Transcript a.curious.relationship.between.E.coli

A Curious Relationship between E. coli and Enterococci
or
Where Do Those Enterococci Come From?
There’s an old adage about war: that it’s 99 % pure boredom and 1 % sheer panic. While I don’t mean to imply that
analyzing water quality data is anything at all like war, those engaged in this work know it to be mostly “been there,
seen it, done that.” Every now and then though, something catches your attention; something makes you sit up and
exclaim “what the #$!@# was that all about?” I intend posting little essays about my own “eureka” moments in
the hope that others might have some thoughts on these subjects; hopefully, someone out there is wondering about
the same thing – or, best of all, knows the answers and will enlighten me.
What follows is a brief introduction to the subject – and then my story.
Members of two bacteria groups, the coliforms and fecal streptococci, are used as indicators of possible sewage
contamination because they are commonly found in human and animal feces. Although they are generally not
harmful themselves, they indicate the possible presence of pathogenic (disease-causing) bacteria, viruses, and
protozoans that also live in human and animal digestive systems. Their presence in streams suggests that pathogenic
microorganisms might also be present and that swimming and eating shellfish might pose a health risk. Since it is
difficult, time-consuming, and expensive to test directly for the presence of a large variety of pathogens, water is
usually tested for coliforms and fecal streptococci instead.
Bacteria are reported as the “most probable number” (MPN) of bacteria in 100 milliliters (100 ml, about 4 ounces)
of water; a statistical test is usually used instead of directly counting bacteria so the actual number remains an
estimate. California Public Health requirements for bacteria counts are complicated and vary somewhat by
jurisdiction. Basically, four separate indicator organisms are used to judge the Public Health risk. For freshwater
recreational use (swimming), the total coliform limit is “no more than 10,000 per 100 ml in a single sample,” for
enterococcus, E. coli and fecal coliforms the single sample limits are 61, 235 and 400, respectively. The State of
California reduces the single sample total coliform limit to 1,000, if the fecal/total coliform ratio is greater than 0.1
(in other words, as long if more than 10 % of the coliforms are of fecal origin).
100,000
concentration (MPN/100 ml)
enterococci
E. coli
10,000
1,000
100
10
AT1
AT2
AT3
CG1
GA1
GA2
GS1
GS2
LC1
MY1
SJ2
SP1
enterococci
E. coli
10,000
1,000
100
10
VR
14
VR
12
VR
11
VR
10
VR
09
VR
07
VR
06
VR
05
VR
04
VR
03
1
VR
01
concentration (MPN/100 ml)
100,000
While preparing an article for the
ChannelKeeper newsletter
(http://www.stream-team.org/) on the
big January 2005 storm that struck
Southern California, I came across
something that aroused my interest: the
relative concentrations of enterococci
during the storm were, almost
invariably, far higher than E. coli.
The two bar charts show these
concentrations: we sampled on Jan. 8 on
the Ventura River (VR, bottom) when
flows were still relatively modest, and
on Jan. 9 in Goleta (top) when things
started to really rip. Notice that only
three Ventura locations had higher E.
coli than enterococci concentrations,
and this occurred at only a single Goleta
site. Notice also that the Goleta
differences are generally greater. I find
it interesting that the higher January 9
flows seem to give relatively lower E.
coli concentrations, and that the sites
with the least differences are either
urban (AT3) or highly contaminated
(VR04). VR14 is a big exception – and
it’s a relatively pristine site with higher
E.coli than enterococci concentrations.
enterococci-SJ023
E. coli-SJ023
TC-SJ023
enterococci-SJ166
E. coli-SJ166
TC-SJ166
1,000,000
25
15
10,000
10
1,000
5
100
11/28/01
12:00 PM
0
11/28/01
6:00 PM
11/29/01
12:00 AM
11/29/01
6:00 AM
11/29/01
12:00 PM
11/29/01
6:00 PM
flow (cfs)
(MPN/100 ml)
bacteria concentration
20
100,000
What brought this all to mind was
this figure, which I made using Santa
Barbara County, “Clean Water,” data
from a storm in Nov. 2001
(http://www.countyofsb.org/project_clea
nwater/). I’ve plotted an approximate
hydrograph (discharge is modeled from
County rainfall data) to give an
indication of when samples were
collected in reference to the storm’s
progress. The county monitored five
locations on San Jose Cr. during this
event; I’ve only plotted 2: one at either
end, SJ023 below Hollister Ave. (the
lower, industrial section), and SJ166 at
N. Patterson above Cathedral Oaks Rd.
(an upper site at the urban boundary).
