one day recovery period
followed. The addition of 0.1 mg/1 ozone in the second phase of
the experiment markedly reduced the ability of the gills to re-
move sodium from the aqueous solution. This reduction in uptake
following ozone exposure also suggests an impairment of normal
physiological functions.
Eutrophicat ion—
The development of standing crops of phytoplankton in pan-
ther Branch is influenced to a large extent by streamflow rates.
High flow rates do not provide adequate detention times for de-
velopment of large standing crops at any given point along the
stream. However, reductions of flow rates and/or pooling in the
stream allow detention times suitable for standing crop develop-
ment, provided nutrients are not limiting for algal growth. De-
velopment of algal populations in this stream, as in other aquat-
ic ecosystems, is influenced by concentrations and availability
of various algal nutrients. This point is not only pertinent to
development of phytoplankton in Panther Branch but also to
aquatic systems which might receive water from this stream.
Thus, it is imperative to have some knowledge of which nutrients
stimulate algal growth in Panther Branch water. Therefore, algal
bioassays were conducted to determine whether nitrogen and/or
phosphorus were limiting for algal growth.
Low-flow water samples collected from sites on Panther
Branch and Spring Creek were used to determine the limiting-
nutrient for algal growth. Aliquots of stream water were inocu-
lated with the algae Selanastrum and spiked with nitrogen and/or
phosphorus as nutrients. The results presented in Figure 38 in-
dicate algal growth was increased by additions of both nitrogen
and phosphorus to water samples and phosphorus was the most im-
portant single limiting nutrient along Panther Branch. The in-
troduction of treated sewage effluent and agricultural runoff
into Spring Creek was probably responsible for the comparatively
larger algal yields in water from this stream. These findings
are in contrast to results derived from bioassays of stormwater.
Algal bioassays were also conducted with water collected
from Panther Branch at various time intervals during the course
of storm events. The stormwater runoff collected below the major
area of construction in The Woodlands (site P-^-30) seemed to
fluctuate in its ability to support the growth of algae (Figure
116
-------
CVJ
O
csi
CO
cr
•=>
o
X
t\J
III
S
00
Q
O
CJ
O
O
00
O
O
.a
(d
CN
M-i
O
^1 0)
«3 fi
-P -H
&H
D O
H
-------
P9
CD
W
co
to
cvj
00
cc
_ -3
O
•31
-
UJ
•v
«eE
__
CO h-
ro
p
3
CO
CD 3
P it!
P
C O
O C
3
CD O.
3 CO
CO 3
O SU
a 3
X >H
CD rd
CD O
C M
O
O O
e co
P< i — i
0) ^
P -H •
bO'—
P 0)
^ CD C
O ,C O
x; p N
CO C
UH il i— 1
O ^.
rQ bC
-P 2 E
OCM
CDOM r-i
4-4 U-t O
cj o >-'
1O
a
O
O
O
o
o
00
o
o
CO
0
o
^*
0
o
OJ
v
o
O
0)
bO
•H
3inNIW /SINOOD
118
-------
UJ
o
o
t
o
I-WATER SAMPLES - SPIKES
2-WATER SAMPLES * 1.0 ppm N
3-WATER SAMPLES * O.O5 ppm P
4-WATER SAMPLES * 1.0 N * 0.05 P
P-IO
P-20
P-3O
P-4O S-09
S-IO
COLLECTION SITE
Figure 38,
Optical densities of Selenastrum capricornutum
after incubation in v/ater from panther Branch.
and Spring Creek (May, 1974).
119
-------
39). Test samples collected early in runoff at P-30 demonstrated
the greatest growth in both control and spiked aliquots. These
samples corresponded to high nutrient stormwaters from the fer-
tilized areas of The Woodlands, not present at site p~10. As
runoff at P-30 progressed there was a decrease in algal cell
yields to similar levels observed for P-10 test samples. Algal
growth was increased by enrichment of stormwater samples with
both nitrogen and phosphorus; however, nitrogen was consistently
limiting for algal growth in stormwater runoff from The Woodlands.
