Presentation outline. Project summary. Project products. Products summary 4/3/2010

Transcription

1 Presentation outline Microbial Source Tracking Sources of E. coli in Upper Fountain Creek 06 April 2010 Presented by Don Stoeckel, PhD to Colorado Dept. Public Health and the Environment E. coliwork group via Webinar from Columbus, Ohio What is microbial source tracking, why is it important, and what is the state of the science? How was microbial source tracking applied to Fountain Creek? Results validation Project summary 16-months of sample collection, 30-month overall effort done primarily by USGS Samples collected from April 2007 to September 2008 I was the technical lead with USGS at the time Involvement at the local, state, and federal levels Pikes Peak Area Council of Governments (lead) Four affected municipalities El Paso County Colorado Department of Public Health and the Environment Two offices of the US Geological Survey Pueblo, Colorado and Columbus, Ohio Project products Collection and analysis of a large number of samples E. coli Microbial source tracking markers Nutrients and field parameters Wastewater organic chemicals Series of public meetings in Colorado Final report to stakeholders in final stages of USGS approval before submittal to Water Research for consideration Products summary The data are publicly available, and provide higher-resolution information about prevalence and patterns of fecal contamination in Upper Fountain Creek. The report is the first known attempt to validate allocation of fecal contamination among sources Prior grey literature attempts were based just on concentrations of markers Basis for my talk at the National Monitoring Conference, Denver, Colorado, the afternoon of April 28, Hold on there, fella back up a minute. what are we talking about? 1

2 UNDERSTANDING QUANTITATIVE MICROBIAL SOURCE TRACKING Why microbial source tracking? There are acceptability limits for fecal contamination of resource waters Most standards are based on the fecalindicator bacteria E. coli(inland) and enterococci (inland or marine) If the standard is violated, what do you do? E. coliand enterococci are found in a lot of hosts The CWA says the concentration must be brought down or the waterway must be reclassified Regardless of source or health risk Regardless of cost Concept of microbial source tracking The intestinal bacteria of animal groups are expected to be different because of: Basic habitat is different in different animals Basal temperature Food supply and characteristics Digestive system and configuration Natural selection in these habitats Competition for space and nutrients Pathogenicity factors (adhesins, toxins) Interactions between microorganisms Prior exposure to agents like antibiotics The MST process 1. Choose host-associated targets that are in the feces of local source groups 2. Characterize reference material (also known as manure and sewage) from local sources 3. Test water for fecal contamination 4. Associate contamination with sources Collection of reference material Library patterns used for MST Phenotypic patterns Sample ARA panel: Antibiotic resistance Carbon utilization Fatty acid profiles Genotypic patterns REP-PCR Macrorestriction followed by PFGE Ribotyping MLST and other sequencing patterns ampicillin (5, 16, 32 µg/ml) cephalothin (5, 16, 32 µg/ml) chlorotetracycline (20, 40, 80 µg/ml) cloramphenicol (5, 16, 32 µg/ml) erythromycin (25, 50, 100 µg/ml) irgasan (=triclosan) (0.01, 0.1, 0.5 µg/ml) kanamycin (1, 5, 16 µg/ml) oxytetracycline (1, 5, 16 µg/ml) penicillin G (25, 50, 100 U) streptomycin (1, 5, 16 µg/ml) sulfamethoxazole (50, 200, 512 µg/ml) tetracycline (1, 5, 16 µg/ml) 2

