WATERMAN - Coastal Water Quality Forecast and ...

Joseph Hun-wei Lee

Hong Kong University of Science and

Technology

Workshop on Smart Urban Water

Systems, HKUST, 12 February 2015

WATERMAN - Coastal Water Quality

Forecast and Management System

Outline

1. Introduction to WATERMAN system

2. Challenges of daily prediction of beach

water quality

3. Daily forecasting of beach water quality

4. WATERMAN operational experience

5. Conclusions

Hong Kong Special Administrative Region, China

Hong Kong Zhuhai

Macao Bridge Hong Kong

Zhuhai

Macao

Pearl

River

Estuary

Shenzhen

Harbour Area Treatment Scheme (HATS)Screening

Plants/pumping

stations

Stonecutters Island

Sewage Treatment

Work (SCISTW)Submarine

outfall

23.6 km deep tunnels

(>100m below ground level)

香港島

九龍

Stonecutters

Island STW Chemically Enhanced Primary Treatment (CEPT) since 2001; 23.6 km of deep tunnels; disinfection since March 2010

Stage 1: Q = 1.4 x 106 m3/d

Stage 2A: Q = 1.8 x 106 m3/d

Environmental discharges in the form of buoyant jets

Single buoyant jet

Dense jets Sediment jets

Rosette jet group

Near field + Intermediate field Far field

Length: km

Time: hr

Length:10-500m

Time: min

Side view

Plan view

Mixing and Transport Processes

Multiple length & time scales

Integral jet model

Experiments

CFD

Shallow water

circulation model

Particle tracking

Challenges of Coastal Water Quality Management

• Coastal water quality prediction is a challenging multi-

scale and multi-disciplinary problem

• Pollution sources located in close proximity to

sensitive receivers

• Dynamic marine environment with complex currents;

highly nonlinear biochemical process

• Water quality data are typically sparse and costly

• Uncertainty in rate coefficients, loading and

boundary conditions

• Need for public accountability and public engagement

Data-driven models

Hydrological model

Field studies

Hydrodynamic model

Meteorological data

Data assimilation

Project WATERMAN海灘水質預報系統的研發

Beachwater quality

forecast

Past data

Laboratory studies

Hong Kong’s beach grading system

香港海灘水質評級系統

Grading

Beach

water quality

泳灘水質

E. coli *

(counts

/100 mL)

大腸桿菌

Minor illnesses

rate **

(cases per 1000

swimmers)

發病率

Water Quality

Objective

Compliance/

Exceedance

1 Good ≤ 24 UndetectableCompliance

2 Fair 25 - 180 ≤ 10

3 Poor 181 - 610 11 - 15Exceedance

4 Very poor > 610 > 15

*Weekly Beach Grading: G. Mean E. coli level of the 5 most recent data (ClnEC5)Annual Beach Ranking: G. Mean E. coli level of all bathing season data (Mar-Oct)

** Skin and Gastrointestinal illnesses (Cheung et al. 1990)

Beach grading based purely on past sparsely sampled data

Volleyball

41 beaches ~10 million visitors in 2012

June-July 2007

The need for a better beach WQ management system

• Resource-intensive for sampling and analysis of E.coli data

(41 beaches x 4-5 times/month x 8 months per year)

• At least 24 hr to obtain E. coli measurement results – delayed

response for pollution events

• Represents the average WQ over the past 1 month - cannot capture

the dynamic beach E. coli variation; no forecast ability!

Environmental factors affecting beach water quality

影響水質的環境因素

Big Wave Bay 大浪灣 Lido Beach 麗都

Main factor

Main factor

Solar radiation Rainfall Tide level Wind speed & direction

Past E. coli level

Water temp.

Salinity

Relationship between E. coli level and environmental factors (2002-2006)

相關大腸桿菌含量和環境因素的相關

降雨 太陽幅射 鹽度

潮汐 向岸風 過去大腸桿菌含量

相關系數

Studies on E. coli decay rate 大腸桿菌衰亡率研究 (實驗和野外研究)

Laboratory studies

F

A B 4

32

1

Bridge

Field studies (during storms)

Sunlight intensitySalinity Water temp.

