

MICROBIOLOGICAL CHARACTERIZATION OF THE WATER AND SEDIMENT
IN KAPAHULU STORM DRAIN SYSTEM AND AT KUHIO BEACH




Kimberly K. Roll and Roger S. Fujioka



Project Completion Report KSDS-4


March 1994


for

The Department of Health
State of Hawaii
P. O. Box 3378
Honolulu, HI 96801




Contract No. ASO Log No. 92-613

Project Period:  April 1, 1992 to December 31, 1994





Principal Investigator:  Roger S. Fujioka
Water Resources Research Center
University of Hawaii
2540 Dole Street
Honolulu, Hawaii 96822


I. MOTIVATION FOR STUDY

A. Use of Bacterial Indicators to Assess Water Quality

	Sewage is the source of many microbial pathogens which are 
transmitted to man by ingestion. When recreational waters are 
contaminated with sewage, the water can serve as the vector for 
the transmission of diseases to man and therefore these sewage-
borne pathogens are also called water-borne pathogens. The 
number and types of sewage-borne pathogens (protozoa, bacteria, 
viruses) are numerous. Moreover, different methods must be 
employed to recover the different pathogens and for some 
pathogens, recovery methods are not yet available. Thus, 
analyzing recreational waters for the presence of all the 
possible pathogens is time consuming, expensive and simply not 
feasible. A feasible way to determine the hygienic quality of 
recreational waters is to determine the concentration of a 
group of bacteria which is naturally found in the feces of man 
and warm blooded animals. These groups of bacteria are called 
indicators of fecal contamination. The criteria (Dutka, 1973, 
Sloat and Zeil, 1987) for selecting a good indicator of fecal 
contamination are:

(1)	It must be consistently and exclusively associated 
with the source (feces, sewage) of the pathogens.
		
	(2)	It must occur in much greater numbers than the 
pathogens.

	(3)	It must not be able to proliferate in the 
environment.

(4)	It must be as stable under environmental conditions 
and to disinfectants as pathogens.

(5)	There must be a simple and unambiguous test for the 
enumeration of the indicator organism.
	
B. Old and New EPA Recreational Water Quality Standards

	Coliform bacteria, characterized as gram-negative, non-
spore forming rods that ferment lactose to form gas within 48 
hours at 37  C was the first group of bacteria used as 
indicator of fecal contamination of recreational waters. 
Initially, the proposed guideline for acceptable swimming water 
in the U.S. was not to exceed 1000-2000 colony forming units 
(CFU)/100 ml of total coliform bacteria (Scott, 1932). This 
guideline was based on the levels of coliforms in Connecticut's 
shoreline water at the time, with 92% of the beaches passing 
the grade. And, conveniently, little intervention was needed in 
meeting these standards (Dufour, 1984). In the mid-1960s, it 
was determined that fecal coliform, a thermotolerant subgroup 
of total coliform, was a better indicator of fecal 
contamination. From 1968 to 1986, the recreational water 
quality standard used in the United States was set at 200 fecal 
coliforms CFU/100 ml (U.S. EPA, 1976).
	In the mid-1970s the Environmental Protection Agency (EPA) 
conducted an extensive microbiological water quality and 
epidemiological study, using improved design to determine the 
predictability of bacterial indicator concentrations in 
recreational waters and incidences of gastroenteritis diseases 
among swimmers. The results of that study showed that 
concentrations of total coliform and fecal coliforms were 
unreliable predictors for water-borne diseases. However, 
concentrations of only enterococci bacteria in marine waters 
and enterococci as well as Escherichia coli in fresh waters 
were reliable predictors for the incidences of gastroenteritis 
diseases among swimmers (Cabelli, 1983).  Due to the findings 
of the Cabelli study, the EPA in 1984 initially established 3 
enterococci CFU/100ml of sample as the standard for marine 
water quality and 20 enterococci CFU/100ml or 77 E. coli 
CFU/100ml for fresh water (U.S. EPA, 1984).  These standards 
were changed in 1986 to make the criteria more closely 
approximate the historically accepted level of 200 fecal 
coliform.  Thus in 1986 EPA revised their recommendations to 35 
enterococci CFU/100ml for marine water and 33 enterococci 
CFU/100ml or 126 E. coli CFU/100ml for fresh water (U.S. EPA, 
1986).

C. Appropriate standards for Hawaii and Tropical Islands

	In Hawaii, Guam, and Puerto Rico, high concentrations of 
fecal indicators have been recovered from environmental waters 
with no known sources of sewage contamination (Fujioka et al., 
1988, Hardina and Fujioka, 1991, Hazen, 1988). For example, 
Hardina and Fujioka (1991) recovered average levels of 7,813 E. 
coli CFU/100ml and 3,220 enterococci CFU/100ml in Manoa stream 
water, far exceeding the Federal standard for fresh 
recreational water of 126 CFU/100ml of E. coli and 33 CFU/100ml 
for enterococci . Moreover, these fecal indicator bacteria were 
present in all soils at high concentrations ranging from 100 to 
10,000 per/100 g of soil. It was therefore, hypothesized that 
these indicator bacteria were multiplying in the soil and were 
being washed into the streams by rain.
	
	If these fecal indicator are naturally present in Hawaii's 
soil and moreover are multiplying in the soils of Hawaii, the 
following two assumptions and criteria for a good fecal 
indicator are not valid in Hawaii: 1. The indicator must be 
consistently and exclusively associated with feces, 2. The 
indicator must not multiply outside of the human intestinal 
tract. These results indicate that the recreational water 
quality standards which are recommended by USEPA are not 
applicable to Hawaii and other tropical islands.

	To find a reliable indicator of recreational water quality 
in Hawaii, Fujioka and Shizumura (1985) tested alternative 
indicators of water quality and determined that Clostridium 
perfringens was superior to the indicators recommended by USEPA 
in Hawaii. C. perfringens is an anaerobic spore-forming bacilli 
which is naturally found in the intestinal tract of humans and 
warm blooded animals.  Since C. perfringens is an anaerobe, it 
cannot multiply in the environment. Moreover, it is the spores 
of C. perfringens which persist in the environment. Since 
spores are extremely resistant, C. perfringens is considered 
too stable and is considered to be a too conservative indicator 
by USEPA.
 	
	However, in tests conducted in Hawaii, C. perfringens was 
found in relatively low numbers as compared to fecal coliforms 
in streams not receiving sewage effluent, while in stream 
samples taken below sewage discharge sites, consistently high 
levels of  C. perfringens were found. In assessing the use of 
C. perfringens, Fujioka and Shizimura, (1985) concluded that it 
was the best indicator for the presence or absence of sewage 
and recommended that for stream water, a standard of not more 
than 50 C. perfringens/100 ml be used. Use of C. perfringens 
was extended to coastal marine waters where it was recommended 
that concentrations of C. perfringens should not exceed 5 
CFU/100 ml and deep coastal waters should not exceed 2 CFU/100 
ml of this bacteria.

	To better determine the quality of coastal waters Fujioka 
(1990) further determined that the aerobic spore forming 
bacteria (bacillus spores) which are naturally found in the 
soil could be used as an indicator that recreational waters are 
contaminated with soil. Since soil is the source of fecal 
coliform, enterococcus, E. coli and bacillus spores, coastal 
waters which contain all of these bacteria can be taken as 
evidence that the source of these bacteria is soil. On the 
other hand, coastal water samples which contain C. perfringens, 
as well as enterococci or E. coli in the absence of bacillus 
spores may be taken as evidence that the source of 
contamination is sewage.

D.  Water Quality of Urban runoff and Storm Drains

	Storm water runoff has for many years been considered as 
nonpoint sources of pollution and therefore the quality of this 
type of water has generally not been monitored. However, 
several studies have shown that storm drain waters can contain 
high levels of fecal indicator bacteria and occasionally 
pathogens as well (Oliveri et al. 1977). Average densities of 
2.38 x 103 fecal coliform CFU/100 ml were recovered by Davis 
(1979) in a study of urban and rural storm water runoff quality 
in Texas.

	Several studies have been conducted on the bacterial water 
quality of storm drains in Hawaii. In 1972, Chun et al 
recovered fecal coliform levels as high as 700/g from dirt and 
dust from three Honolulu streets. Rain wash these dirt and soil 
into storm drains. In 1978, Young, reported a range of 463 - 
2.0 x 103 fecal coliform/100ml and 6.3 x 103 - 7.9 x 103 fecal 
streptococcus/100ml (MPN) in Honolulu's storm drains. In a more 
recent study, Fujioka (1988) recovered a range of 4.7 x 103 - 
1.8 x 105 fecal coliform CFU/100ml and 1.2 x 103 - 1.1 x 105 
fecal streptococcus CFU/100ml in rural and urban storm drains. 
In 1988, Fujioka reported that fecal indicator bacteria were 
recovered from streams in pristine areas of the mountains and 
generally increase as the stream water pass through urbanized 
areas where storm drains discharge into streams.

	In 1990, Fujioka documented that high concentrations of 
fecal indicator bacteria in storm drains and streams were 
adversely affecting the quality of coastal waters which receive 
these waters. In that report, Fujioka concluded that fecal 
indicator recovered from storm drains and streams of Hawaii 
which do not receive sewage effluents should have less of a 
public health significance as compared to fecal indicator 
recovered from sewage. This same issue was addressed by Cabelli 
who was responsible for designing the EPA studies which led to 
the presently accepted water quality standards using 
enterococci and  E. coli. Cabelli (1989) reported that the 
water quality standards are only effective for waters polluted 
by municipal wastewater and sewage sludge discharges. The 
standard is not applicable when bodies of water are 
contaminated with fecal indicator bacteria whose sources are 
bathers themselves, sanitary wastes from boats, storm water 
runoff and direct discharges from lower animals.
	

II. IDENTIFYING A PROBLEM IN HAWAII

A. New Recreational Water Quality Standards for Hawaii

	In 1986, USEPA recommended that every state change the 
marine recreational water quality criteria and standards from a 
geometric mean of 200 CFU/100ml of fecal coliform to a 
geometric mean of 35 enterococci CFU/100ml. The federal 
standard of 35 enterococci CFU/100ml correlated with 19 cases 
per 1000 swimmers. This was unacceptably high for the State of 
Hawaii. As a result, the state of Hawaii in 1990 established 
its marine recreational water quality standard at 7 
enterococci/100 ml based on the EPA prediction that at a 
geometric mean of 7 enterococci/100 ml, the predictable and 
acceptable disease rate was 10 diseases per 1000 swimmers (DOH, 
State of Hawaii, 1990). Of all the states in the U.S., this is 
the most restrictive recreational water quality standard. 
However, the state of Hawaii has still retained the old 
standard of 200 fecal coliform CFU/100 ml for inland waters 
designated for recreational use because of evidence of 
naturally high concentrations of indicator bacteria in Hawaii's 
streams.