3 Problems with libraries Error rates are unacceptably high The most realistic validation studies showed rates of correct classification that were better than random, but still very low (25% to 40% correct) Failed to discriminate major from minor contributors in poop soup tests Cost of collecting and evaluating large numbers of isolates Tens of thousands of isolates are required for reasonable representation Genetic markers used for MST Cultivation based Enrichment (esp, toxin genes) Host-specific markers within isolates Phage types Whole DNA based Bacteroidales R. coprophilus Bifidobacteria Novel fecal anaerobes from Field and others, 2003 Microbial source tracking markers Many host-associated markers have been proposed Based on 16S rdna phylogeny (co-evolution) Research groups include Drs. Field, Layton, Wuertz, Okabe, Reischer, Gilpin Based on host-microbe interactions Research groups include Drs. Santo Domingo, Shanks, Carson, Rose, Scott, Sadowsky, Call Based on host-specific viruses Research groups include Drs. Girones, Jofre, Jiang, Noble (apologies to research groups that were omitted or not fully attributed) Problems with genetic markers Not correlated with public health outcomes Not yet correlated with regulatory fecal indicators Limit of detection may be above or below the trigger point of fecal contamination where fecal-indicator based regulations are exceeded In other words, MST markers for human contamination can be detected when regulations are not exceeded Does simple detection mean there is a problem??? Analytical sensitivity Low level fecal contamination is permitted by U.S. criteria Analytical limits of detection Human-associated marker Research by McLellan (Bower et al AEM 71:8305) detection of HF183 in diluted sewage by PCR 0.2 to 82 CFU/100 ml E. coli (n=14) This research (unpublished data) Detection of qhf183 in sewage by qpcr Detection limit 4 copies/5 ul reaction 78 to 4,800 CFU/100 ml E. coli(geometric mean 480, n=30) 3

4 qpcr begins process of quantification End-point PCR, such as Bacteroidales markers Enterococcus faecium human-associated esp Bac32 General HF134 Human HF183 Human Quantitative PCR (qpcr) provides both presence and abundance data for markers Test sample MST markers do not have absolute host specificity Statistical sensitivity and specificity of markers Presence-absence data (Stoeckel and Harwood 2007, AEM 73:2405) MAR, CUP, ribotyping, PFGE, rep-pcr also included in table Various other markers also included in table Sensitivity and Specificity Quantitative data General and human-associated marker in target and alternate hosts Marker AllBac (n) qhf183 (n) BacHum (n) log (copy per g dry wt) Limit of Detection N 8 to 11 5 to 7 5 to 7 --(LOD varies with level of inhibition, extraction efficiency, and dilution) Human (46) 9.32 (46) 9.51 (46) Bird (11) 10.0 (1) 11.0 (1) Pet (10) 7.67 (9) 8.52 (9) Ruminant (16) 6.65 (7) 6.74 (9) Stoeckel, Stelzer, and Dick. Manuscript in preparation, values subject to change Approach for quantitative evaluation Under current US regulations, contamination is measured using fecal-indicator bacteria For effective use of MST data for management, host-associated marker data must be interpreted in the same terms Governing concept: E. coli feces Marker feces E. coli water Marker water Study area and sites Fountain Creek 10 Below Ruxton Creek 11 At Schryver Park 12 Near Colorado Springs (USGS gage) 13 Below Camp Creek Fountain Creek 14 Above 21st Street SUMMARY OF RESEARCH 1 At Green Mountain Falls 2 Below Crystal Creek 3 Below Wellington Gulch 4 Below Cascade Creek 5 Above French Creek mile below French Ck 7 Above Cavern Gulch 7.9 Above Ruxton Creek 15 Below 8th Street 7.9 Ruxton Creek Above Pikes Peak Cog 8 Railway 9 At mouth Monument Creek 4

