Bathing Waters - Measurement and modelling of coastal bathing water quality in Wales: do we need a radical new approach to regulation to deliver resilient coastal communities?
Presentation: Farmer-led climate adaptation - Project launch and overview by ...
Kay david day_3_session_8
1. Bathing Waters
Measurement and modelling of coastal
bathing water quality in Wales: do we need
a radical new approach to regulation to
deliver resilient coastal communities?
Mark Wyer, Carl Stapleton Paul Brewer, Bill
Perkins and David Kay
Session 8
11:00 to 11:15
19th September 2019
Environment Evidence: Marine Evidence (Wales) Conference 2019
6. ≤40 This range is below the
NOAEL in most
epidemiological studies.
<1% GI illness risk
<0.3% AFRI illness risk
41–200 The 200/100 ml value is
above the threshold of
illness transmission
1–<5% GI illness risk
0.3–<1.9% AFRI illness risk
201–500 This range represents a
substantial elevation in
the probability of all
adverse health
outcomes
5–10% GI illness risk
1.9–3.9% AFRI illness risk
>500 Above this level, there
may be a significant risk
of high levels of minor
illness transmission.
>10% GI illness risk
>3.9% AFRI illness rate
The WHO Microbiological Guidelines
9. The
Background
2015 Prize from the UK Water Industry Foundation for the
Highest Impact Research in the Water Sector in the area of
Software Development as outlined in the 2014 UK Research
Assessment Exercise.
10. Where do we need to be?
Real-time prediction of bathing water
(1.5-5.4 billion UK£ and we keep present Blue-Flag numbers)
black box
hydrodynamic
12. Problems (outwith Scotland)
• Model calibration data
– ‘Bathing Day’ is the modelling unit
– Spot compliance samples provide the
calibration data
• Diurnality introduces variation and increase
model error reducing explained variance
• Censored data (< and >) and measurement
imprecision in cfu and/or MPN counts would
further reduce model utility
• Data precision?
13. USEPA Experience
USEPA issues two reports in
2010 a review (2010a) and
modelling (2010b) report.
They reported explained
variances of 20-40% in US
applications of this modelling
approach.
Is this too low for public
health Advice?
14. Solutions
• Characterise the ‘bathing
day’ water quality for model
building
– multiple sampling events
during daylight
• 07:00 to 19:00
– Measure FIOs with enhanced
accuracy through the bathing
day
• Triplicate enumeration / >100+ml
filtered
22. FIO Diurnality
• UV Irradiance is the main predictor
• Temporal pattern examination prudent
• Two sets of comparisons were made:
– 61 days (10/05-28/09/2011), split into 07:00-11:00 and 11:30-
16:00 groups – Student’s t-test
– 24 days (18/07-07/09/2011), split into 07:00-11:00, 11:30-
15:00, and 15:30-19:00 groups - ANOVA
23. Hourly Compliance outcomes – all data
Hour (GMT)
rBWD E. coli
Outcome
rBWD
enterococci
Outcome
rBWD Overall
Outcome
07:00 Sufficient Poor Poor
08:00 Sufficient Poor Poor
09:00 Sufficient Poor Poor
10:00 Good Poor Poor
11:00 Good Sufficient Sufficient
12:00 Good Sufficient Sufficient
13:00 Good Sufficient Sufficient
14:00 Good Good Good
15:00 Sufficient Sufficient Sufficient
16:00 Sufficient Poor Poor
17:00 Sufficient Poor Poor
18:00 Sufficient Poor Poor
19:00 Sufficient Poor Poor
Compliance outcome changes
through the sampling day
24. Model performance tested against
a new data set collected in 2014
‘no deterioration in performance’
25. Cem aes Bay – Water Quality
Modelling
February 2018
Dr Mark Wyer, Professor David Kay
26. Dependent variables (Y)
• Water sampling
60 x 12 hr days
½ hourly sampling – 27 samples per day (inc. AQC)
1620 samples analyzed
1 litre samples
28. Results – enterococci
• Large variation ≈ 2 orders in each day (max: 3.6 orders)
• Elevation in response to event conditions – even
relatively small events
30. Predictor variables (X1 … Xn)
• Velocity x channel cross section area
• Discharge calculation
31. Intestinal enterococci model Type 2/01 60
row daily matrix
Predictor Variable Coefficient
a (Constant) -0.931
X1
Log10 Afon Wygyr Max. Q on sampling day
(m3)
1.296
X2 Log10 Rainfall in previous 24 Hrs.+1 0.551
X3
Mean Wind Sector (16 point) on sampling
day (Rad.)
0.090
X4 ETR in previous 12 Hrs. (MJ/m2) -0.037
X5 Mean Air Temp. on sampling day (˚C) -0.205
X6 Mean Air Temp. in previous 24 Hrs. (˚C) 0.197
Adjusted r2 = 76.3%
32. Intestinal enterococci model Type 2/01
Sign outcome using GM 34 cfu/100 ml threshold:
08:00 – 7.75% Good/92.25% Poor
11:00 – 38.73% Good/61.27% Poor
14:00 – 78.87% Good/21.13% Poor
33. Acclimatize Study Beaches
• 2017 Cemaes Bay, Anglesey
– At the time the only “Poor” beach in Wales
– Acclimatize successfully generated a prediction
model - implemented in 2018
– The model helped to secure a “Sufficient”
classification in 2018 – No “Poor” beaches in
Wales!
34. Acclimatize Study Beaches
• 2018 New Quay North (Dolau), Ceredigion
– One of five sites with a “Sufficient” classification in 2017
– A popular family holiday beach – remained “Sufficient” in
2018
– Field work successfully completed – modelling in progess
35. Acclimatize Study Beaches
• 2018 Traeth Gwyn New Quay, Ceredigion
– One of five sites with a “Sufficient” classification in 2017
– Progressive deterioration from “Excellent” in 2015 –
returned to “Good” in 2018
– Field work successfully completed – modelling in progess
36. Results - Swansea 2011
• Large variation ≈ 1.4 orders in each day
(max: 3.1 orders)
38. Results - Cemaes 2017
• Large variation ≈ 2 orders in each day (max: 3.6 orders)
• Elevation in response to event conditions – even
relatively small events
39. Results – New Quay North 2018
• Large variation ≈ 2 orders in each day (max: 3.9 orders)
• Elevation in response to event conditions
• High concentrations (>1000) on “normal” days
40. Results – Traeth Gwyn
• Large variation ≈ 2 orders in each day (max: 3.4 orders)
• Elevation in response to event conditions
• Occasional high concentrations (>1000 cfu/100 ml) on “normal”
days
41. Results Nolton Haven (2019)
• Note that 12 of 15 samples are at the Lower Limit of Detection
42. Impact with WHO and EU
https://www.who.int/water_sanitation_health/
publications/who-recommendations-to-
european-water-directive/en/