B. Review of beach water quality by State DOH

	In 1991, the Hawaii State Department of Health reviewed 
the historical enterococcus and fecal coliform data (1973 - 
1990) from nine sites along Waikiki beach, spanning the 
coastline from Ala Moana Bridge to the Elk's Club Beach 
(Harrigan, 1991). A decreasing trend in the levels of the 
indicator bacteria was observed in a southeasterly direction 
away from the Ala Wai Canal and toward the Diamond Head end of 
Waikiki. However, one exception to this trend was near Kuhio 
Beach, a popular swimming area enclosed by a low seawall and 
adjacent to where the Kapahulu Storm Drain System empties into 
the ocean. Over the 5-year period that enterococci data from 
this area were reviewed, 46 % of the samples exceeded the 
Hawaii State Marine Recreational Standard of 7 enterococci 
CFU/100ml. In 1990 Kuhio Beach was found to "chronically" 
exceed the standard with 71% of the geometric means calculated 
above the standard.  The term chronic refers to situations when 
50% of the geometric means calculated in a calendar year exceed 
the state standard.
	
	As a result of the elevated concentrations of enterococci 
at Kuhio Beach, the DOH was obliged to determine the source of 
enterococci, the risk to humans associated with the elevated 
counts of enterococci in the water at Kuhio Beach and to make 
recommendations for management actions if necessary.  However, 
in reviewing the conditions at Kuhio Beach, it was clear that 
the most obvious source of indicator bacteria was not sewage, 
which is a definite source of water borne diseases but was the 
Kapahulu storm drain which has an outlet under the jetty 
bordering the southeastern end of Kuhio Beach.  In studies 
conducted earlier by Water Resources Research Center of the 
University of Hawaii, it was clearly demonstrated that storm 
drains in Hawaii contain very high concentrations of fecal 
indicator bacteria and especially enterococci (Fujioka et al., 
1988).  Moreover, beaches that receive discharge of stream or 
storm drains can expect to have elevated concentrations of 
indicator bacteria, occasionally exceeding the marine 
recreational water quality standard (Fujioka 1990). Thus, the 
most obvious source of enterococci recovered from Kuhio Beach 
is the storm drain water for Kapahulu storm drain system. 
However, data to clearly document that the Kapahulu storm drain 
is the source of the enterococci in Kuhio Beach is not 
available.

	Clearly, the most important question is whether the source 
of enterococci bacteria in the waters at Kuhio Beach signals a 
health risk to people using that water. The USEPA study upon 
which the enterococci marine recreational water quality 
standard is based provided evidence that at marine beaches with 
known sources of sewage contamination, elevated concentrations 
of enterococci in recreational waters will result in an 
increase in gastrointestinal illness among swimmers in that 
water (Cabelli, 1983). In a follow up study by the USEPA, it 
was shown that if the source of indicator bacteria (E. coli, 
enterococci) polluting a body of recreational water is not 
sewage but a non-point source presumably of animal waste, the 
elevated concentrations of indicator bacteria in recreational 
water was not a reliable predicator of illness among swimmers 
(Calderon et al., 1991).

	In summary, it was documented that waters at Kuhio Beach 
contain concentrations of enterococci often exceeding the new 
recreational water quality standard of 7 enterococci CFU/100ml.  
Although there is no direct data, the available information 
strongly indicate that the major source of enterococci bacteria 
at Kuhio Beach comes from the discharge of storm drain water 
from the Kapahulu storm drain. Although, elevated 
concentrations of enterococci in recreational waters is 
interpreted by the USEPA to be associated with increase in 
illness among swimmers the conditions at Kuhio Beach present 
circumstantial evidence that this may not be the case.  
However, additional evidence is required to strengthen the 
interpretation of water quality at Kuhio Beach.  In this 
regard, two study approaches are required to provide additional 
information.  First, there is a need to monitor the levels of 
indicator bacteria in the Kapahulu Storm Drain System and 
determine if this is a source of indicator bacteria recovered 
from Kuhio Beach.  Second, there is a need to conduct a water 
quality and epidemiological study similar to that conducted by 
the USEPA to measure actual disease incidences among the 
swimmers at Kuhio Beach.




III. OBJECTIVES OF STUDY	

	A microbiological team was established to complete the 
following three separate objectives of the multiphasic studies 
of this project:

	1.  To conduct a microbiological water quality assessment 
of the Kapahulu Storm Drain System and to determine the sources 
of indicator bacteria recovered from the Kapahulu Storm Drain 
System.

	2.  To determine the impact of this storm drain on the 
bacterial quality of water at Kuhio Beach.

	3.  To provide the water quality data at Kuhio Beach for 
the epidemiological study which is described in a companion 
study.


IV. STUDY SITES AND SAMPLING STATIONS

A.  Establishing the Study Area and Sampling Sites

	Mr. Chew Lun Lau of the Department of Public Works (DPW), 
City and County of Honolulu (CCH) was initially consulted to 
obtain blue prints and information about the Kapahulu Storm 
Drain System (KSDS). A crew from CCH showed members of the 
microbiological team the sampling stations at the KSDS and how 
to open the storm drain covers to sample the waters. The 
cooperation and information from the DPW and the Department of 
Parks and Recreation, CCH were invaluable to this study.

	A sketch of the study area which includes the Kapahulu 
Storm Drain System (KSDS), Kuhio Beach area, and the locations 
of the sampling sites are outlined in Figure 1.  The figure 
clearly shows that two major tributaries of the KSDS contribute 
to the water being discharged into Kuhio Beach.  The west 
branch of the tributary collects water primarily from urbanized 
area and was designated the "urbanized tributary".  The east 
branch of the tributary collects water primarily from the 
Waikiki Hotel area and was designated the "hotel tributary".  
Sampling sites were selected to characterize the water in the 
KSDS and the water at Kuhio Beach.

B.  The Urbanized tributary of Kapahulu Storm Drain System 
(KSDS)

	Site 12.  This site collects the water draining from the 
urbanized community before it enters the Zoo area.  Located on 
Monsarrat Ave at the north end of the Waikiki Shell parking 
lot, it is an open gully which receives runoff from the Diamond 
Head area via a channel running under Kapiolani Park.  Under 
dry conditions water trickles through the pipe under Monsarrat 
Ave. toward the Zoo area.  The area surrounding site 12 is 
overgrown with weeds and became more so over the course of the 
study.

	Site 11.  This site is approximately 100 meters downstream 
of site 12 and is located approximately mid-point of the 
Honolulu Zoo between the turtle and monkey displays.  This site 
was selected because there was some concern that waste waters 
from the Waikiki Zoo were entering the KSDS.  Water from this 
site represents storm drain water from the urbanized area and 
the contributions of storm drain water from the Zoo area.

	Site 10.  This site is approximately 200 meters downstream 
from site 11 and is located in the Zoo parking. Water from this 
site contains the storm drain water from Site 11 and also from 
storm water draining the Kapahulu Avenue area.

C.  The Hotel Tributary of KSDS

	Site 8.  This site collects the water draining from a 
Waikiki hotel area and is located on the east sidewalk of Ohua 
St., approximately 75 meters up from Kalakaua Ave.

	Site 9.  This site is approximately 150 meters downstream 
from site 8 and is at the intersection of Paoakalani Street and 
Kalakaua Avenue. It is a low spot and therefore is under tidal 
influence.  Thus, water from this site represents water from 
the hotel tributary area mixed with some intruding ocean water.

D.  Ocean Sites Near the Discharge of KSDS

 	Site 5.  This site is at the mouth of the discharge site 
for the KSDS into the Kuhio Beach area and is located at the 
end of the stone jetty.  Samples of water were taken at the 
large opening at the end of the jetty.  Water from this site 
represents the initial dilution of the storm drain water from 
the KSDS with the ocean water.

	Site 4.  This site is located approximately 75 meters west 
of Site 5 and is seaside of the seawall between Sites 1 and 2.  
This site is susceptible to contamination from Site 5 but the 
water exchange rate at this site is good.

	Site 6.  This site is located just east of Site 5 and can 
be expected to receive some of the water from Site 5. Water 
circulation at this site is good.

E.  Sites Within the Enclosed, Swimming Area of Kuhio Beach

	Site 1.  This site represents the eastern half of the 
enclosed portion of Kuhio Beach and is used extensively for 
swimming. There is an opening of the seawall near Site 5 which 
allows water from Site 5 to enter this enclosed area.

	Site 2.  This site represents the western portion of the 
enclosed area of Kuhio Beach and is a popular swimming area.  
This site is approximately 100 meters from Site 5 and the sea 
wall near this site has a 20 meter opening to increase water 
circulation.
	
F.  Control, Unimpacted Sites

	Site 3.  This site is located approximately 200 meters 
northwest from Site 5 and is west of Site 2. This area is 
outside of the seawall enclosing Kuhio Beach and therefore has 
good water circulation. Thus, this site is not expected to be 
measurably impacted by storm drain water from Site 5 and this 
site was selected as a control, non-storm drain impacted site.

	Site 7.  This site is near Queen's Surf Beach which is 
approximately 200 meters southeast of Site 5. There is good 
water circulation at this site and no discharge of storm water 
near this site. Thus, this site was selected as a control, non-
storm drain impacted site.


V. METHODOLOGY

A. Microbiological Analysis of Water Samples

	 For microbiological analysis, surface water was collected 
in sterile polyethylene bottles.  Samples were stored in an 
iced cooler and transported to the laboratory to be analyzed 
within 6 hrs.