5 Sites for 1-year intensive monitoring Routine monitoring Monthly and semi-monthly Project results summary After observed sources of fecal contamination were addressed (broken pipes), E. colistill were above regulatory limits in Fountain Creek E. colidensities were highest (up to 61,000 MPN/100 ml) during the tourism season downstream from Manitou Springs Contrary to expectations, human sources were not implicated as the major causes of exceedances Human, cow, dog, cat, elk excluded as causes Sheep, bear, horse probably not causes Birds could not be excluded as causes MST markers VALIDATION TESTS AllBac a marker present in most fecal sources (nonspecific) qhf183 a marker associated with human sources of fecal contamination BacHum a second marker associated with human sources of fecal contamination BoBac a marker associated with ruminant sources of fecal contamination (cattle, deer, elk, etc). In-house tests at USGS Prepared samples were analyzed blind High degree of accuracy in presence/absence BoBac was detected in sample 1 because it is carried at low concentration in cat fecal material QC Blind 1 QC Blind 2 QC Blind 3 QC Blind 4 Source Cat and human Cattle Horse Human Observed E. coli >24,000 24, qhf183 Detected Not detected Not detected Detected Marker BacHum Detected Not detected Not detected Detected detected BoBac Detected Detected Not detected Not detected E. coli In-house test, quantification QC Blind 1 QC Blind 2 QC Blind 3 QC Blind 4 Observed E. coli >24,000 24, Human Estimated Ruminants 0 42, added to Pets 620, test mixture Other Human 62,000 ND ND 7,900 Calculated Ruminant 67, ,000 ND ND upper limit Pets 1,300, ,000 4,700 3,500 Preliminary data from Stoeckel, Stelzer, Mau, Stogner, and Muzzy in review ND, marker not detected, value presented is the threshold above which marker should have been detected Petsvalues in italicsbecause no pet-associated markers were tested. The value is based on pets carrying human- and ruminant-associated markers Marker Ecoli. feces,0.9 water, source water, adj Markerfeces, adj,0.1 5

6 2007 Sanitary Survey Results Summary Winter flow - Oct 22 Survey - Sept 13 FOUNTAIN CREEK SAMPLES Sample Event Survey - Aug 14 Low flow - Jul 30, Jul 31 Survey - Jul 17 Survey - Jul 5 Snowmelt: May 22, Jun 11 7A Monument Creek Sample Location Ruxton Creek < 126 MPN/100 ml MPN/100 ml > 630 MPN/100 ml Evaluation of Fecal Contamination Sanitary Survey May to September, 2007 E. coli density (MPN/100 ml) 10,000 1, Measurement, Fountain Creek Measurement, Ruxton Creek Upper detection limit Colorado standard Site identifier E. coli density (MPN/100 ml) 100,000 10,000 1, Fecal contamination summary Cool season, Fountain Creek Warm season, Fountain Creek Cool season, Ruxton Creek Warm season, Ruxton Creek Colorado Standard Site Identifier Seasonality Data collected at all sites MST results AllBac general fecal contamination Cool season, Fountain Creek Temp (deg C) Temp E. coli 1.E+06 1.E+04 1.E+02 1.E+00 E. coli density (MPN/100 ml) Log(adjusted AllBac concentration (copy/100 ml) Warm season, Fountain Creek Cool season, Ruxton Creek Warm season, Ruxton Creek Site identifier Sample date 6

7 MST results qhf183 Associated with human contamination MST results BoBac Associated with ruminant contamination Log(adjusted qhf183 concentration (copy/100 ml) Cool season, Fountain Creek 7 Warm season, Fountain Creek Cool season, Ruxton Creek 6 Warm season, Ruxton Creek Site identifier Log(adjusted BoBac concentration (copy/100 ml) Cool season, Fountain Creek 7 Warm season, Fountain Creek Cool season, Ruxton Creek 6 Warm season, Ruxton Creek Site identifier Seasonal synthesis The human element E. coli Lower in cool than warm weather increase gradually upstream to downstream MST markers (general, human- and ruminant associated) about the same in cool as warm weather increase gradually upstream to downstream Nutrients About the same in cool as warm weather Wastewater organic chemicals Present when both E. coliand human-associated MST markers increased Though human-source fecal contamination was not commonly detected, there were specific instances when human contamination was evident. Among samples for which Human-Nonhuman determination was possible 25 of 39 had E. coliabove the 126 MPN/100 ml Colorado water-quality standard (high samples were targeted) 5 of the 25 showed indications that human sources could be the primary contributor of fecal contamination (3 of 10 dates) 2 of the 25 showed indications that human sources could be a minor contributor of fecal contamination May 01, 2008 Fountain Creek below 8 th St (site 15) E. colidensity Above Ruxton Creek (site 7.9) 80 MPN/100 ml Below Ruxton Creek (site 10) 61 MPN/100 ml Below 8 th Street (site 15) 2,400 MPN/100 ml Elevated MST markers for human fecal contamination at site 15 Elevated ammonium at site 15 Wastewater chemicals detected at site 15 (8 of the 10 chemicals detected across 64 samples in this study) May 01, 2008, Ruxton Creek (site 9) E. colidensity 380 MPN/100 ml MST markers elevated (especially Human) Ammonium notably higher than on other dates at this site Wastewater chemicals detected in this sample (8 of the 10 chemicals detected across 64 samples in this study) 7

