| Final
Report on the BST Click to
view final report of this project in PDF format.
The
Bacterial Source Tracking Project
Over
the past decade, there have been an increasing numbers of
reports of E. coli in lakes and streams in Indiana
and nationwide. E. coli is a bacterium
that is found in human and animal fecal material in relative
high numbers. It is a health concern because its presence
in water is an indicator of fecal contamination. E. coli can
enter receiving waters from one of three routes: overflow
from combined sewer systems, drainage from septic tanks and
contributions from animal fecal material. The first two situations
represent direct health concerns due to potential exposure
to disease causing organisms by what is known as the oral-fecal
route. The latter situation is also significant due to the
potential for transfer of pathogens from animals to humans.
Furthermore, in proposing remediation activities, it is important
to be able to distinguish between animal contributions and
human contributions to the total E. coli load to receiving
waters.
Elevated
levels of E-coli have long been identified through monitoring
conducted by the St. Joseph River Watershed Initiative, the
Fort Wayne - Allen County Department of Health, and the Indiana
Department of Environmental Management (IDEM). However, to
accurately determine whether the sources of contamination
are human (from failing septic systems or sewage treatment
facilities), domestic animal, livestock or wildlife, additional
research is necessary. The BST program is an effort to produce
a database of characteristics of known fecal contaminants
that can then be compared to water samples taken from throughout
the watershed.
The
St. Joseph River Watershed Initiative was awarded a Section
319 grant (01-383) in 2002 which funds the Bacterial Source
Tracking (BST) project, a joint effort of the Initiative and
the Biological Sciences Department of Indiana University –
Purdue University Fort Wayne (IPFW). Dr. Deborah
Ross is the chief investigator for
the project.
While
E. coli is the most common indicator of fecal contamination,
other bacteria, such as fecal enterococci are also used as
indicators. Currently available methods for pinpointing the
source of fecal contamination in the environment fall into
two major categories: use of antibiotic resistance patterns
and use of molecular biological techniques. The latter have
proved their worth in identification of pathogenic strains
in cases of food borne disease where the number of samples
needed is limited. However, they are time consuming and expensive.
Furthermore, a database linking individual strains of fecal
bacteria and sources of pollution (i.e. human, waterfowl,
swine, cow, etc.) does not exist. Development of such a database
would be prohibitively costly and time consuming. The use
of antibiotic resistance patterns, however, offers speed,
reproducibility, low expensive and a proven track record.
In
the antibiotic resistance method, which utilizes replica plating
technology, bacterial strains are isolated from the environment,
and characterized as to their sensitivity to a range of antibiotics.
The basis for this method is that if bacteria have been exposed
to a given antibiotic, they will develop resistance to it;
it they haven't been exposed, they will not be resistant.
Thus the growth pattern of bacterial strains from water is
matched against standard strains from known sources. Such
a database has been developed by researchers in Virginia,
using fecal enterococci as the indicator of choice. With this
database, researchers were able to identify inputs of fecal
bacteria from a farm as resulting from uncontrolled access
of cattle to the stream in question, and the remediation of
this situation, installation of fencing, was simple and expedient.
Researchers in Florida
have used this method in subtropical waters.
The
following graph illustrates the percentage of SJRWI weekly
grab samples that exceeded the MCL for E. coli from 1996-2003
at Site 100, the confluence of Cedar Creek with the St. Joseph
River, just north of Fort Wayne.
The
objectives of the current SJRWI study are to develop a database
similar to those used in Virginia and Florida for the St.
Joe River watershed. It was initially proposed to determine
whether or not the Virginia database was suitable for the
situation in northern Indiana . It was felt that the development
of our own database might be too time consuming to complete
within the proposed time limits. Upon consultation with the
Virginia researchers, it was decided to go ahead and develop
the database from sources specific to northern Indiana . This
would guarantee the applicability of our database to our own
watershed. The concerns regarding the time frame were due
to an overestimation of the time required to determine the
antibiotic resistance patterns of strains of enterococci from
known sources. The first step in development of the database
was to determine the significant sources of fecal contamination
in the St. Joseph River watershed. Advice from the SJRWI management
and county extension agents resulted in the following list
of sources: human, domestic pet, swine, beef cattle, dairy
cattle, horses and wildlife (raccoons, deer, etc). A sampling
schedule was developed which would result in the collection
of samples and their return to the laboratory where they could
be immediately sampled for fecal enterococci.