	The enumeration of bacteria in the ocean and storm drain 
water samples was performed using the membrane filtration 
technique as outlined in Standard Methods of Water and 
Wastewater 17th ed. (APHA, 1989).  For recovery of enterococcus 
bacteria the membrane was initially placed on Difco mE agar and 
incubated for 48 hr at 41  C. The membranes were then 
transferred to Difco Esculin Iron Agar (EIA) plate and 
incubated at 41  C for 20 min.  Positive enterococci colonies 
were pink to red colonies that developed a black of reddish-
brown precipitate on the under side of the filter. For recovery 
of fecal coliforms, membranes were placed on Difco mFC media 
and incubated for 24 hr at 44.5  C in a water bath. Positive 
fecal coliform colonies were blue colonies. For recovery of E. 
coli, membranes were placed on Difco mTEC media, and incubated 
for 2 hr at 30  C as a resuscitation step, followed by a 22 hr 
incubation in a water bath at 44.5  C. The membrane was then 
transferred to a filter pad saturated with urease.  After 15 
min. positive E. coli colonies were yellow or yellow-brown 
colonies.  For recovery of C. perfringens, the membranes were 
placed on mCP media (Bisson and Cabelli, 1979) and incubated 
for 24 hr under anaerobic conditions at 45  C. Positive  C. 
perfringens colonies turn from a yellow color to pink on 
exposure to ammonium hydroxide.  For the recovery of bacillus 
spores, 200 ml of water samples were initially pasteurized in 
water bath at 63  C for 1 h to inactivate vegetative cells but 
to allow bacterial spores to survive. The membrane was then 
placed onto mTGA media and incubated at 37  C for 48 hr. 
Positive bacillus spores developed into black colonies.
B.  Microbiological Assay of Soil

	Sterile 500 ml plastic containers were filled 
approximately half way with soil samples which were kept on ice 
and transported to the lab and analyzed with in 4 hr.

	The Most Probable Numeration (MPN) method (APHA, 1989) was 
used to enumerate fecal coliforms, C. perfringens and fecal 
streptococcus in the soil samples.  10 g of samples were added 
to 99 ml of sterile phosphate buffer solution to obtain a 
concentration of 0.1 g/ml.  A series of concentrations (.01, 
.001, and .0001) were then made by transferring 10 ml of the 
higher concentration to 90 ml of sterile phosphate buffer 
solution until the lowest concentration was obtained.  In order 
to obtain the MPN/ Index/g for the bacteria, the following 
equation was used for calculation:

	MPN Index / 100 ml  =  MPN Index  x  10 / largest volume 
tested (1 ml = 0.1 g, thus 100 ml = 10 g)

Therefore:	MPN Index / 100ml = MPN Index / 10 g, and 
dividing the MPN Index by 10 will give an MPN Index/g for the 
sample.

	EC+MUG medium was used in the presumptive phase for fecal 
coliform.  Tubes were incubated for 24-48 hr. at 44.5  C.  All 
positive tubes (growth and acid) within 48 hr were confirmed 
(growth and acid) using EC broth and incubated at 44.5  C.  The 
MUG fluorescence test was also used to confirm the presence of 
E. coli in all positive presumptive tubes. Sulfite-Polymyxin-
Sulfadiazine (SPS) broth was used in the presumptive phase for 
C.  perfringens. Tubes were anaerobically incubated at 37  C 
for 24 hr.  Positive tubes (turbidity or growth) were streaked 
on mCP agar (Bisson and Cabelli, 1979) in the confirmed phase 
and incubated at 41  C for 24 hr.  Red to burgundy colonies 
upon exposure to ammonium hydroxide vapors were confirmed as C. 
perfringens.  Azide-dextrose broth was used in the presumptive 
phase for fecal streptococci.  All tubes were incubated at 35 
 C for 24-48 hr.  Positive tubes (turbidity or growth) were 
streaked on PSE agar for confirmation and incubated at 35  C 
for 24 hr.

	To enumerate indicator bacteria from sand samples the 
method previously reported by Oshiro (1990) was used.  Elutions 
of the sand were made by adding 90 ml of sterile phosphate 
buffer to 90 g of sand sample, which was shaken vigorously for 
10 seconds and let sit for a minute before the supernatant was 
poured into a sterile container.  Another 90 ml of sterile 
buffer was added to the same sand sample and again shaken and 
the supernatant poured into the same sterile container.  The 
eluate was then analyzed for levels of  enterococci, E. coli, 
fecal coliform, C. perfringens and bacillus spores using the 
methods to enumerate bacteria in the ocean and storm drain 
water samples described above.

C.  Measurement of Physical Parameters in Water Samples	

	Acid washed borosilicate bottles were used to collect 
water samples for reactive phosphorus and pH analysis.	 
Reactive phosphate mg/L (PO4) was measured within 24 hr using 
the HACH method and the HACH DR 3000 Spectrophotometer. The 
values were converted to mg/L reactive phosphorus by dividing 
by three. pH was measured (within 6 hours of collection) using 
the Orion pH meter (Model 811).

	Dissolved oxygen was measured at the site using a Yellow 
Springs Instrument Co. Model 57 Dissolved Oxygen Meter. 
Salinity was also measured on site using a refractometer.

D.  Measurement of Dye As Tracer

	The organic fluorescent dye tracer, fluorescein, which is 
non-toxic at low levels was selected because it has a low 
sensitivity to both salinity and temperature changes, and does 
not adsorb to walls. Although the fluorescence of this dye is 
reduced in a low pH environment and has a high photo decay rate 
(Smart and Laidlow, 1977), these conditions are minimized when 
the tracer is used in a storm drain.

	 The storm drain and sewer line samples were collected in 
15 ml borosilicate glass containers and kept in the shade.  The 
samples were processed within 6 hours.	The levels of 
fluorescein in the water samples were determined visually and 
measured using a Turner Filter fluorometer (Model 111).  The 
fluorometer measures the relative intensity of light emitted 
from a liquid containing fluorescent material and the amount of 
fluorescein in proportion to the amount of fluorescent material 
in the liquid.  Calibration curves were constructed for each 
sample, with known amounts of fluorescein.  Because of the high 
background fluorescence levels in storm drain waters, 
calibration curves were constructed for each sample, by 
measuring the sample water and by adding various known amounts 
of fluorescein dye to sample water as compared to distilled 
water.


VI. RESULTS: QUALITY OF WATER AND SEDIMENT IN KSDS

	The quality of the water and sediment in the Kapahulu 
Storm Drain System (KSDS) was determined by obtaining and 
analyzing 18 water and sediment samples from five carefully 
selected sites (see Figure 1). Samples were analyzed for 
indicator bacteria (enterococci, E. coli and fecal coliforms, 
C. perfringens, bacillus spores) and some chemical parameters 
(pH, salinity, dissolved oxygen, reactive phosphorus). The 
results of these analyses are summarized in Table 1, Figure 2 
and detailed in Appendices 1 - 5.

A.  Storm Drain from Hotel Tributary of KSDS

   1.  Site 8. Water samples from Site 8 represent storm drain 
water produced from hotels and other retail outlets which are 
concentrated in the Waikiki area. Water at this site appeared 
stagnant and characterized by low salinity (Average: 0.9 ppt), 
indicating that water at this site was essentially fresh and 
was not being mixed with ocean water or groundwater. Of all the 
storm drain sites tested, water at this site contained the 
highest reactive phosphorus level (Average:1.26 mg/l) 
indicating that run off from this hotel site contain elevated 
concentrations of nutrients. Sources of phosphorus from hotel 
areas which can be expected to enter storm drains are food 
products, cleaning solutions, and fertilizers.

	Geometric mean concentrations of fecal indicator bacteria 
(enterococci: 890 CFU/100 ml, E. coli: 9,291 CFU/100 ml, fecal 
coliform: 24,081 CFU/100 ml) at Site 8 greatly exceeded the 
recreational water quality standards based on these indicator 
bacteria. However, the low concentrations of C. perfringens (11 
CFU/100 ml) indicate that the source of these indicator is 
environmental rather than from sewage or feces of animals. The 
concentrations of bacillus spores (24 CFU/100 ml) was moderate 
and indicative of contribution of soil. In assessing the 
concentrations of indicator bacteria from site 8, it should be 
noted that water at this site was not diluted by brackish 
water.

	At Site 8, the water samples often had a turbid appearance 
and a foul odor. Moreover, there was a period when the water 
had an opaque milky color appearance and this water foamed when 
mixed. On July 2, 1993, the City and County was called to 
investigate the discolored storm drain water.  As a result of 
this investigation, an illegal hook up to the storm drain 
system was discovered.  A sump in the basement of a near-by 
hotel was found to lead directly into the storm drain line and 
was found to contain the same milky white substance.  The 
illegal connection was sealed and since that time water from 
site 8 has not appeared milky.

   2.  Site 9. This site is just downstream of Site 8 and just 
before the storm drain water from the hotel area merges with 
the transport and discharge of all storm from the KDSD into the 
ocean water near Kuhio Beach. Water samples from Site 9 were 
characterized by elevated salinity (Average: 13.8 ppt) 
indicating that storm drain water at this site is brackish. 
Since Site 9 is at a low elevation, some ocean water can be 
expected to intrude to this site. Water from this site 
contained acceptable levels of dissolved oxygen (Average: 5.4 
mg/l) and pH (7.73) and slightly elevated concentrations of 
reactive phosphorus (Average 0.331 mg/l).

	Geometric mean concentrations of all fecal indicator 
bacteria (enterococci: 241 CFU/100 ml, E. coli: 671 CFU/100 ml, 
fecal coliform: 1,492 CFU/100 ml) were lower at Site 9 than at 
Site 8 but still exceeded the recreational water quality 
standards based on these indicator bacteria. The low geometric 
mean concentration of C. perfringens (5 CFU/100 ml) in the same 
water samples indicates that the source of these fecal 
indicator bacteria were environmental rather than from sewage 
or animal feces. Elevated geometric mean concentrations of 
bacillus spores (70 CFU/100ml) in water samples from Site 9 may 
reflect the accumulation of dirt and debris between Sites 8 and 
9. In assessing the lower concentrations of all fecal indicator 
bacteria at Site 9 as compared to Site 8, it should be noted 
that water from Site 9 has been diluted with ocean water and is 
brackish. Moreover, it is well known that inactivation of 
indicator bacteria occurs faster in brackish water than in 
fresh water.

B.  Storm Drain from Urbanized Tributary of KSDS

  1.  Site 12. Water from Site 12 represents storm drain water 
produced by an urbanized community. Water flow from Site 12 was 
usually a slight trickle as it flowed under Monsarrat Ave and 
the water usually had a brown and turbid appearance. Water 
samples from this site had an average salinity of 11.8 ppt 
indicating that the storm drain water from this urbanized area 
was also brackish. The relative elevation of Site 12 was still 
low enough to allow ocean water at high tides to intrude into 
this area. Water samples had acceptable levels of pH (Average: 
7.78) and low dissolved oxygen (Average: 2.2 mg/l). The average 
concentrations of 0.757 mg/l of reactive phosphorus in the 
water at this site was elevated indicating that fertilizers or 
perhaps organic debris were being added to this storm drain 
system.