Concentrations at in-between sampling
points fall, more or less, between these
two extremes. The first sample is a prestorm sample; and the results are what
we might expect: lower concentrations
for all indicators at the upper boundary
of Goleta development, higher below
Hollister at the bottom end.
The last sample appears to have been taken just past the storm peak. Notice that the upper end now has higher total coliform and
enterococci concentrations than the lower sampling point (the lower location was sampled about 20 minutes before the upper). There is
almost no development above this location except an orchard. Notice too that E. coli concentrations are lower than enterococci values,
and that the difference becomes greater as the storm progresses. Indeed, at the time of the last samples, E. coli numbers were decreasing
as enterococci increased: for the last sampling, enterococci concentrations are 1.5 orders of magnitude higher than E. coli at the most
undeveloped site. That this location, at the peak of the storm, had the highest enterococci numbers is quite intriguing. If I wasn’t
supposed to know better, I’d say that these enterococci concentrations will be relatively unrelated to pathogenic concentrations, that they
are being flushed out of soil and other relatively innocuous environments, and that the high numbers indicate high survivability, and
even reproduction, of enterococci in these environments.
bacteria concentration
80,000
enterococci: Dec-02
E. coli: Dec-02
enterococci: Feb-03
E. coli: Feb-03
60,000
40,000
20,000
0
Milpas
Serena
San
Pascual
West Side
Drain
Victoria
Drain
Haley
Haley below Montecito
above Filter
Filter
This is City of Santa Barbara, “Clean Creeks Project” data
(http://www.santabarbaraca.gov/Government/Departments/Parks_and_Recreation/Creeks_Division_Main): samples collected during
storms from various drains within the City and at locations along Mission Creek. I don’t know at what time these samples were
collected, nor in what order, but the December event was very big storm on Dec. 16 (peak flows ~ 1200 cfs), and the Feb. sampling
occurred during the second pulse of another reasonably sized storm on Feb. 12 (flows ~ 200 cfs). Again there is a pattern, stronger for
the second event than for the first, of higher enterococci than E. coli concentrations.
I have a working hypothesis here: Only in heavily contaminated areas, or in dense urban clusters, will we see enterococci stormflow
concentrations equal to, or less than, E. coli concentrations. The greater the percentage of undeveloped or agricultural crop land, the
higher enterococci concentrations will be relative to E. coli numbers. I think the cause has to do with relatively greater survivability
and reproduction of enterococci in the mild Santa Barbara climate.
600
concentration (MPN/100 ml)
2002-2004 E. coli geomean
2002-2004 enterococci geomean
400
200
SP
1
SJ
2
SJ
1
Y1
M
LC
1
S2
G
S1
G
A2
G
A1
G
1
G
C
AT
3
AT
2
AT
1
0
E. coli geomean
enterococci geomean
400
200
VR
15
VR
14
VR
13
VR
12
VR
11
VR
10
VR
09
VR
08
VR
07
VR
06
VR
05
VR
04
VR
03
VR
02
0
VR
01
2001-2004 geomeans
concentrations (MPN/100 ml)
600
If we look at non-storm samples,
we see the opposite picture: E. coli
concentrations are usually higher
than enterococci. The EPA single
sample limits for freshwater contact
recreation reflect this: 235 for E.
coli vs. 61 for enterococci. The
figure shows geomeans (for the time
spans indicated) for Channelkeeper
samples; these represent baseflow
averages since it is extremely rare
for a storm to coincide with a
sampling day.
Notice that almost all the Goleta
samples show higher E. coli than
enterococci concentrations – these
reflect urban nuisance waters and
agricultural runoff. The only site
that has higher enterococci
concentrations is Maria Ygnacio:
which only flows for a week or so
after big storms. We see the same in
Ventura: most sites have higher E.
coli numbers except those that are
(1) relatively pristine, (2) feature
golf-course watering, or (3) only
flow during and shortly after storms:
represented by sites VR09 through
VR15.