鹽度 水溫 光照度Decay rate

Bacterial decay modeling

Bacterial Loading Current, Turbulent Mixing潮汐流、紊流混合

E. coli concentration(大腸桿菌含量)

Far-field Hydrodynamic

Model

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

20.0 25.0 30.0

Salinity (ppt)

Dep

th (

m)

Neap

Spring

Near-field modelJETLAG

Submarine outfalls

Dynamic coupling(DESA)

Project WATERMAN3D Hydrodynamic Model of Hong Kong waters(Water Research, Chan et al. 2013)

Input for deterministic forecast

Pearl River flow珠江流量(ANN模型)

- Estimated using rainfall with ANN model

BoundaryTide 邊界潮位

Wind 風速、風向-Waglan daily mean(except for Tsuen Wan)

Boundary Salinity 鹽度 Bacterial decay rate細菌衰亡率

Solar radiation 太陽輻射

- Estimated from Predicted UVI

Water Temp. 水溫

- Meas. North Point

0

0.5

1

1.5

2

2.5

3

3.5

0 6 12 18 24

MJ/m2/h

Time (hr)

Sunny,

UV=13

Mild, UV=7

Overcast,

UV=2

Rainfall induced E.coli loading降雨引發的細菌污染- Empirical correction using previous 3-

day rainfall

-15 -10 -5 0 5 10 15(m)

Jet trajectories

Depth = 10m

Sigma Layers

E.coli

0.E+00 5.E+05 1.E+06

cnt/100mL

-0.5 -0.25 0 0.25 0.5

10 20 30

Velocity

Salinity

(m/s)

Tidal current

Velocity/salinity 3D jet trajectories Computed E.coli profile

by far field model

Near-far field coupling

Longitudinal transect of computed E.coli concentration field

HATS

Tsuen Wan

Update (Hindcast)for yesterday

Initial condition for Today

Forecast for Today

Beach Water Quality indices

(11 am)

昨天實測的氣象數據Measured weather conditions yesterday (4 am)-Wind-Solar radiation-River discharge (ANN model)

今天預測的氣象數據Predicted weather conditions for today-Wind, flow yesterday-Predicted solar radiation from forecast UVI (6am)

Daily Deterministic Beach Water Quality Forecast3D模型水質預報流程

Empirical correction for rainfall induced load

更新昨天預測

今天的初始值

今天的水質預測

預報水質等級

GIS-based 3D Visualization System

GEM

10

100

1000

10000

Feb-1 Feb-6 Feb-11 Feb-16 Feb-21 Feb-26

E.c

oli

(#

/10

0m

L) Predicted

Measured

HATS

Disinfection Suspended

(污水消毒暫停)Gemini Beach

雙仙灣

大腸桿菌含量預測

實測

610

180

Beach WQ improvement

Dynamic variation of E.coli, Gemini Beach, Feb 2010雙仙灣海灘的大腸桿菌含量的變化

E.Coli exceeds standard

16 Jun, 2011Overcast with rain 密雲有雨Solar Rad. = 5.0 MJ/m2/dHigh water to ebb tide 高潮

6 Jul, 2011Sunny day 晴天Solar Rad. = 28.3 MJ/m2/dFlood tide 漲潮

24

180

610

0.0

0.5

1.0

1.5

2.0

2.5

3.0

1

10

100

1000

10000

0:00 6:00 12:00 18:00 0:00

Tide (m)E.coli (#/100mL) GEM

Prediction

Field data

Tide

0

100

200

300

400

500

600

700

800

900

1000

0:00 6:00 12:00 18:00 0:00

Solar Rad. (W/m2)

24

180

610

0.0

0.5

1.0

1.5

2.0

2.5

3.0

1

10

100

1000

10000

0:00 6:00 12:00 18:00 0:00

Tide (m)E.coli (#/100mL) GEM

Prediction

Field data

Tide

0

100

200

300

400

500

600

700

800

900

1000

0:00 6:00 12:00 18:00 0:00

Solar Rad. (W/m2)

Model – data comparison, Gemini Beach 大腸桿菌 大腸桿菌

太陽輻射 太陽輻射

Forecast Accuracy* in 2010-201220 Key Beaches

82-95%

83-98%

85-93%

* Accuracy for forecasting the

compliance/exceedance of beach water

quality objective (180 cnt/100mL)

(N ~ 100 per beach)

74-93% (5 open beaches)79-90% (by 3D model)

總結

• WATERMAN海灘水質預報系統應用到水文、水動力及水質等先進數學模型，和每天的大量水文氣象數據同化，利用考察和實驗數據對模型預測進行大量科學驗証。

• 2011-2013年，WATERMAN系統在預報泳灘水質達標／超標的整體準確率高達74-98%。

• WATERMAN系統已被用於研究優化HATS消毒劑的用量，和應變突發性的海灘污染事件（如污水管破裂）。

• WATERMAN海灘水質預報系統是國際上最先進的同類系統之一，將是未來香港邁向智慧城市不可或缺的一部分。