	Geometric mean concentrations of fecal indicator bacteria 
(enterococci: 3,975 CFU/100 ml, E. coli: 6,270 CFU/100 ml, 
fecal coliform: 5,961 CFU/100 ml in water samples from Site 12 
greatly exceeded the recreational water quality standards based 
on the concentrations of these bacteria. The elevated 
concentrations of C. perfringens (147 CFU/100 ml) in these same 
water samples indicate that sewage or feces from animals is a 
major contribution to the storm water at Site 12. Urbanized 
communities are characterized by pets whose feces often are 
washed into the storm drain. Elevated concentrations of C. 
perfringens in storm drains as compared to free flowing streams 
in Hawaii were previously reported by Fujioka (1990). Thus, 
animal pet feces is the most likely source of the elevated 
concentrations of C. perfringens. The low concentration of 
bacillus spores (4 CFU/100 ml) indicated less contribution of 
soil.

   2.  Site 11. This site is downstream of Site 12 and within 
the Honolulu Zoo. Thus, run off from parts of the zoo also 
discharge into this drainage area. More water was observed at 
this site than at Site 12 indicating that the water values 
observed at Site 12 may be diluted before it reaches Site 11. 
Water from Site 11 was characterized as being brackish (average 
salinity: 15.2 ppt) with acceptable average values for pH 
(7.81) and dissolved oxygen (3.3 mg/l). The average 
concentrations of reactive phosphorus (0.188 mg/l) was 
relatively low and lower than at Site 12.

	Geometric mean concentrations of fecal indicator bacteria 
(enterococci: 376 CFU/100 ml, E. coli: 572 CFU/100 ml, fecal 
coliform: 851 CFU/100 ml, in water samples from Site 11 were 
much lower than in water samples obtained from Site 12 but 
still exceeded the recreational water quality standards based 
on these indicator bacteria. The observation of lower 
concentrations of fecal indicator at Site 11 as compared to 
Site 12 is probably due to the greater volume of water flowing 
at Site 11 which reflect a greater dilution. If the zoo is the 
source of water responsible for this dilution at Site 11, it is 
evidence that storm drain water from the zoo is not a major 
source of fecal indicator bacteria. The slightly lower 
concentrations of C. perfringens (58 CFU/100 ml) at Site 11 as 
compared to Site 12 may also reflect the dilution effect at 
Site 11.

   3.  Site 10. This site is downstream of Site 11 and located 
in the Zoo parking lot. Thus, the water in the storm drain at 
this site represents the storm drain water flowing from Site 11 
but is also mixed with storm drain water draining the Kapahulu 
Avenue area as well. The average values for salinity (16.1 
ppt), the pH (7.77), the dissolved oxygen (3.5 mg/l) and 
reactive phosphorus (0.174  mg/l) in the water samples obtained 
from Site 10 were very similar to that at Site 11. Both Sites 
11 and 12 are low elevations which will allow ocean water to 
intrude to these sites under high tide conditions.

	The geometric mean concentrations of the fecal indicator 
bacteria (enterococci: 1,510 CFU/100 ml, E. coli: 2,089 CFU/100 
ml, fecal coliform: 2,145 CFU/100 ml) were much higher at Site 
10 than at Site 11 and greatly exceeded the recreational water 
quality standards based on the concentrations of these 
indicator bacteria. The concentrations of C. perfringens (31 
CFU/100 ml) and bacillus spores (33 CFU/100 ml) were of 
moderate levels. These results probably reflect the 
contribution of storm water flowing from Site 11 (urbanized 
area) and mixed with storm drain flowing from the Kapahulu 
Avenue area (urbanized/commercial zone).

	It should be noted that the highest counts of bacteria in 
the storm drains were recovered on 9/12/92, the day after 
hurricane Iniki hit the islands. This was an unusual day but 
reflects a natural event which can result in extraordinarily 
high concentrations of fecal indicator bacteria in storm drains 
and detrimental impact on the quality of beach waters 
throughout the islands. Rain and resulting high wind and waves 
contribute to the transport of more forms of pollutants into 
storm drain during these storm events. On this sampling day, 
sand was actually washed up onto Kalakaua Ave indicating that 
sea water, with force, had pushed itself inland and had 
resulted in cleaning out (suspending) much of the sediment in 
the storm drains.

VII. SOURCES OF INDICATOR BACTERIA IN THE KSDS

A.  Animal Feces from the Honolulu Zoo
	
	There has been much speculation that the fecal wastes from 
the Honolulu Zoo were being discharged into the storm drain and 
was the source of most of the indicator bacteria found in the 
KSDS. However, high concentrations of fecal indicator bacteria 
were recovered in all storm drain sites, even those not 
impacted by discharge from the Zoo area. These results are 
taken as evidence that there are sources of fecal indicator 
bacteria other than from the Honolulu Zoo. In this regard, the 
reports of environmental sources (stream, soil) of indicator 
bacteria in Hawaii have been previously reported (Fujioka et 
al, 1988; Hardina and Fujioka 1991).  However, since the 
Honolulu Zoo is an obvious source of fecal wastes, a study was 
conducted to determine its contribution to the storm drain.

	All experiments at the Zoo was conducted with the 
cooperation of Mr. Lloyd Shimazu who showed us the sites and 
answered all questions on operations at the Zoo. It was 
determined that the Honolulu Zoo discharges the fecal wastes of 
animal into a dedicated sewer line which transport this source 
of waste into the sewage line for the City of Honolulu. Thus, 
the animal waste from the zoo should not be entering the storm 
drain. To directly test this hypotheses, fluorescein dye was 
added as a single slug (one minute dose) into the Zoo sewer 
line at two sites at two separate times and water samples were 
taken from storm drains in the Zoo grounds which cross the 
sewer line down stream from the dye injection sites to 
determine if the dye enters the storm drain. Samples were also 
taken from the sewer line down stream from the sites where the 
dye was added and water samples from this site were also taken. 
These water samples were taken before the addition of the dye 
and at least every 10-15 minutes after the addition of the dye. 
All water samples were visually assessed for dye as well as 
measured for fluorescence to detect low levels of dye. A map of 
the zoo showing the dye injection sites and the sampling sites 
in the sewer line and the storm drain lines are shown in Figure 
3. It should be noted that measurable background fluorescence 
was detected in storm drain water (0.0314 to .226 ppb) and in 
sewage samples (0.662 ppb) even before the fluorescein dye was 
added. Thus, background levels of natural fluorescence are 
present in storm water and sewage.

	  Fluorescein dye was first added into sewer line within 
the zoo near the gharial pool at 10:05 AM on August 30, 1993. 
The results of monitoring the storm drain and the sewage 
sampling sites are summarized in Table 2 and graphically 
displayed in Figure 3. In Figure 3, 0 minutes represents 10:00 
AM or 5 minutes before the addition of the dye to the first 
sewer line site within the zoo at 10:05 AM. As shown in Figure 
3, it took approximately 17 minutes before the dye was detected 
at a level almost 1000 times over background at the sewer line 
test site on Monsarrat Avenue. Most of the dye flowed past the 
Monsarrat sewer line site between 30 and 60 minutes after the 
dye was added, reaching concentrations almost 1,000,000 times 
above background. During this period, the dye in the sewage 
line was easily visible to the naked eye and dye was not 
visible at the storm drain monitoring sites. Since the flow in 
the sewer line at the zoo was low and inconsistent, fresh water 
was pumped into the line after the addition of the dye to 
enhance the transport of the sewage and the dye through the 
sewage line. Using this method, most of the visible dye was no 
longer present at the sewage site 60 minutes after the addition 
of the dye at the injection site near the gharials.

	A second dose of fluorescein dye was injected into the 
sewer line behind the Sun bear cage at 11:15 AM (t = 75 min.) 
and water samples at the storm drain sites and the sewage line 
site again monitored for fluorescence.  The results are 
summarized in Table 2 and graphically shown in Figure 3. In 
this experiment, a sharper and smaller peak of dye was observed 
at the sewage site approximately 13 minutes after the injection 
of the dye (85 minutes on Figure 3. The sharper peak represents 
the addition of less dye in a shorter period of time, a 
decision made after our knowledge from the first dye injection 
experiment. During these time periods the water at the storm 
drain 1 (SD1) test site had fluorescein levels ranging from 
.0000224 ppm to .000113 ppm while fluorescein levels at and 
storm drain 2 (SD2) ranged from .0000928 ppm to .002098 ppm. At 
storm drain 2, there was an increase in fluorescence at the 90 
minute time period which could be interpreted as increase of 
fluorescence due to dye or natural background levels of 
fluorescence. Since the level of fluorescence detected was 
below the natural background levels of fluorescent dye, it was 
concluded that this increase was within the range of background 
levels of fluorescence as shown by the background fluorescence 
at Storm drain sites 1 and 2 measured on two separate days (see 
Figure 3).

	In summary, the results of these experiments clearly 
showed that the sewage from the zoo was being transported 
directly to the City and County of Honolulu's major sewage 
line. Moreover, that there was not a direct link between the 
Zoo's sewer line and the storm drain system.  However, our 
study design does not rule out the possibility of a slow or 
indirect contamination between the sewage line and the storm 
drain line. The advantage of using fluorescein dye was the ease 
of visualizing the dye and the sensitivity of the measurements. 
The disadvantage of this dye is the presence of low levels of 
natural background fluorescence in sewage and storm drain 
water. Thus, in future dye studies, selection of dyes should be 
based after determining the background levels of fluorescence 
in the test samples.

B.  Soil as a Source of Indicator Bacteria in KSDS

	In a previous study, Hardina and Fujioka (1991) reported 
that high levels of indicator bacteria are naturally present in 
Hawaii's soil and concluded that soil is the primary source of 
the indicator bacteria recovered from streams. Moreover, that 
rain becomes the carrier of these indicator bacteria. Run off 
from soil is a major contributor to storm water. It was thus 
hypothesized that soil is a major source of the bacteria found 
in storm drains. Soil as a source fecal indicator bacteria is 
exacerbated when pigeons roost in trees above a grassy area or 
walk on the grounds of the grassy area. Areas fitting this 
description are available inside and outside the Honolulu Zoo.

