Michael Quinnell Senior Oceanographer 30 th June 2009 1 Background - - PowerPoint PPT Presentation
Michael Quinnell Senior Oceanographer 30 th June 2009 1 Background to author Introduction to review undertaken Overview of numerical models 5 identified points of concern Summary Any questions Degree in Ocean Sciences
1
Background to author Introduction to review undertaken Overview of numerical models 5 identified points of concern Summary Any questions Degree in Ocean Sciences Post-graduate qualification in Coastal Engineering 12yrs experience in commercial oceanography and marine
geophysics
CMarSci, MIMarEST
2
Review of Sewerage Scheme EIS Statements, 2006 & 2008, ‘Outfall
Model Report’ sub-section
Review comments report submitted Dec 2008 EIS presents a detailed investigation into a wide range of environmental
impacts
Concern identified in 5 key study areas of Outfall Model Report affecting
validity of predictions of effluent dispersion
Several numerical models used in EIS to simulate effluent dispersion
3
A computer program which gives a report of predicted predicted conditions, e.g. a conditions, e.g. a weather forecast weather forecast
Effluent dispersion models
Output: map of effluent dispersion
Inputs: wind, current, effluent concentration, effluent volume
Inputs and model therefore influence outputs
Output of one model can be the input to another
IMPORTANT: Accuracy of model output
is linked to accuracy of model inputs inputs… …
INPUT (wind, temperature…) NUMERICAL MODEL (software program) OUTPUT (weather forecast)
4
Average flow volumes v. Maximum flow volumes
2008 report models used predicted average daily outflow volume from
pipe (i.e. not max volume)
Models using average daily volume allow prediction during average pipe
It would be more appropriate to model average & maximum outflow
conditions
This could be more informative to determine if effluent would disperse
5
2006 report undertook (at DCC’s request) modeling of effluent
dispersion of plant failure event (i.e. maximum concentration of untreated effluent)
2008 report omitted this modeling of plant failure event How untreated effluent disperses during plant failure event not modeled
therefore not known if would effluent disperse onto bathing beaches if plant fails
6
Remember: accuracy of output of a dispersion model is limited by accuracy of input… Conclusions:
EIS modeled average outflow volume Therefore we cannot determine the impacts on beach, bathing area and
shellfish area etc when above average outflow is experienced Recommendations:
Effluent dispersion modeling and particle tracking modeling should be
re-run using maximum outflow volume and maximum concentration events (i.e. plant failure) and results re-assessed
7
How often is acceptable for effluent concentrations at a location to be surpassed?:
extreme conditions
e.g. currents, winds, rainfall, river outflow; plant operating efficiency, failure frequency, effluent volumes, concentrations
25yrs? Not stated in report
8
How frequently effluent concentrations at a location are exceeded is
important in determining acceptability of such extreme conditions
E.g. Qu. On how many occasions in a year is effluent disperse close to or
Conclusions:
Return period of dispersion model results presented not known (to the
reader)
Plant design return period for extreme events not known
Recommendations:
Return period of dispersion model results presented and plant design
return period for extreme events should be reported
9
1 (DWF) x 8800 (PE) x 0.225 (PE) = 1980m3/day 1.5 (DWF) x 8800 (PE) x 0.225 (PE) = 2970m3/day 3 (DWF) x 8800 (PE) x 0.225 (PE) = 5940m3/day
10
3DWF value was used in effluent dispersion and particle tracking models.
If calculation is incorrect this represents under-estimation of 17% of projected flow
Remember: accuracy of output of a dispersion model is limited by accuracy
Conclusion:
Methodology of calculating 3DWF value is questioned
Recommendation:
Accuracy of calculation of 3DWF value should be independently confirmed
11
Recommended values of coefficient based on water clarity. Outfall location categorized by general description - EIS report quotes
coefficient used close to that recommended for “coastal waters in the UK”
Waters in Lough Foyle are shallow, surrounded by mud flats, frequently
mod to high winds and waves. Dye detection survey report states “very high” concentrations
Therefore not considered representative of typical “coastal waters in the
UK” >> a more conservative coefficient should be used
90 coefficient used for bacterial
12
T90 coefficient key in predicting extents of dispersion of effluent. If
coefficient in model incorrect then effluent spreads further at high concentrations than presently predicted Remember: accuracy of output of a dispersion model is limited by accuracy
Conclusion:
Accuracy of T90 coefficient for Lough is questioned
Recommendation :
Accuracy of T90 coefficient should be independently confirmed Measured during current survey
13
Is current data input to effluent dispersion model accurate?:
Current flow directly affects where effluent is moved to
>>directly affects accuracy of final effluent dispersion model
Methodology:
(i) Obtain suitable current flow model (ii) Undertake current measurements (iii) Validate model by comparison and adjustment of results with current flow model (iv) Input current flow model results to effluent dispersion model (v) Obtain predicted movement of effluent
14
EIS results of comparison reported “very good correlation” between
measured and modeled current direction and duration data, and “good correlation” between measured and modeled current speed data during spring tides for all locations
Current flow model therefore not adjusted and input to final effluent
dispersion model
Author believes poor correlation between measured data and model
data
Author believes quality of current measurement program is poor
(data quality, sampling rate, profiling)
15
Tidal duration Current speed Current direction
16
Tidal duration Current speed Current direction
17
Tidal duration Current speed Current direction
18
Tidal duration Current speed Current direction
19
The Author’s beliefs are confirmed with comments in drogue tracking and dye
tracing survey reports: “…some of the flood releases had to be curtailed because the drogues were heading shorewards…” “…Once the dye reached Moville the patch elongated significantly and followed a more inshore track towards Carrickarory Harbour …” “…It was not possible to define the inshore boundary of the patch during the sixth run as the dye was in very shallow water adjacent to Hazel’s Bridge …” “…had a significant onshore component, such that the inshore boundary of the patch was in contact with the shore before the dye had reached Greencastle…” “…necessary to recover the drogue twice during this survey, firstly because it moved into very shallow water, and secondly as it was in the entrance to the
Harbour …”
20
Measured Modeled Flood spring tide current directions
21
Accuracy of current data input to effluent dispersion model key to predicting
extents of effluent dispersion. If current data input incorrect then effluent spread differently to presently predicted
Measured data shows currents with increased shoreward setting
component
Accuracy of current flow model data input to dispersion model is
questioned
22
Recommendations:
Current survey is repeated for higher accuracy, increased sampling rate:
acoustic Doppler profiling current meter instruments (ADCP)
Current flow model re-validated with new current survey data Effluent dispersion model re-run Consideration given to a 3D model to better model changes in current
profile
23
Observe effects of extreme conditions upon environment (max outflow
volume, max concentration events, include plant failure)
Return period of modeled results should be confirmed and reviewed Methodology of calculation of 3DWF value should be confirmed Accuracy of T90 bacteria decay coefficient should be reviewed, and
turbidity measurements made to confirm
Current survey should be repeated using higher accuracy ADCP
equipment; Then current flow model re-validated; Then effluent dispersion model re-run; Consideration given to using 3D hydrodynamic model
24
25