Data from storm events at Hunting Bayou were similar to
those described above, indicating a stimulation of algal growth
with additions of nitrogen or both nitrogen and phosphorus to the
majority of water samples and slight or no stimulation with only
phosphorus spikes (Figure 40). The marked reduction in algal
growth, even with combined nitrogen and phosphorus, in the 4:30
sample could have been due to the presence of some toxic agent or
the absence of some essential trace element. Additional data ob-
tained from bioassays of stromwater from Westbury Square (Figure
41) also indicated that nitrogen was the limiting nutrient for
algal growth in stormwater runoff.
Porous Pavement
Rainwater Storage and Quality—
The water depth under the porous pavement in both the sand
subbase and the general storage layer during a period of rainfall
from 3/6/76 to 3/8/76 is shown graphically in Figure 42. The
depth of the sand layer is 33 in- (84 cm),, and the layer shows
saturation resulting from failure of the French drain to remove
water from the lower levels. The gravel layer responded to rain-
fall by storing water as shown by the increased water depth in
the gravel layer followed by a gradual decrease in depth as the
water was drained from the top of the sand layer. The height of
the stored water and the time to reach the peak height depend on
the quantity and intensity of the rainfall. This is best illus-
trated in Figure 42 in which seven separate rainfall events oc-
curred over a 60 hour time span. The first three events, total-
ing 1116 ft3 (31.6 m3), were spread over approximately five hours
and produced an even response with a peak of approximately six in.
The subsequent event at 21 hours involved only 283 ft3 (8m3) of
rainfall and increase in stored water height of 1.8 in. (4.6 cm).
For the two major storm events occurring at 28 and 48 hours, the
water level had not drained sufficiently and the increase in
stored water height exceeded the measuring probe length. The
time required for the majority of stored water to drain away was
approximately 10 hours.
The quality of the water in runoff from the standard pave-
ment and in the gravel and sand layers beneath the porous pave-
ment was monitored during six storm events. As expected, there
was a general flushing effect for the runoff water with initially
120
-------
*r~
1
*i
101
(Ml
* 1
TO]
OJ|
-1
*r
rol
—1
*c_
toT
M|
*E
Sevif^
5 o HT!
ff 0. 4. L-
— c> — toT"
^C
o o o p 9.
(Siuiooi/6uu) 1H9QM AHQ
— ^
2
• 1
UJ
J k
-------
CO
UJ
it:
CL
CO
i
_l
CL
^
CO
ce
ft
*<
•$•
i
CL
o-c
2 P C
g g- 4
a. °-
o. 10 2
/-\ O <~
2 § -
+ +• 4
_J -l -
Q. Q. a
2 2
< ^ <
CO. CO V
on on a,
UJ Ul U
• * i_
§ si
^ ^ ^
OJ K> ^
CMC
1
^•c
CVlf
^*
^
1
«•[
[Ml
*l
(Ml
Tl
CVJI
-
t
cad
•1
rol
-I
1
rol
— L
rol
— 1
i
— L
rol
i
1
rol
ro
__
•
•
:
srl
ro
•«•«•
CM
^^MNH
}
n
Q
CO
o
o
U>
S:
vC
O
CM
O
—
00
o
9
m
CM
en
LU
Q
a
8
UJ
CL
cn
^ VD
0) OJ
-P
rtJ .d
SC)
P* U
»-i ^1
M (d
O S
-P -—•
to
S O
fd
PQ
ti
-P
03
fd
C
Q)
Q)
CO
£ O
O G
M H
-t-i
CO
CO
1H9I3M
122
-------
O
O
C£
UJ
Q
7.0
6.0
5.0
4.0
3.0
2.0
1 - CONTROL
2- 1.0 N
.1. 3- 0.05 P
2
1
3
4
I
-
3
4- 1.0 N* 0.05 P
1
2
3
4
1
2
4
3
•••
2115-2215 , 2300-0045, OII5-O215, 0245-0315
WESTBURY SQUARE
TIME OF COLLECTION (hrs)
Figure 41. Growth of Selenastrum in stornwater
runoff from Westbury Square (May 8,
1975).
123
-------
UJUJUJ< •
>CE>Q.