	For this study soil samples were obtained from the 
following three sites: Site 1: grassy area within Honolulu Zoo 
grounds known to be heavily used by pigeons. Site 2: grassy 
area outside of Zoo entrance and known to be heavily used by 
pigeons. Site 3: grassy backyard of private residence near 
Kapiolani Park which was not used by pigeons. The results of 
analyzing the soil samples are summarized in Table 3 and show 
first of all that all three soil samples contained 16,000 
MPN/g of fecal streptococcus. The concentrations of fecal 
coliform in soil samples from Sites 1 and 3 were similar 
(1,000-1300 MPN/g of soil) and higher at Site 2 (16,000 MPN/g 
of soil). These results may reflect the fact that Site 2 is 
protected from sun and drying by the spreading banyan tree. 
Similar levels of C. perfringens (300-500 MPN/g) were recovered 
from soil samples from Sites 1 and 2 with lower levels from 
soil samples from Site 3 (90 MPN/g). These results indicate 
that all soil samples in Hawaii whether it is used by pigeons 
or not contain high levels of fecal coliform and fecal 
streptococci bacteria and much lower levels of C. perfringens. 
The feces of pigeons have been previously determined to contain 
high levels of fecal coliform and fecal streptococci but low 
levels of C. perfringens. Thus, soil in areas used by pigeons 
can be expected to have high concentrations of indicator 
bacteria. However, soils not used by pigeons must still be 
considered sources of indicator bacteria. It therefore, can be 
expected that when it rains, bacteria from soil everywhere will 
be washed into the storm drain systems.


VIII. IMPACT OF KSDS ON KUHIO BEACH

	Kuhio Beach is characterized by being essentially enclosed 
by the rock walled jetty to the east and a line of breakers to 
the south and west. These man-made walls result in an enclosed 
body of water which is calm and a major reason why this beach 
has one of the highest density of swimmers. However, this 
enclosure results in poor water circulation and is one of the 
reason why the quality of water at this beach has not been able 
to consistently meet the state of Hawaii marine recreational 
water quality standard of 7 enterococci/100 ml. More recently, 
the storm drain water being discharged from the jetty near 
Kuhio Beach has been suspected as a source of contaminating the 
water within Kuhio Beach.

	To determine the extent at which the storm water from the 
Kapahulu Storm Drain System (KSDS) was contributing to the 
fecal indicator bacterial concentrations in the ocean waters 
near Kuhio Beach, water samples from selected ocean sites 
(Sites 1, 2, 5 and 7) at Kuhio Beach and Queen's Surf Beach 
were sampled concurrently with the five storm drain sites 18 
times over the course of the study. Water samples from sites 3, 
4 and 6 were concurrently sampled with the storm drain sites on 
15 days. Location of each of the sites are shown in Figure 1 
and each site is described in detailed in the Methodology 
Section. The results of analyzing each water sample for the 
various bacteria and chemical parameters are summarized in 
Table 4 and detailed in Tables 5-13.

A.  Control Sites. To assess the impact of the discharge of 
storm drain water near Kuhio Beach, there is a need to 
determine the quality of ocean water sites which are relatively 
near, used for the same purpose but not likely to be impacted 
by the storm drain discharge for comparative purposes. Two of 
these control sites were selected. The first control site is 
Site 3 which is located farthest west of Kuhio Beach and just 
outside the breakers which demarcate the boundaries of Kuhio 
Beach. There is good ocean current circulation at this site and 
many people use this area for swimming. The second control site 
is Site 7 (Queen's Surf Beach) which is located farthest east 
from Kuhio Beach and is also a popular swimming site. Due to 
ocean currents and its location, this site was considered least 
likely to be impacted by storm drain run off into the ocean.

	The chemical parameters of water samples from these two 
control sites (Sites 3,7) were similar and typical of coastal 
water quality. The results of averaging all the data are 
summarized in Table 4 and show that water at these two sites 
had comparable salinities (33.2 vs 33.6 ppt), pH (8.05 vs 
8.15), dissolved oxygen (7.15 vs 6.89 mg/l) and reactive 
phosphorus (0.19 vs 0.15 mg/l).

	 The geometric mean concentrations for all bacteria in 
water samples from these two control sites (Sites 3, 7) are 
summarized in Table 4. The results show that the concentrations 
of enterococci (1.5 vs 1.9 CFU/100 ml), E. coli (2.1 vs 1.1 
CFU/100 ml), fecal coliform (4.2 vs 2.2 CFU/100 ml), C. 
perfringens (0.2 vs 0.1 CFU/100 ml) and bacillus spores (0.7 to 
0.9 CFU/100 ml) are relatively low and comparable in water 
samples from Sites 3 and 7. Thus, based on the geometric mean 
concentrations of five bacteria in 15 to 18 water samples, it 
can be concluded that the water at these two sites readily met 
all existing recreational water quality standards, including 
Hawaii's very restrictive standard of a geometric mean of 7 
enterococci/100 ml. Based on the data obtained, one can 
conclude that the quality of water at Sites 3 and 7 is good and 
is unpolluted. The data support the selection of Sites 3 and 
Site 7 as good control sites.

	Decisions on whether a site meets recreational water 
quality standards must follow USEPA guidelines on the frequency 
of analyzing water samples from a given site. Officially, the 
guidelines state that five water samples should be taken every 
six days over a 30 day period. This criteria was never achieved 
in this study. However, we calculated the geometric mean 
concentrations of all 15 or 18 water analyzed per site and 
compared that with the recreational standard.

	Besides determining water quality based on geometric mean, 
the quality of water at each site should be examined to 
determine the number of individual water assays which exceeded 
the given recreational water quality standard. This kind of 
data (Table 5-9) can be used to compare relative quality from 
one site to another as a measure of sporadic pollution events. 
When this approach is taken, Site 3 exceeded the enterococci 
standard in 2 of the 15 sampling days (14 CFU/100 ml) while 
Site 7 exceeded the enterococci standard 3 of the 18 sampling 
days (range: 11 - 56 CFU/100 ml). None of the samples from Site 
3 or Site 7 exceeded the old recreational water quality 
standard of 200 fecal coliform/100 ml. These results point out 
the difficulty in consistently meeting the new state of Hawaii 
marine recreational water quality standard of 7 enterococci/100 
ml as compared to the EPA recommended standard of 35 
enterococci/100 ml and especially the old fecal coliform 
standard.
	
	In summary, based on averaging the measurements for 
various parameters in 15 to 18 water samples, the quality of 
water at Site 3 and Site 7 was good and acceptable for 
recreational use. These results support our decision in 
selecting Sites 3 and 7 as control sites. Although, these sites 
were designated as unpolluted, there were individual days when 
the concentrations of enterococci at these sites exceeded the 7 
enterococci standard used by the State of Hawaii for marine 
recreational standard. Values at these sites will be used to 
compare similar values obtained at the selected test sites 
within and near Kuhio Beach.
B.  Storm Drain Water Ocean Discharge Site. Site 5 represents 
the ocean site closest to the point where all of the storm 
drain water from the KSDS initially enters the ocean. As a 
result, Site 5 can be assumed to be the site for zone of 
initial dilution of the storm drain water and water samples 
from this site should show the greatest impact from the storm 
water. The expected number of indicator bacteria at Site 5 will 
be a function of the volume of storm drain water being 
discharged and the extent to which the ocean water will be able 
to mix and dilute the storm water. Measurements of water 
salinity at Site 5 can be taken as evidence of the mixing and 
dilution of storm drain water by ocean water.

	The results of averaging all the water samples analyzed at 
this site are summarized in Table 4. Based on the chemical and 
physical parameters (salinity, pH, dissolved oxygen, reactive 
phosphorus), the values obtained at Site 5 was similar to that 
obtained from the control sites (Site 3, Site 7). These results 
indicate that during these 18 days of sampling, the amount of 
fresh water from the storm drain had been effectively diluted 
and mixed by ocean water to the extent that the measured 
chemical parameters could not determine whether water from this 
site was impacted by the storm drain. In these same water 
samples, the geometric mean concentrations of fecal indicator 
bacteria (enterococci: 4.5 CFU/100 ml, E. coli: 10.3 CFU/100 
ml, fecal coliform: 16.9 CFU/100 ml) were definitely higher 
than at Sites 3 and 7 indicating a measurable impact of the 
storm drain water at Site 5. However, this impact was minor as 
the geometric mean concentrations of the fecal indicator 
bacteria were well below the recreational standards based on 
these bacteria. Geometric mean concentrations of  C. 
perfringens (0.5 CFU/100 ml) and bacillus spores (2.0 CFU/100 
ml) at this site were so low that its significance was 
difficult to assess.
	
	As mentioned earlier, the same data can be assessed by 
examining the results of each sample to determine the frequency 
or the individual days when water quality at Site 5 exceeded 
the recreational standard. The results of the analysis of water 
samples for each of the individual days are tabulated in Table 
5-9 and show that of all the marine water sites, Site 5 
exceeded the 7 enterococci/100 ml level most frequently (6/18 
water samples) with concentrations of enterococci of 7, 29, 35, 
48, 53 and 107 CFU/100 ml. Thus, on individual days, the 
concentrations of enterococci at Site 5 was substantial 
indicating that on a given day, the discharge of the storm 
drain water can greatly impacted on the quality of water at 
Site 5. The results in Table 5 also show that no enterococci 
were recovered on 4 of the 18 days indicating that on those 
days, there was no measurable impact of the storm drain 
discharge at Kuhio Beach.

	In the interpretations of the results of water quality at 
Site 5, it must be recognized that good circulation of ocean 
current occurs at this site. Although, we previously determined 
that storm drain water in the KSDS contained consistently 
elevated concentrations of fecal indicator bacteria, the impact 
on the quality of ocean water at Site 5 is dependent to a large 
extent on the volume of the storm drain water and how well this 
storm water is mixed with ocean water. Based on our 
observation, the flow of storm water is generally low and 
sporadic. If the volume of storm drain water being discharged 
into the ocean is small, the dilution and circulation effects 
of ocean water will minimize the impact on the quality of the 
ocean water. However, when the volume of the storm drain 
increases such as following a rainfall event or by the 
discharge of sump water into the storm drain, a significant and 
measurable impact on the quality of water at Site 5 can be 
expected.