1,000
(MPN/100 ml)
E. coli concentration
10,000
100
10
E. coli-MC00
1
Mar-02
Jun-02
Sep-02
Dec-02
E. coli-RS02
Mar-03
Jun-03
E. coli-MC07
Sep-03
Dec-03
1,000
(MPN/100 ml)
enterococci concentration
10,000
100
10
1
Mar-02
Jun-02
enterococci-MC00
enterococci-RS02
enterococci-MC07
Sep-02
Mar-03
Sep-03
Dec-02
Jun-03
Dec-03
Again, this is City of Santa
Barbara “Clean Creeks Project”
data: 5 point moving geomean
averages at three sampling locations
along Mission Creek (MC00 is at
Montecito St, at the tidal limit;
MC07 is at the Mission Cyn. Bridge
(a relatively pristine upper
catchment area with some
residential development); and RS02
is Rattlesnake at Skofield Park (allundeveloped Forest Service land).
However, unlike the previous data,
these are almost all baseflow
samples; in other words, samples
taken between storm periods and
during the April to September dryseason. While the higher-elevation,
more pristine and less developed
sites show lower concentrations for
both indicator organisms (as
expected), enterococci
concentrations are noticeably higher
than E. coli at the two relatively
undisturbed locations. At times E.
coli concentrations at RS02 and
MC07 are an order-of-magnitude
lower.
1,000
(MPN/100 ml)
bacteria concentration
10,000
100
10
1
Mar-02
Jun-02
Sep-02
E. coli-MC00
enterococci-MC00
E. coli-RS02
enterococci-RS02
Dec-02
Mar-03
Jun-03
Sep-03
Dec-03
I’ve re-plotted data from the last figure on a single graph to better show the contrast between enterococci and E. coli
concentrations at Rattlesnake (RS02, an undeveloped site) and Mission at Montecito St. (MC00, in the downtown area at the tidal
limit). Again, these are 5 point moving geomeans. Notice at Montecito Street, where we are dealing mainly with urban nuisance
waters, E. coli concentrations are higher than enterococci – in line with the respective EPA limits (interestingly, the ratio of the
respective limits is 3.85, i.e., 235 divided by 61, while the average ratio of the data is 3.61). However, E. coli concentrations at
Rattlesnake are lower than enterococci; the average ratio between E. coli and enterococci is 0.41 – not quite, but almost – the
reciprocal of the Montecito Street ratio.
25
E.coli/entero-SJ023
E.coli/entero-SJ062
E.coli/entero-SJ166
FC/TC-SJ023
FC/TC-SJ062
FC/TC-SJ166
flow
0.1
20
15
10
0.01
5
0.001
11/28/01
12:00 PM
0
11/28/01
6:00 PM
11/29/01
12:00 AM
11/29/01
6:00 AM
11/29/01
12:00 PM
11/29/01
6:00 PM
flow (cfs)
indicator bacteria ratio
1
Finally, this again is County of
Santa Barbara “Clean Water”
data from the Nov. 2001 storm on
San Jose Creek. This was a
relatively small storm which
followed the biggest storm of the
year five days earlier; the
hydrograph is modeled flow
below Hollister and should be
considered approximate. I’ve
plotted bacteria ratios for three of
the locations the County sampled
(the other two show similar
relationships): one at either end,
SJ023 at the end of Kellog, below
Hollister Ave., and SJ166 at N.
Patterson above Cathedral Oaks
Rd., and the third at Hollister
itself.
The concentration ratios are E. coli to enterococci, and fecal coliform to total coliform (FC/TC). The FC/TC ratio is a standard
California test for water quality; ratios greater than 0.1 indicate a greater probability of fecal contamination and intestinal illness, and
trigger a lower total coliform limit: reduced from 10,000 to 1,000 MPN/100 ml. The County did not measure FC, so I’ve multiplied
E. coli concentrations by 1.7 to estimate FC (1.7 is the ratio between the Calif. FC and EPA’s E. coli limits, 400 and 235 MPN,
respectively, and it implies that 60 % of the fecal coliforms in a sample were E. coli).
The ratios at each site track each other extremely well (except for a single point), and the relationships between locations are
maintained. If we assume that both ratios indicate the relative probability of fecal contamination, the story they tell is reasonably
logical and confirms what we would typically envision happening: ratios decreasing from pre-storm levels with the first flush of
runoff from relatively clean impervious surfaces, increasing early on the rising hydrograph limb, and then again decreasing as the
storm reaches and passes its peak. During the main part of the storm, the upper-elevation site shows about five times (half an order
of magnitude) less contamination than the lower locations. If enterococci are behaving like total coliform concentrations, it is not
unreasonable to believe they also must originate in sources not directly associated with fecal contamination.