<£LOtO
to
•»-
o
o
CO
IO
u>
CM
to
**
H-
N-
IO
•*-•
H-
to
CO
CM
to
•*"•
**-
JO
« w™
tLJ?
O m. O
— o
HH/S3HONI
UJO
cc
LU
<
CO
0.
o
o
(O
in
to
in
CO
a:
m rs
to o
o
to
in
CM
in
CO
-------
high concentrations of contaminants which then rapidly decreased.
The rapidity if change is not evident in the water underlying the
porous lot. Table 28 presents a summary of stormwater quality
data collected during a 0.44 in. (1.12 cm) rainfall on 2/20/76,
which is exemplary of the monitroing effort. SOC and soluble COD
are generally lower in the water under the porous lot than in the
runoff from the standard pavement, but the reverse is true for
conductivity, TKN, and NH3. This is apparently a result of the
failure of the French drain to completely drain the sand layer
beneath the porous lot. The lower COD and TOC values, together
with the high Kjeldahl and ammonia nitrogen, indicate than an
anaerobic digestion process is occurring in the standing water.
Lead values are generally lower in the water under the porous lot
and, in most cases, would be acceptable in aquifers. The few
cases where high lead values were found in the underlying layers
do indicate that before porous paving is used for recharge of
aquifers by percolation information must be obtained on the ability
of soils to remove lead from leachates.
Wet Skid Resistance—
A modification of the locked wheel test was used to deter-
mine the coefficient of sliding friction between a standard
passenger vehicle tire (size 6.50 x 13) and the test pavements,
both wet and dry. In the case of the porous pavement, tests were
made on a section of the original paving and on a repaired section.
Test results for three series are shown in Table 27,
TABLE 27. COEFFICIENT OF FRICTION
Test Date Paving and Condition
Reaction Force
(Ibs)
Coefficient
of Sliding
Friction (u,)
12-3-75
2-5-76
4-28-76
Old porous - dry
Old porous - wet
New porous - dry
New porous - wet
Standard - dry
Standard - wet
Old porous - dry
Old porous - wet
New porous - dry
New porous - wet
Standard - dry
Standard - wet
Old porous - dry
Old porous - wet
New porous - dry
New porous - wet
Standard - dry
Standard - wet
43.2
60.8
47.2
52.8
52.0
42.4
58.4
68.0
59.2
64.0
71.2
57.6
52.5
60.0
57.0
60.0
55.5
51.0
.605
.851
.661
.739
.728
.594
.818
.952
.829
.896
.997
.806
.735
.840
.798
.840
.777
.714
1 Ib . = .138 nt
125
-------
TABLE 28. SUMMARY OF STORMWATER QUALITY FOR POROUS
PAVEMENT STORM ON 2/20/76
Conventional
1
Constituent
pH
Sp . Cond .
Soluble COD
SOC
TKN
NH3
NO3
N02
TP
ortho P
Pb
Zn
Gravel
X
8.1
457
35.
15.
2.5
1.5
.12
.014
.10
.06
.05
.18
Porous
Layer
s
0.2
14.
5.8
4.4
.42
.41
.14
.017
.02
.02
.03
.39
Pavement
Sand
3c
8.0
542
35.
13.
2.7
1.33
.03
.004
.50
.48
.03
.27
Layer
s
.04
12.
3.7
2.8
.46
.28
.01
.002
.08
.06
.01
.18
Pavement Runoff
x
7.8
108
60.
30.
1.2
.15
.36
.013
.11
.06
.31
.34
s
.23
54.
35.
16.
.42
.09
.17
.009
.04
.02
.43
.21
n = 11
n = 11
n = 10
all measurements in mg/1 except
specific conductance in
2 _
in pH units and
= mean
s = standard deviation
n = number of samples
126
-------
In most cases the standard, dense paving exhibited dry skid
resistance better than or equal to the two porous pavements. The
reverse was true when wet paving was tested with the porous pave-
ments being clearly superior. For standard pavement, the coef-
ficient of friction under wet conditions was, as expected, lower
than under dry conditions. However, a consistent behavior pat-
tern for the porous pavements, for which there is currently no
verified explanation, is the increase of the coefficient of
friction on wet pavement over that on the dry pavement. One
possible explanation is that in the case of the wetted pavement
surface dust layers have been washed through the pavement and
the rougher surface .then comes into full play.