C.  Sampling Sites Within Enclosed Area of Kuhio Beach.

	For sampling purposes, Kuhio Beach was divided into two 
sections. Site 1 was the sampling site representing the body of 
water in Kuhio Beach which was closer to Site 5 (the storm 
drain discharge site) and Site 2 was the sampling site 
representing the half of Kuhio Beach which was farther away 
from Site 5. The results of concurrently analyzing the quality 
of the water in the storm drain and within Kuhio Beach are 
summarized in Table 4. Based on the averages of the chemical 
and physical measurements of the 15-18 marine water samples, 
the water samples at Sites 1 and 2 had comparable values for 
salinity (33.4 vs 33.4 ppt), pH (8.02 vs 7.92), and dissolved 
oxygen (6.44 vs 6.39 mg/l). The average reactive phosphorus 
value was slightly higher at Site 1 (0.034 mg/l) as compared to 
Site 2 (0.014 mg/l). This can only be taken as suggestive 
evidence that that storm water is having a measurable impact at 
Site 1.

	The geometric mean concentrations of the five bacteria in 
all of the water samples are summarized in Table 4 and show 
similar concentrations of enterococci (2.0 vs 2.3 CFU/100 ml), 
E. coli (4.9 vs 6.5 CFU/100 ml), fecal coliform (12 vs 10.5 
CFU/100 ml), C. perfringens (0.3 vs 0.2 CFU/100 ml) and 
bacillus spores (2.4 vs 2.2 CFU/100 ml) at Sites 1 and 2. Thus, 
based on geometric mean concentrations of fecal indicator 
bacteria in 15 - 18 water samples, the quality of water within 
Kuhio Beach is good and readily met Hawaii's strict 
recreational standard of 7 enterococci/100 ml.

	Another way to assess the impact of the storm drain water 
within Kuhio Beach is to determine the frequency the individual 
water samples exceeded the recreational water quality standards 
of 7 enterococci/100 ml. The results for each of the 15 to 18 
sampling days as tabulated in Tables 5-9 show that at Site 1, 3 
of 18 water samples exceeded the enterococci standard with 
concentrations of 7, 27, and 60 CFU/100 ml. At site 2, the 
frequency of exceeding the enterococci standard was 4 of 18 
water samples at concentrations of 7, 10, 10, and 14 CFU/100 
ml. These results indicate that on individual days, 
concentrations of enterococci exceeding the standard of 7 
enterococci/100 ml can be expected within Kuhio Beach at Sites 
1 and 2. During these same sampling days, Site 5 exceeded the 
enterococci standard 6 of the 18 days and three of these days 
correlated with the elevated concentrations observed at Sites 1 
or 2. These results are suggestive that the source of the 
enterococci at Sites 1 and 2 is water from Site 5 (storm 
water).

D.  Sampling Sites outside Kuhio Beach

	To determine the transport of the storm water outside of 
the Kuhio Beach area after it has been discharged into the 
ocean near Site 5, two other sampling sites east and west of 
Site 5 were selected. Site 4 is just outside the breaker of 
Kuhio beach and 75 meters west of Site 5 while Site 6 is just 
east of Site 5. Due to the predominating wave action, water at 
Site 5 can be expected to be transported to Sites 4 and 6.

	The average chemical parameters of water samples obtained 
from Sites 4 and 6 are summarized in Table 4 and show 
comparable values for salinity (33.4 vs 32.9 ppt), pH (8.07 vs 
8.21), and dissolved oxygen (7.66 vs 7.38 mg/l). These values 
are similar to water samples obtained from the control sites 
(Sites 3, 7), indicating that the water at Sites 4 and 6 have 
been mixed well with ocean water. Although slightly elevated 
average concentrations of reactive phosphorus were observed at 
Site 4 (0.041 mg/l) and Site 6 (0.022 mg/1) as compared to Site 
3, it was difficult to interpret this data since the average 
reactive phosphorus at Site 5 was low (0.015 mg/l).

	The geometric mean concentrations of the five bacteria in 
water samples obtained at Sites 4 and 6 are summarized in Table 
4. The results show that the bacterial concentrations at Site 4 
were comparable to the values obtained from the control Sites 3 
and 7 and were lower than at Site 6. At Site 6, slightly 
elevated concentrations of enterococci (3.5 CFU/100 ml), E. 
coli ( 7.0 CFU/100 ml), and  fecal coliform (11 CFU/100 ml) 
were observed. However, the concentrations of these bacteria at 
Site 6 were well below the recreational water quality standard 
based on the concentrations of these bacteria. The 
concentrations of C. perfringens (0.4 vs 0.5 CFU/100 ml) and 
bacillus spores (1.7 vs 3.3 CFU/100 ml) at Sites 4 and 6 were 
low but slightly higher than the control Sites 3 and 7. These 
results are consistent with the observation that water from the 
storm drain is transported both east and west of the discharge 
point (Site 5). Due to the wave current, more of the storm 
water is transported east or away from the opening to Kuhio 
Beach.

	Assessment of the results of individual water samples for 
Sites 4 and 6 are summarized in Table 5-9 and show that none of 
the 15 water samples from Site 4 exceeded the 7 enterococci 
/100 ml standard while 3/15 water samples from Site 6 exceeded 
this standard with concentrations of 37, 61 and 308 CFU/100 ml. 
The extremely high count of 308 CFU/100 ml was obtained in 
water sample taken the day after Hurricane Iniki and does not 
represent a typical day. The overall results support the 
conclusion made earlier that most of the storm drain water 
being discharged at Site 5 is transported toward Site 6 rather 
than Site 4.

IX. MONITORING BEACH WATER QUALITY FOR EPIDEMIOLOGICAL STUDY

	One of the most serious questions related to the discharge 
of the Kapahulu storm drain water into the ocean water is 
whether this practice results in substantially increasing the 
incidences of diseases among swimmers at Kuhio Beach. To 
address this question, a 16 month long epidemiological/water 
quality study was initiated. In this section we report the 
water quality monitoring data obtained by the microbiological 
team in support of the epidemiological study. The results of 
the epidemiological findings are contained in a companion 
report by authored by Morens, Roll and Fujioka (KSDS-5/1994).

 	To support the epidemiological study, water samples from 
four sites were analyzed for enterococci, E. coli, fecal 
coliforms, C. perfringens and bacillus spores on the same days 
that the epidemiological team were interviewing swimmers at 
Kuhio Beach. Selected water samples were also analyzed for 
staphylococci bacteria, reactive phosphorus, pH and salinity. 
The primary swimming sites within Kuhio Beach (Site 1, Site 2), 
and the mouth of the storm drain discharge site (Site 5) were 
sampled on 164 days, with both morning and afternoon samples 
being taken on 141 of those days. Site 7, the control site, was 
sampled on 146 days with morning and afternoon sampled taken on 
133 of those days.

A.  Site 7. This popular swimming site (Queen's Beach) was 
selected as the control site and is characterized by absence of 
nearby storm drain discharge. The monthly as well as the 
cumulative geometric means for all the indicator bacteria in 
water samples obtained from this site over the sixteen month 
period are summarized in Table 14. The results show that the 
cumulative geometric means over the entire 16 month period for 
all of the indicator bacteria (enterococci: 1.4 CFU/100 ml, E. 
coli: 1.5 CFU/100 ml, fecal coliform: 2.0 CFU/100 ml, C. 
perfringens: 0.4 CFU/100 ml bacillus spores: 2.9 CFU/100 ml) 
were low. Moreover, none of the 16 monthly geometric means at 
Site 7 exceeded the most restrictive recreational water quality 
standard of 7 enterococci/100 ml. The geometric mean 
concentrations for enterococci were similar whether the samples 
were taken in the morning (1.4 CFU/100 ml) or in the afternoon 
(1.3 CFU/100 ml). These results indicate that the quality of 
water at Site 7 was excellent and suitable for recreational 
use.

B.  Site 5. This is the site where the storm water from the 
Kapahulu storm drain is discharged into the ocean. It is 
located outside of Kuhio Beach. This area is generally not used 
for swimming but some swimmers have been observed to jump off 
the pier into the ocean water at this site and surfers do ride 
their boards into this area. The results of the quality of 
water at this site is summarized in Table 15 and show that the 
cumulative geometric mean over the entire 16 month period for 
enterococci was 7.1 CFU/100 ml or equal to the marine water 
recreational water quality standard. However, this standard was 
exceeded in 6 of the 16 months at this site. The six 
consecutive months (September - February) represents the rainy 
months of the season and most likely reflect the greater volume 
of storm drain water flowing into the ocean. The geometric mean 
for enterococci obtained from the morning samples (8.4 CFU/100 
ml) was higher than the geometric mean obtained from this same 
site during the afternoon (6.7 CFU/100 ml).

	For E. coli the cumulative geometric mean over the 16 
month period was 9.4 CFU/100 ml which is higher than that of 
enterococcus but well below the federal fresh water 
recreational standard of 126 E. coli/100 ml. The geometric mean 
concentration of E. coli was higher during the morning (11 
CFU/100 ml) than in the afternoon (8.7 CFU/100 ml). For fecal 
coliform, the cumulative geometric mean was 14 CFU/100 ml which 
was higher than that of E. coli but well below the old 
recreational water quality standard of 200 fecal coliform/100 
ml. The geometric mean concentration of fecal coliform was 
higher during the morning (18 CFU/100 ml) than in the afternoon 
(11 CFU/100 ml). For C. perfringens the cumulative geometric 
mean was 0.68 CFU/100 ml which was below the 5 CFU/100 ml 
guideline suggested for beach water. The geometric mean 
concentration of this bacteria was very low and similar in the 
morning (0.72 CFU/100 ml) or in the afternoon (0.69 CFU/100 
ml). For bacillus spores the cumulative, the morning and the 
afternoon geometric means were similar and ranged from 5.0 to 
5.2 CFU/100 ml.

	In summary, based on the cumulative geometric means of all 
the indicator bacteria, the water at Site 5 contained highest 
concentrations of fecal coliform (13.9 CFU/100 ml), followed by 
E. coli (9.4 CFU/100 ml), enterococci (7.1 CFU/100 ml), 
bacillus spores (5.2 CFU/100 ml) and lowest concentrations of 
C. perfringens (0.7 CFU/100 ml). The same relative 
concentrations of these bacteria were also measured in the 
storm drain (see Table 1). These results indicate that the 
source of the indicator bacteria at Site 5 is the storm water. 
Moreover, since the cumulative mean concentration of 
enterococci at Site 5 was equal to the marine recreational 
water quality standard of 7 enterococci/100 ml, it must be 
concluded that the discharge of storm water had a deleterious 
impact on the quality of water at Site 5. This impact of the 
storm drain water at Site 5 was more evident when the geometric 
mean concentrations of enterococci were exceeded during the six 
winter months at concentrations ranging from 8 to 30 CFU/100 
ml).