The
initial sampling consisted of dilution of a weighed amount
of fecal material (10 g) in physiological saline and plating
of a volume (0.1 ml) of the appropriate dilutions onto a bacteriological
medium designed to enhance to growth of enterococci. Dilution
of the sample is necessary into order to obtain separation
of bacterial cells within the fecal material such that individual
cells will grow into isolated, defined colonies which appear
red. These colonies can then be transferred into individual
wells on a microwell plate. Each well contains a volume of
0.2 ml of medium which will turn black when enterococci are
grown. Thus, individual bacterial cells can be cultured and
tracked. Following growth in the microwells, a replica plating
device can be used to transfer the bacteria from the microwells
to bacteriological growth medium containing various concentrations
of antibiotics.
For
the SJRWI database, from 1 to 5 concentrations of 9 antibiotics
were used for a total of 30 combinations. Because of the small
volumes involved and the use of the replica plater, 96 bacterial
strains can be inoculated onto the 30 media in a matter of
15 to 20 minutes. After two days' growth, strains are scored
for growth or nongrowth on each of the antibiotic concentrations.
This method is so rapid that it was possible to test over
1000 bacterial strains within 6 weeks. The next step is to
enter each strain into a computer program which is capable
of performing discriminant analysis. JMP IN was selected because
the researchers in Virginia
had used this program and our use of it would make our two
databases compatible. In this statistical program, analysis
of variance is first performed on the dataset, then strains
from known sources are compared and grouped based on similarity.
The goal is to be able to separate strains from a given source
based on the susceptibility to each of the 9 antibiotics.
The
project was initiated in spring of 2001 with a grant from
the Fort Wayne Community Foundation, which funded setup of
the database and baseline sampling. Testing continued throughout
the summer and fall of 2002 under the IDEM grant, resulting
in further refinement of the ability to positively identify
fecal contaminant sources. During the summer 2002 sampling
period, antibiotic resistance analysis was performed twice
(mid summer and late summer) on 18 of the monitoring sites
established within the watershed, plus one additional site
on the St. Joseph River
near the point at which water is withdrawn for the City of
Fort Wayne . All
of the samples from both dates displayed either human, geese
or horse antibiotic resistance profiles as the major source.
A significant overlap in identifying horse, dairy cattle and
human sources remained to be further refined.
Two
other factors which have been reported to affect the analysis
are flow rate and distance downstream from a contaminant source.
Therefore, urban sites where combined sewer overflow (CSO)
is likely to impact the water quality have been identified
and sampled during and after a storm event to determine the
impact of combined sewer overflow on the observed antibiotic
resistance patterns. Other efforts focused on identifying
locations in the watershed where particular sources of fecal
bacteria are likely to impact the water (i.e. a livestock
operation located near a stream or ditch). A sampling regime
has been designed to take samples immediately above the site,
and then at various distances downstream of the site, to determine
how far a site might be expected to impact the results of
antibiotic resistance analysis.
During
the summer of 2002 and the first part of 2003, all of the
monitoring stations were sampled a total of three times. Results
of these analyses indicate that the major sources of contamination
are human and geese, with geese particularly prominent in
late August-early September.
In
addition, five monitoring sites were selected for a more frequent
(every three weeks) sampling schedule to determine the optimum
number of samples to take from a particular site in order
to obtain the most accurate picture of pollution sources at
that site. This part of the project is not yet complete.
Another
aspect of the project is to examine individual tributaries
to determine if there are changes in sources along the length
of the tributary. For this part of the project, we identified
four tributaries with known sources of fecal contamination:
Dibbling Ditch, Garret City Ditch and the lower Cedar Creek,
Nettle Creek, and a series of short tributaries in the Grabill
area. Preliminary analyses of these samples indicate that
BST can detect changes in sources of pollution along the length
of a watershed as well as changes in land use patterns. For
example, there is a pronounced horse/domestic pet pattern
in bacteria from tributaries in the lower St. Joseph River,
which is consistent with Amish farming practices. The Garrett
City Ditch shows a time variation in sources, with a significant
human contribution in some but not all samples, which is consistent
with a sewage contribution to the Garrett City Ditch from
the overloaded sewage treatment plant (renovations and upgrades
to the treatment plant were completed in late 2003 to correct
this situation). These examples indicate that by combining
knowledge of land use with BST, an accurate picture of sources
fecal pollution to a tributary can be obtained.
Given
this ability to detect sources along a tributary, it is important
to examine the watershed in more detail and use both BST analysis
and land use information to pinpoint pollution sources and
work with landowners to find ways to reduce or eliminate the
pollution. The SJRWI has included several locations in the
Cedar Creek among its current monitoring sites, but these
do not cover the complete watershed; in particular, the Little
Cedar Creek and Willow Creek have not been analyzed for BST.
|