Noise Levels
Traffic noise levels (Table 29) were determined using two
different vehicles, both equipped with standard steel belted
radial tires. Measurements were made early in the morning to
avoid external noise interference.
TABLE 29. NOISE LEVELS 1 mi = 1.6 km
paving
Vehicle
Speed (mph)
Noise Level!(dB)
Old porous
New porous
Standard
Old porous
New porous
Standard
Old porous
New porous
Standard
1
1
1
2
2
2
2
2
2
15
15
15
15
15
15
20
20
20
A62
A57
A64
B68
B68
B71
B69
B69
B73
The porous pavings are less noisy than standard paving.
1 A scale consistent with human ear. B scale filters to
accentuate lower frequency sound.
Porosity
In situ porosity was determined on the porous pavements_by
grouting in place a 6 in. diameter tube having two scribed lines
5 in. apart. Water was placed in the tube and the time required
for the water level to pass the marks was measured. Results are
calculated in in./sec of water transmission. The section of new
porous pavement gave the best water transmission than that in the
original pavement. Water transmission rates in the new pavement
were normally 0.55 in./sec (1.4 cm/sec) while the original sec-
tions exhibited a normal porosity of 0.38 in./sec (0.96 cm/sec).
Sections of original paving were found to be clogged. These sec-
127
-------
tions were primarily near the curb or near the curbing at the
upper side of the lot opposite the curb. The clogging near the
curb resulted from mud and dirt from the wheels; of contractors'
trucks, while the clogging of the upper lot was; chiefly from
cement dust from curb construction and find sand and silt washed
from the grass plot before the grass gave adequate cover. The
partially clogged or completely clogged areas were marked for
studies of cleanability.
Cleaning and Maintenance
It was recognized early in the program that maintenance and
cleaning of porous pavement to prevent or alleviate clogging
would be a factor in the application of such peivements. Sections
of the porous pavement which were clogged were cleaned by various
methods. No method was satisfactory on fully clogged pavement
and only a superficially clogged section, showing a water penetra-
tion of 0.38 in/sec (0.96 cm/sec), could be restored to normal
operation. The best method for cleaning was brush and vacuum
sweeping followed by high pressure water washing of the pavement.
Vacuum cleaning alone, once the pavement is clogged, was found to
have no effect. The oils in the asphalt bind c'.irt and only
abrading and washing techniques are effective in removal. By
removing fractional thicknesses of paving, it was observed
the clogging to a depth of 0.5 in (1.3 cm) was sufficient to
prevent water penetration.
Damage to pavement porosity results chiefly from abuse dur-
ing the early life of the paving. Normally, pa.ving is carried
out while heavy construction and earth moving is continuing in
the area and is subjected to mud and dirt from contractor ve-
hicles for up to several months. Continual passage of these
vehicles serves to compact dirt into the pores. Porosity can be
retained only if the paving is cleaned daily by sweeping and high
pressure water washing.
Once a large area of porous pavement is fully clogged it
cannot be adequately cleaned and the paving must be removed to a
depth where the clogging is not evident and new porous paving
filled in. In extreme cases, the affected area of the porous
topping must be removed and new topping put down.
Both the standard and porous pavements showed signs of the
effect of heavy vehicle passage and of power steering damage.
Heavy vehicles tended to leave tire depressions in either lot on
hot days while power steering damage was evidenced by small cir-
cular depressions in the lot surface. The latter damage occurs
when wheels are turned when a vehicle is not in motion. The use
of a lower penetration asphalt for pavements should offset the
damage to paving in warm climates.
128
-------
Pavement Deflection
Pavement deflections in porous and standard surface lots
were measured on four different occasions. Measurements were
made of pavement acceleration and time duration of acceleration
as it deflected on passage of the test vehicle. Test results
indicate there are no real differences in the magnitude of the
deflections, but the time duration of the deflection was always
longer for the porous pavement than for the standard pavement
with correspondingly lower "g" levels. This long response may be
responsible for the lessened tire noise observed in the sound
level tests.