C.  Site 1. This is the site within the enclosed area of Kuhio 
Beach which is closest to Site 5 (see Figure 1). It is a 
popular swimming site and swimmers at this site were 
interviewed by the epidemiological team. The data summarizing 
the quality of water at this site are summarized in Table 16 
and show that the cumulative geometric mean concentration of 
enterococci over the 16 month period was 4.9 CFU/100 ml. This 
was well below the recreational standard of 7 enterococci/100 
ml. However, this standard was exceeded in four of the sixteen 
months. The four months (October to January) represented the 
rainy season and the same months when concentrations of 
indicator bacteria at Site 5 showed an increase. Thus, the 
evidence indicate that the source of the elevated 
concentrations of enterococci at Site 1 was the storm drain. 
This became apparent during the winter months when increase 
rainfall results in increasing the volume of storm water and 
indicator bacteria being discharged into Site 5. Some of the 
water at Site 5 which is contaminated with indicator bacteria 
such as enterococci is then transported to Site 1. The 
cumulative geometric mean concentrations of enterococci 
obtained from morning samples (3.9 CFU/100 ml) was slightly 
lower than the geometric mean obtained from this same site 
during the afternoon (4.9 CFU/100 ml).

	For E. coli the cumulative geometric mean over the 16 
month period was 6.6 CFU/100 ml, well below the fresh 
recreational water quality standard of 126 E. coli/100 ml and 
none of the monthly geometric mean approached this standard. 
The geometric mean concentrations of E. coli was higher in the 
morning samples (7.3 CFU/100 ml compared to the afternoon 
samples (5.9 CFU/100 ml). For fecal coliform, the cumulative 
geometric mean was 9.2 CFU/100 ml but well below the old 
recreational standard of 200 fecal coliform/100 ml and none of 
the monthly geometric mean approached this standard. Higher 
geometric mean concentrations of fecal coliform were recovered 
from this site in the morning samples (10.6 CFU/100 ml) as 
compared to the afternoon samples (9.3 CFU/100 ml). For C. 
perfringens, the geometric mean concentrations in all samples, 
in the morning and afternoon samples were approximately 0.5 
CFU/100 ml which was well below the 5 CFU/100 ml guideline 
suggested for beach water. For bacillus spores, the cumulative 
geometric mean was 4.9 CFU/100 ml with geometric mean of 4.6 
CFU/100 ml in the morning and 5.0 CFU/100 ml in the afternoon.

	In summary, based on the cumulative geometric mean 
concentrations, Site 1 contained highest concentrations of 
fecal coliform (9.2 CFU/100 ml) followed by E. coli (6.6 
CFU/100 ml), followed by enterococci and bacillus spores (4.9 
CFU/100 ml) and finally by C. perfringens at only 0.5 CFU/100 
ml. This is the same relative concentrations of indicator 
bacteria as was observed at Site 5 and provide additional 
evidence that the source of indicator bacteria at Site 1 was 
the storm drain water. The overall quality of water at Site 1 
met the stringent Hawaii marine recreational water quality 
standard of 7 enterococci/100 ml. However, this standard was 
exceeded during four of the sixteen months indicating that the 
storm drain does impact on the quality of water at Site 1.

D.  Site 2.  This site is also within the enclosed area of 
Kuhio Beach but west of Site 1 and farther away (100 meters) 
from Site 5. As with Site 1, Site 2 is heavily used by swimmers 
and swimmers at this site were also interviewed by the 
epidemiological team. The data summarizing the quality of water 
at this site are summarized in Table 17 and show that the 
cumulative geometric mean concentration of enterococci over the 
entire 16 month period was 3.5 CFU/100 ml. This was well below 
the level observed at Site 1 and well below the Hawaii marine 
recreational standard of 7 enterococci/100 ml. However, this 
standard was exceeded in three of the sixteen monthly means. 
These three months (September, October, November) represent the 
rainy season and are the same months when elevated 
concentrations of enterococci were recovered from Site 5 and 
Site 1. These results indicate that the source of elevated 
enterococci at Site 2 is the same source (storm water) which 
are affecting Sites 5 and Site 1. The results indicate that the 
indicator bacteria originate from storm water which is 
discharged into the ocean at Site 5 and transported to Sites 1 
and then to Site 2. The cumulative geometric mean concentration 
of enterococci obtained in the morning (3.5 CFU/100 ml) was 
similar to that  obtained from afternoon (3.3 CFU/100 ml) 
samples.

	For E. coli, the cumulative geometric mean over the 16 
month period was 4.5 CFU/100 ml, well below the fresh 
recreational water quality standard of 126 E. coli/100 ml and 
none of the monthly geometric means exceeded this standard. The 
geometric mean concentrations of E. coli during the morning 
samples (5.9 CFU/100 ml) was higher than in the afternoon 
samples (2.8 CFU/100 ml). For fecal coliform, the cumulative 
geometric mean was 7.0 CFU/100 ml, well below the old standard 
of 200 fecal coliform/100 ml and none of the monthly geometric 
mean approached this standard. Higher geometric mean 
concentrations of fecal coliform were recovered from the 
morning samples (8.8 CFU/100 ml) as compared to the afternoon 
samples (4.8 CFU/100 ml). For C. perfringens, the cumulative 
geometric mean, as well as the mean recovered from the morning 
and afternoon samples were low and very similar ranging form 
0.4 to 0.5 CFU/100 ml. This was well below the 5 CFU/100 ml 
guideline for beach waters. For bacillus spores, the cumulative 
geometric mean was 5.6 CFU/100 ml with geometric mean of 4.5 
CFU/100 ml in the morning samples and 6.1 CFU/100 ml in the 
afternoon samples.

	In summary, based on cumulative geometric mean 
concentrations, Site 2 contained highest concentrations of 
fecal coliform (7.0 CFU/100 ml) followed by bacillus spores 
(5.6 CFU/100 ml), followed by E. coli (4.5 CFU/ 100 ml), 
followed by enterococci (2.5 CFU/100 ml) and finally by C. 
perfringens at 0.4 CFU/100 ml. Thus, the relative 
concentrations of the fecal indicators at Site 2 were similar 
to that at Site 1 and Site 5 but the measured concentrations 
were lower. These results provide additional evidence that the 
source of indicator bacteria at Site 2 was the storm drain 
water. Moreover, that the quality of water at Site 2 was better 
than at Site 1 and generally met the Hawaii marine recreational 
water quality standard. However, this standard was exceeded 
during three of the sixteen months. As with Site 1, the months 
with the elevated enterococci represented the rainy season and 
indicate that the source of the enterococci is storm water 
which initially impacts Site 5 and is then transported to Site 
1 and then to Site 2. The lowest concentrations of indicator 
bacteria at Site 2 most likely represent greater dilution and 
some inactivation as the indicator bacteria is transported from 
Site 5 to Site 2.


X. SOURCES OF INDICATOR BACTERIA RECOVERED FROM KUHIO BEACH

A.  Sewage as source of indicator bacteria. Presence of fecal 
indicator bacteria in water initially suggests that the source 
of the fecal indicator bacteria is sewage. However, a sanitary 
survey of the Kuhio Beach area indicated that sewage was not 
being discharged into the waters at Kuhio Beach. It should be 
noted that based on USEPA studies (Cabelli, 1983, Dufour, 
1984), there will be a predictable number of diarrheal diseases 
among swimmers as the concentrations of fecal indicator 
bacteria from a sewage source increases.

B.  Storm drain as source of indicator bacteria. The storm 
drain water from the KSDS was designed to be discharged into 
the ocean near Kuhio Beach near Site 5. The results of this 
study have documented the presence of high concentrations of 
fecal indicator bacteria in the storm drain water of the KSDS 
and discharge of these bacteria at Site 5 near the Kuhio Beach. 
The same relative concentrations of indicator bacteria observed 
in the storm drain was also observed at Sites 5, Site 1 and 
Site 2. Moreover, the monthly geometric means of these fecal 
indicator increased at all three of these sites during the 
rainy season. These results support the known effects of 
rainfall which washes more fecal indicator bacteria into the 
storm drain, increases the volume of water in the storm drain, 
and thereby increasing the load of pollution at Site 5.

	Based only on the expected concentrations of enterococci 
at Sites 5, 1 and 2, it may be prudent to designate these sites 
unsuitable for swimming during the winter months or anytime it 
rains. However, the increase in the concentrations of 
enterococci at Sites 5, 1 and 2 does not automatically mean 
there will be an increase in disease incidences as predicted by 
the USEPA studies. Those studies which established the 
enterococci standard documented that at recreational sites 
where point source of sewage discharge were present, the 
incidences of diseases increased as the concentrations of 
enterococci in the water increased. However, in a more recent 
USEPA study, Calderon et al (1991) determined that increase in 
enterococci in water from non-point source without a source of 
sewage did not result in the corresponding increase of 
diarrheal diseases among swimmers using that water. Since the 
source of indicator bacteria at Sites 5, 1 and 2 is from the 
storm drain, which represents non-point source rather than a 
sewage source, the conditions more closely approximates the 
study of Calderon et al (1991) rather than the Cabelli study 
(1983). The very low concentrations of C. perfringens at Sites 
5, 1 and 2 also indicate that sewage is not a major source of 
enterococci at these sites. These results suggest that the 
concentrations of enterococci at Kuhio Beach may not be related 
to increased concentrations of swimming associated diarrheal 
diseases. However, only the results of an epidemiological study 
can resolve the health risks related to the concentrations of 
enterococci at Kuhio Beach.

C.  Swimmers as sources of staphylococci bacteria. In a 
previous study, Charoenca and Fujioka (1993) reported a 
correlation between the number of swimmers at Kuhio Beach and 
the concentrations of staphylococci bacteria in the water. 
Swimmers were concluded to be sources of staphylococci bacteria 
in waters at Kuhio Beach. Moreover, the increase in 
staphylococci in the water at Kuhio Beach was correlated with 
the increase in numbers of reported skin infections among 
children swimming at this beach.