8 August 12, 2008, upper sites on Fountain Creek (sites 7.9 and 10) E. coli densities Near Green Mountain Falls (site 2) <100 MPN/100 ml Upstream from Ruxton Creek (site 7.9) 460 MPN/100 ml Downstream from Ruxton Creek (site 10) 740 MPN/100 ml Downstream from 8 th Street (site 15) >24,000 MPN/100 ml MST markers Consistent with human sources at sites 7.9 and 10 but not at site 15 No notable increases in nutrient data Wastewater chemicals were not analyzed on this date August26, 2008, above Ruxton Creek (site 7.9) E. coli densities Near Green Mountain Falls (site 2) 62 MPN/100 ml Upstream from Ruxton Creek (site 7.9) 170 MPN/100 ml Downstream from Ruxton Creek (site 10) 730 MPN/100 ml Downstream from 8 th Street (site 15) >2,400 MPN/100 ml MST markers Human-associated markers high at site 7.9 but not 10 Human sources could partially (less than half) explain contamination at site 15 No notable increases in nutrient data Wastewater chemicals were not analyzed on this date What does this mean? Existing management of human, pet, and ruminant in the study area limited fecal contamination from these sources to low levels These sources tended to be detected only when E. coliwas within standards More resources allocated to managing these controllable sources would not have led to Upper Fountain Creek meeting standards Under current conditions, Upper Fountain Creek is not expected to meet regulatory limits during warm weather months Birds are the only tested source that carries E. coli but not MST markers Bird markers are not yet available to confirm birds as causes of E. coli exceedances Pigeons Manitou Springs Rough calculations Based on the concentration of E. coliin Fountain Creek, the flow, and the concentration of E. coli in pigeon feces The combined defecation of 16 to 420 birds directly to Fountain Creek is sufficient to increase E. coliby 1,000 MPN/100 ml Numerous birds were seen in the area Power washing bird droppings from concrete to storm drains was observed Conclusions Fountain Creek continued to exceed E. coli standards in the summer of 2008 Controllable sources (human, pets, cattle) generally were not indicated as major sources of fecal contamination Bird sources, particularly pigeons, may be a source of the detected fecal contamination 8

9 The bottom line Understanding contamination sources, as provided by this study, will help water resource managers to effectively use resources to address measured fecal contamination Though the study identified birds as potential sources (especially pigeons), there is no direct test for bird-source contamination Future steps Continuation of E. colimonitoring would be useful to measure effectiveness of management strategies that might be implemented Continuation of this analysis could be carried downstream to Pueblo and in other major tributaries, such as Monument Creek Though pigeons were implicated in Upper Fountain Creek, other sources may be more relevant in other reaches Summary This type of research can dramatically enhance effectiveness of TMDL and other water management plans 2007 land use survey overwhelmingly pointed to human or other controllable sources of fecal contamination Exceedances during recreational season supported this concept Only through use of quantitative MST and other hosttargeting analysis were humans, pets, cattle and ruminant wildlife excluded as likely sources of fecal contamination in this study Questions? 9