Pavement Survey
Visits were made to four locations where porous paving is in
use. The locations at the University of Delaware, Newark, Dela-
ware, Bryn Mawr Hospital, Bryn Mawr, Pennsylvania, and at The
Marine Sciences Consortium, Lewes, Delaware were in good con-
dition with little or no indication of pavement clogging. The
paving at a Travelodge in Tampa, Florida showed considerable
clogging from sand and silt caused by passage of trucks on the
nearby road and inadequate curbing to prevent surrounding soil
from washing onto the lot. Degradation of the surface was evi-
dent in the form of patches of loose stone and gravel.
It is evident from the damage at the Tampa lot and in the
lot at The Woodlands that a lower penetration asphalt should be
used in the topping, especially in warm areas. Again, it is felt
that clogging can be minimized by proper use of curbing to pre-
vent surrounding soil from washing onto the lot surface. Recent
installation of porous paving at The Franklin Institute parking
lot in Philadelphia, Pennsylvania has further pointed up the need
for close control of contractor vehicles on a newly installed
lot. It was necessary to sweep away caked mud from vehicle
wheels and then wash the affected area with high pressure water.
Traffic Paint Visibility
Two colors of traffic paint, white and yellow, with and
without addition of glass beads, were applied to the pavements.
They were photographed under equal incident lighting intensities
at night'under wet and dry conditions and the reflected light in-
tensity was determined from image densities on the negatives.
The white marking paint showed a slight superiority in visibility
over the yellow whether or not glass beads were added. No real
difference in the reflected light intensity from either pavement,
wet or dry, was evident. The test, however, did not evaluate
glare from oncoming vehicle lights which may obscure reflected
light from paint on a wet, non-porous pavement.
129
-------
1.
2.
3.
REFERENCES
Field, Richard, and Pauline Weigel. Urban Runoff and Com-
bined Sewer Overflow. JWPCF, 6,: 1108, 1973.
Wolman, M. G., and A. P. Schick. Effects of Construction on
Fluvial Sediment, Urban and Suburban Areas of Maryland.
Water Resources Research, _3_: 451, 1967.
Guy, H. P., and G. E. Ferguson. Sediment in Small Reser-
voirs Due to Urbanization. j. Hydraulic Div., ASCE, 88,
HY2; 27, 1962.
4. Anderson, D. W., and S. D. Faust. Changes in Quality of
Water in the passaic River at Little Falls, New Jersey.
USGS 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, 40; 1861, 1968.
6. Storm and Combined Sewer Overflows Section, Environmental
Impact of Highway Deicing, U.S. EPA Report # 11040GKK06/71.,.
1971
7. Bryan, E. H. Quality of Stormwater Drainage 'from Urban
Land. Paper presented at 7th American Water Resources Con-
ference, Washington, D. C., 1971.
8. Heaney, J. p., and R. H. Sullivan. Source Control of Urban
Water Pollution. JWPCF, 43: 571, 1971.
9. Sartor, J.D. and G.B. Boyd. Water Pollution Aspects of
Street Surface Contaminants. U.S. EPA Report # EPA-R2-72-081,
1972.
10. Claudon, D. G., D. I. Thompson, E. H. Christenson,. G. W.
Lawton, and E. C. Dick. Prolonged Salmonella Contamination
of a Recreational Lake by Runoff Waters, Appl. Microbiol.,
21. (5): 875-877, May, 1971.
11. Tafuri, Anthony N. Pollution from Urban Land Runoff. News
of Environmental Research in Cincinnati, U.S. Environmental
130
-------
Protection Agency, Cincinnati, Ohio, April 11, 1975.
12. Harms, L. L., P. Middaugh, J. N. Dornbush, and.j. R. Ander-
sen. Bacteriological Quality of Surface Runoff from Agri-
culture Land - Part II. Water and Sew. Works, 122 (11)-
71-73, November, 1975. ~r~~~
13. Harms, L. L., P. Middaugh, J. N. Dornbush, and J. R. Ander-
sen. Agricultural Runoff Pollutes Surface Waters - Part I
Water and Sew. Works, 122 (10): 84-85, October, 1975.