	In the present epidemiological study, illness due to 
diarrheal diseases as well as skin diseases were asked. An 
attempt was made to determine the contribution of staphylococci 
by swimmers by analyzing six water samples from Sites 1, 2, 5 
and 7 for concentrations of staphylococci. The results of the 
presumptive staphylococci concentrations are summarized in 
Table 18 and show that Site 1 and Site 2, which consistently 
had higher numbers of bathers contained the highest geometric 
mean concentrations of staphylococci.  Site 1 had a geometric 
mean of 643 CFU/100 ml in the morning samples and 402 CFU/100 
ml in the afternoon samples. All 12/12 (100%) of samples from 
Site 1 exceeded the suggested guideline of not more than 100 
staphylococci/100 ml (Favero et al., 1964).  Site 2 had the 
next highest geometric mean concentrations of staphylococci 
with 139 CFU/100 ml during the morning samples and 84 CFU/100 
ml during the afternoon samples. Fifty percent (6/12) of the 
water samples from Site 2 exceeded 100 staphylococci/100 ml. At 
Site 5, the mean concentrations of staphylococci was 39 CFU/100 
ml in the morning and 41 CFU/100 ml in the afternoon with only 
1/8 or 8% of the samples exceeding the 100 staphylococci/100 ml 
level. Site 7, the control site had the lowest geometric mean 
concentrations of staphylococci with 21 CFU/100 ml in the 
morning samples and 26 CFU/100 ml in the afternoon samples. 
However, 3/12 (25%) of the samples at this site exceeded the 
100 staphylococci/100 ml guideline.

	To determine if swimmers were sources of staphylococci, a 
statistical analysis based on correlation coefficient indicated 
that the number of swimmers were correlated to the 
staphylococci concentrations in the water at sites 1 and 2.  
The results (Table 20) show a significant correlation at Site 1 
(R=0.823, p=0.0443) in afternoon samples, and at Site 2 
(R=0.921, p=0.009) in morning samples. These results and the 
observation of lower concentrations of staphylococci at Site 5 
and higher concentrations of staphylococci at Sites 1 and 2 
support the conclusion that swimmers do contribute to levels of 
staphylococci bacteria in the ocean water.		

D.  Sand as source of indicator bacteria. In previous studies, 
Fujioka and Oshiro (1990) reported that sand was a major source 
of indicator bacteria for the water at Hanauma Bay. To 
determine whether sand is a major source of indicator bacteria 
at Kuhio Beach, dry and wet sand samples from Sites 1, 2, and 7 
were analyzed for the various indicator bacteria on three 
different days. The results of the sand analysis are summarized 
in Table 19 and show that dry sand samples contain much higher 
concentrations of indicator bacteria than wet samples. These 
results are similar to that observed in earlier Hanauma Bay 
study and reflect the observation that soil in the sand is the 
source of the bacteria and wet sand has been washed free of 
soil by the surf. Dry sand contain more soil and therefore more 
indicator bacteria.

	At Site 1, dry sand contained geometric mean 
concentrations of 1,323 CFU/100 g of enterococci, 297 CFU/100 g 
of E. coli, 3,088 CFU/100 g of fecal coliform, 23 CFU/100 g of 
C. perfringens and 2,837 CFU/100 g of bacillus spores. At Site 
2, dry sand contained geometric mean concentrations of 3,679 
CFU/100 g of enterococci, 4.4 CFU/100 g of E. coli, 63 CFU/100 
g of fecal coliform, 72 CFU/100 g of C. perfringens and 1,054 
CFU/100 g of bacillus spores. At Site 7, dry sand contained 
geometric mean concentrations of 2.2 CFU/100g of enterococci, 
1.1 CFU/100 g of E. coli, 0 CFU/100 g of fecal coliform, 6.2 
CFU/100 g of C. perfringens and 88 CFU/100 g of bacillus 
spores.

	The results of sand analysis showed that more indicator 
bacteria are present in the sand at Kuhio Beach (Site 1, 2) 
than at Site 7. Based on the study conducted at Hanauma Bay, 
the source of the bacteria in the sand is soil. This is 
supported by the observation that the sand at Kuhio Beach 
visibly contains more dirt than the sand at Site 7. Two factors 
may increase the soil content of the sand at Kuhio Beach. 
First, there are more swimmers who will track more soil to the 
sand at Kuhio Beach. Second, due to the breakers, wave action 
at Kuhio Beach is minimized and the soil in the sand cannot be 
washed out to sea as can be expected at Site 7. Thus, the 
source of some of the indicator bacteria in the water at Kuhio 
Beach must come from the soil in the sand. However, the 
contribution of sand to the actual numbers of fecal indicator 
recovered from the waters in Kuhio Beach is not known.

E.  Swimmers or people as sources of indicator bacteria. 
Swimmers are known sources of indicator bacteria and more 
importantly known sources of various types of water borne 
pathogens. In an enclosed area such as Kuhio Beach, swimmers 
can be expected to contribute to the pollution load in the 
beach water. The results of the staphylococci study showed that 
swimmers at Kuhio Beach are contributing to the high levels of 
staphylococci at Kuhio Beach.  However, the source of 
staphylococci is the skin of people whereas the source of fecal 
indicator bacteria from people is feces.

	To determine the impact of swimmer density to the 
concentrations of fecal indicator bacteria, a correlation 
coefficient assessment was made. The results of this assessment 
indicated no correlation between density of swimmers and 
excessive concentrations of indicator bacteria in the waters at 
Kuhio Beach. One possible explanation for this finding is that 
the contribution of fecal indicator bacteria from swimmers are 
much less than the contribution of these indicator bacteria 
from the storm drain. This condition was supported by the 
following observations. First, the concentrations of fecal 
indicator  was higher at Site 5 than at Sites 1 and 2. Second, 
evidence was obtained that the source of the indicator bacteria 
recovered from water samples at Kuhio Beach was the storm drain 
water being discharged at Site 5. These results show that the 
contribution of indicator bacteria in Kuhio Beach is 
predominantly controlled by the storm drain and masks the lower 
contribution of indicator bacteria from other sources such as 
people and beach sand.






























REFERENCES

American Public Health Association, American Water Works 
Association, Water Pollution Control Federation.  1989.  
Standard Methods for the Examination of Water and 
Wastewater, 17th edition. American Public Health 
Association, Washington D.C.

Bisson, J.W., V.J. Cabelli.  1979.  Membrane filter enumeration 
method for Clostridium perfringens.  Appl Environ 
Microbiol. 37:55-66.

Cabelli, V.J.  1983.  Health effects criteria for marine 
recreational waters.  USEPA Report EPA-600/1-80-031.

Calderon, R.L., E. W. Mood, A.P. Dufour.  1991.  Health effects 
of swimmers and nonpoint sources of contamination water.  
International Journal of Environmental Health Research.  
1:21-31

Charoenca, N., R. Fujioka.  1991.  Assessment of Staphylococcus 
bacteria in Hawaiis marine recreational waters.  Wat Sci 
Tech.  27:283-289.

Chun, M.J., R.H.F., Young, G.K. Anderson.  1972.  Wastewater 
Effluents and Surface Runoff Quality.  Water Resources 
Research Center Tech. Report No. 63.

Davis, E.M.  1979.  Maximum Utilization of Water Resources in 
Planned Community Bacterial Characteristics of Stormwaters 
in Developing Rural Areas.  USEPA Report EPA-600/2-79-
050f.

DOH, State of Hawaii.  1992.  Chapter 11-54 Hawaii 
Administrative Rules on Water Quality Rules.

Dufour, A.P.  1984.  Health effects criteria for fresh water.  
USEPA Report EPA600/1-84-004.

Dutka, B.J.  1975.  Coliforms are an inadequate index of water 
quality.  J Environ Health.  36:39-46.

Favero, M.S., C.H. Drake, C.B. Randall.  1964.  Use of 
staphylococci as indicator of swimming pool pollution.  
Public Health Reports.  79:61-70.

Fujioka, R.S., L.K. Shizumura.  1985.  Clostridium Perfringens, 
a reliable indicator of stream water.  J Water Poll Cont 
Fed. 57:986-992.

Fujioka, R.S., K. Tenno, S. Kansako.  1988.  Naturally 
occurring fecal coliforms and fecal streptococci in 
Hawaiis freshwater streams.  Toxicicity Assessment. 
3:613-630.


Fujioka, R.S.  1990.  Evaluation of Clostridium perfringens as 
a suitable indicator for recreational water quality 
standard.  Project Completion Report.  ASO LOG No. 89-355.  

Hardina, C.M., R.S. Fujioka.  1991.  Soil:  The environmental 
source of E. coli and enterococci in Hawaii's streams.  
Environ.  Toxicology. 6:185-195.

Harrigan, J.  1991.  Bacteriology on the Relationship Between 
Monitoring Data and the Marine Recreational Waters 
Standard Part A: Waikiki Beach, Oahu.  Internal Department 
of Health Document.

Hazen, T.C.  1988.  Fecal coliforms as indicators in tropical 
waters:  A review.  Toxicity Assessment:  International 
Journal.  3:461-477.

Oliveri, V.P., K.W. Cornelius, K. Kawata, J. E. Smith.  1977.  
Microorganisms in Urban Stormwater.  USEPA Report EPA-
600/2-77-087.

Oshiro, R.  1990.  Sand, soil, and pigeon dropping: Sources of 
indicator bacteria in the waters at Hanauma Bay beach, 
Hawaii.  Paper presented at the 90th Annual meeting of 
American Society for Microbiology.

Scott, W.J.  1932.  Survey of Connecticut's Shore Bathing 
Waters.  Public Health Engr. 19:316.

Sloat, S., C. Zeil.  1987.  The use of indicator organisms to 
assess the safety of public water.  HACH Technical Center 
for Applied Analytical Sciences.

Smart, P.L., I.M.S. Laidlaw.  1977.  An Evaluation of Some 
Fluorescent Dyes for Water Tracing.  Water Resources 
Research.  13:15-33.
 
U.S. Environ. Prot. Agency.  1976.  Quality criteria for water 
(EPA/440/9-76/023).  Washington D.C.: US EPA.

U.S. Environ. Prot. Agency.  1984.  Water quality criteria: 
Request for comments.  Fed Reg. 51:8-12-8016.

U.S. Environ. Prot. Agency.  1986.  Ambient water quality 
criteria for bacteria - 1986 (EPA440/5-84-002). Washington 
D.C.: US EPA. 

Young, R.  1978.  Urban storm water runoff in Hawaii.  
International Symposium on Urban Storm Water Management.  
University of Kentucky, Lexington, Kentucky. July 24-27.   


0-6

	

0-5
 