14. Edmondson, W. T. Eutrophication in North America, in:
Eutrophication: Causes, Consequences, Correctives, Anon
(ed.). Proceedings of NAS-NRC Symposium, Washington, D. C.,
1969 -
15. Sylvester, R. O. An Engineering and Ecological Study for
the Rehabilitation of Green Lake. University of Washington
Seattle, Washington, I960. ' '
16. Hasler, A. W. Eutrophication of Lakes by Domestic Drainacre
Ecology, 28; 383, 1947.
17. Biggar, j. w,, and R. B. Corey. Agricultural Drainage and
Eutrophication. In: Eutrophication: Causes, Consequences
Correctives, Anon (ed.). proceedings of NAS-NRC Symposium,
Washington, p. c., 1969. *
18. Cooper, C. P. Nutrient Output from Managed Forests. In:
Eutrophication Causes, Consequences, Correctives, Anon (ed )
proceedings of NAS-NRC Symposium, Washington, p. C., 1969.
19. Jackson, D. P., and H. F. A. Meier. Variations in Summer
Phytoplankton Populations of Skaneateles Lakes, New York
Assoc. Theor. Appl. Liranol., proa., 16: 173, 1966.
20. palmer, M. C. Algae in Water Supplies of the United States
In: Algae and Man, Jackson, D, F. (ed.). Plenum press.
New York, 1964.
21. Schwimmer, D., and M. Schwimmer. Algae and Medicine, in-
Algae and Man, Jackson, D. F, (ed.), Plenum Press, New York,
22. Rose, E. T. Toxic Algae in Iowa Lakes. Iowa Acad. Sci
Proc., 60.: 738, 1953. '
23. Olson, T. A. Water Poisoning - A Study of Poisonous Algae
Blooms in Minnesota. Am. J. Health, 50: 883, ij.960.
24. Moore, G.. T., and K. F. Kellerman. A Method of Destroying
or Preventing the Growth of Algae and Certain Pathogenic
131
-------
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
Bacteria in Water Supplies. US.DA Bureau Plant Indus-
tries Bulletin 64, 1904.
Anon (ed.). Eutrophication: Causes, Consequences, Cor-
rectives. Proceedings of NAS-NRC Symposium, Washington,
D. C., 1969.
Kerr, P. C., et al_. The Interrelation of Carbon and Phos-
phorous in Regulating Heterotrophic and Autotrophic Popula-
tions in Aquatic Ecosystems. Proceedings of 25th1Purdue
Industrial Waste Conference, West Lafayette, Indiana, 1970.
Keuntzel, L. E. Bacteria, Carbon Dioxide, and Algal Blooms.
JWPCF, 41: 1737, 1969.
Lee, F. An Approach to the Assessment of the Role of Phos-
phorous in Eutrophication. Paper presented at American
Chemical Society, Los Angeles, California, 1970.
Shapiro, j.
1970.
A Statement on Phosphorous. JWPCF, 42 ; 772,
American Chemical Society.
Chemical Basis for Action.
D. C., 1969.
Cleaning Our Environment: The
Amer. Chem. Soc., Washington,
Poche, R. M. A Baseline Census and Development of a Monitor-
ing System for Important Animal Species on the Proposed
Woodlands Site,'Montgomery County, Texas. Consultant's Re-
port to The Woodlands Development Corp., April, 1973.
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., Mary, 1973.
Hass, 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 Deve-
lopment Corp., July, 1972.
Kendrick, W. W., and D. Williams. Soil Survey of The Wood-
lands. Consultant's Report to The Woodlands Development
Corp., July, 1973.
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.
Winslow, D. E., J. A. Veltman, and W. H. Espey, Jr. Natural
Drainage Systems. Paper presented at Spring Session of the
Texas Section of the ASCE, Beaumont, Texas, March 29, 1974.
132
-------
37. Dugan, G. L., and P. H. McGauhey. Protecting Our Lakes:
Wastewater Treatment is Not Enough. Paper presented at
46th Annual Conference of W.p.c.F., Cleveland, Ohio,
September 30-October 5, 1973.
38. Schicht, R. j., and p. 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, 197$.
39. 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*
40. Ward, C.H., and J. King. "Eutrophication Potential of Surface
Waters in a Developing Community, " Draft Final Report fpr
Maximum Utilization of Water Resources in a Planned Community,
EPA Research Grant #802433, October, 1976.
41.
42.
43.
44.
45.
46.
47.
MacEntee, F. J. A Preliminary Investigation of the Soil
Algae of Northeastern Pennsylvania. Soil Sci., 110s 313-
317, 1970. - ".
Hofstetter, A. M. A Preliminary Report of the Algal Flora
from Selected Areas of Shelby County. Jour. Tenn. Acad. of
Sci., 43; 20-21, 1968.
MacEntee, F. J., G. Schreckenberg, and H. C. Bold. Some
Observations on the Distribution of Edaphic Algae. Soil
Sci., 114; 717-719, 1972.
Olivieri, V., C. Kruse, K. Kawata, J. Smith. "Microorganisms
in Urban Stormwater, " EPA^600/2-77-087, MERL, Cincinnati,
July, 1977.
Diniz, E. V. and W. H. Espey, Jr. "Maximum Utilization of
Water Resources in a Planned Community; Application of the
Storm Water Management Model: Volume I," EPA-600/2-79-050c, 1979.
Metcalf & Eddy, Inc., et al., Storm Water Management Model, in
four volumes, "Final Report," (Vol. I: USEPA Report No. 11024
DOC 07/71, p. 352); "Verification and Testing," (Vol. II:
USEPA Report No. 11024 DOC 8/71, p. 172;) "Users Manual,"
(Vol. Ill: USEPA Report No. 11024 DOC 10/71, p. 249).
U.S. Corps of Engineers. Urban Runoff: .Storage and Treatment
and Overflow Model "STORM." Hydrologic Engineering Center
Computer Program 723-58-L2520, U.S. Army, Davis, California,
May, 1974. ^
133
-------
48. Nash, J. G. The Form of the Instantaneous Unit Hydrograph.
Int. Assoc. of Sc. Hydrol., Pub. 45, Vol. 3_: 114-121, 1957.
49. Horton, R. E. An Approach Towards a Physical Interpretation
of Infiltration Capacity. Soil Sc. Soc. of America, Proc.,
5.: 399-417, 1940.
50. McKee, J. E., and H. W. Wolf. Water Quality Criteria.
May, 1963.
51. Davis, Ernst. "Microbiological Quality of Stormwater Runoff
in The Woodlands," Draft Final Report for Maximum Utilization
of Water Resources in a Planned Community, EPA Research
Grant #802433, October, 1976.
52. Report of the Commission on Water Quality Criteria. FWPCA,
April, 1968.
53. Texas Water Quality Standards. Texas Water Quality Board,
Austin, Texas, February, 1976.
54. Bender, R. J. Ozonation, Next Step to Water purification.
Power, 114 (8): 58-60, August, 1970 (Abs.).
55. Glover, G.E. and Herbert, G.R. Microstaining and Disinfection
of Combined Sewer Overflows. U.S. EPA Report f 11023EVO
06/70, June, 1970.
56. Moffa, P.E., et al. Bench-Scale High-Rate Disinfection of
Combined Sewer Overflows: With Chlorine and Chlorine
Dioxide. U.S. EPA Report #670/2-75-021, June, 1970.
57. Davis, E.M., L.W. Whiteheard, and J.D. Moore. Disinfection.
Jour. Water Poll. Control Fedn., 46 (6): 1181-1191,
June, 1974.
58. Davis, E.M., J.D. Moore, D. Casserly, J. Petros, and W.
DiPietro. Disinfection. Jour. Water Poll. Control Fedn.,
47^ (6): 1323-1334, June, 1975.
59. Horton, R.K. Aquatic Life Water Quality Criteria. Envirn.
Sci. and Tech., 11; 888, 1967.
60. Tsai, Chu-fa. Effects of Chlorinated Sewage Effluents on
Fish in Upper Patument River, Maryland. Chesapeake Sci. 9,
2_: 83, 1968.
61. Arthur, J.W., and J.G. Eaton. Chloramine Toxicity to the
Amphiphod, Gammarus Pseudolimneaus and the Fathead Minnow,
Pimepephales promelas. Jour. Fish Res. Bd. Can., 28: 1841,
1971.
134
-------
62'
Los Angeles, California, June 27-28, 1972
135
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-600/2-79-050a
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
MAXIMUM UTILIZATION OF WATER RESOURCES IN A PLANNED
COMMUNITY
Executive Summary
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.
I. PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Environmental Science & Engineering
Rice University
P.O. Box 1892
Houston, Texas 77001
1O. PROGRAM ELEMENT NO.
1BC822 SOS 2 Task 02
11. CONTRACT/GRANT NO.
802433
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory—Gin, OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
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 Section, FTS 340-6674,
(201) 321-6674 '
16. ABSTRACTgtormwater £rom four watersheds in the Houston area was monitored over a three
year period. Land use in the watershed included undeveloped forest, developing forest,
iully-developed residential and mixed commercial-residential.
Chemical parameters monitored included suspended solids, oxygen demand, organic carbon,
.itrogen, phosphorous, dissolved oxygen, pH, specific conductance and chlorinated hydro-
carbons. Indicator and pathogenic bacterial species were enumerated as well as aquatic
and edaphic algae species. Disinfectant demand and algal bioassays were also conducted.
Relationships have been developed between stormwater runoff quality, quantity and4land
use in an effort to predict pollutant loads. The appearance of a "first flush"' is depen
dent on the parameter measured and watershed characteristics. Rainwater quality contri-
mtes significantly to stormwater pollutant loads, especially in urbanized areas. Modi-
Eying effects of natural biological processes on nitrogen content in the runoff and ef-
fects of the hydrological regime on nutrient limitations were observed. The effective-
less of storage lakes, very positive in the case of suspended solids, were also observed
Ehe Storm Water Management Model (SWMM) was modified to describe the processes occurring
in the watersheds and allowed for (1) separate sewer systems, (2) effects of urbaniza-
tion on base flows, (3) performance efficiency and cost effectiveness of natural drain-
age systems, (4) four additional water quality parameters (COD, Kjeldahl nitrogen,
vibrates, phosphates), (5) hyduologie effects of
17.
KEY WORDS AND DOCUMENT ANALYSIS
a.
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Water quality, Pesticides, Pavements,
Disinfection, Urbanization, Chemical
analysis
Demonstration watersheds,
Hydrologic data, Hydrologi
models, Overland flow,
Fish toxins, Eutrophicatioi
Water sampling, Porous pav J
nent, The Woodlands, Storm
Water Management Model
13B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport)
TTNfTLASSTFIED
21. NO. OF PAG.ES
150
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (Rev. 4-77)
136
4 U.S. GOVERNMENT PBINIINO OFFICE: 1080 -657-060/5430
-------
-------
-------
-------
H
1
en
o
o
^^
K)
VD
\
O
Ul
o
p
g.3 =5
S S. c
0) & CX
^l O
S" ^ ^
5 ^ S.
^° S S"
5:52.
lYi "j
aa n>
H
o> a
~- c:
O
? r^.
3 ^
*"*' a
C3 O'
O)
a- "~-
2 ^
5. 1
?s
§
3 ^-
o. 5
3 w
!=:
O C5
»l
CD Ct>
03 £
§~
^ ^^
Cb CD
l!
s 1
• to
CD
§
^-
CO
0)
Cr-
•^
cb
.^
Q)
CB
m
D
O -n -n
-0 O o
H 3 OD
C < C
Z > 0)
3 m Z
< ._ m
5»tf
38
m °
0)
T3
(0
O
IE
Tl
CO O
O C
o 3.
7T 3"
i
O
w
(A
51 O m
" c? I
CD 33
o i
5' S
o w
" S
(/> to
CD 3
ffl O
3 Q. O
3" • H
§ D m
3 CD O
o> < -j
Up
^3 ?
CD 3 5*
-* G)
gT 5- m
o
c
(0
m
z
3J "D
i s
1-s
rn —j rn
? ? i
CO D D
8 S S
H m
m 0)
O -n
I 5
C)
m
Z
n
-------