mark s. dortch, phd, pe, d.wre

27
1 Modeling Water and Sediment Contamination of Lake Pontchartrain following Pump-out of Hurricane Katrina Floodwater Mark S. Dortch, PhD, PE, D.WRE Mark S. Dortch, PhD, PE, D.WRE Retired Research Civil Engineer, Retired Research Civil Engineer, Environmental Laboratory, ERDC, USACE Environmental Laboratory, ERDC, USACE Senior Water Resources Engineer, Moffatt & Senior Water Resources Engineer, Moffatt & Nichol Nichol Dortch, M.S., Zakikhani, M., Kim, S.C., and Steevens, J.A. 2008, in J. Env. Mgt, 87, 429-442.

Upload: ethel

Post on 04-Feb-2016

36 views

Category:

Documents


0 download

DESCRIPTION

Modeling Water and Sediment Contamination of Lake Pontchartrain following Pump-out of Hurricane Katrina Floodwater. Mark S. Dortch, PhD, PE, D.WRE Retired Research Civil Engineer, Environmental Laboratory, ERDC, USACE Senior Water Resources Engineer, Moffatt & Nichol. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Mark S. Dortch, PhD, PE, D.WRE

1

Modeling Water and Sediment Contamination of Lake

Pontchartrain following Pump-out of Hurricane Katrina Floodwater

Mark S. Dortch, PhD, PE, D.WREMark S. Dortch, PhD, PE, D.WRERetired Research Civil Engineer, Environmental Retired Research Civil Engineer, Environmental

Laboratory, ERDC, USACELaboratory, ERDC, USACESenior Water Resources Engineer, Moffatt & NicholSenior Water Resources Engineer, Moffatt & Nichol

Dortch, M.S., Zakikhani, M., Kim, S.C., and Steevens, J.A. 2008, in J. Env. Mgt, 87, 429-442.

Page 2: Mark S. Dortch, PhD, PE, D.WRE

2

Background• Interagency Performance Evaluation Task Interagency Performance Evaluation Task

force (IPET) was commissioned by Congress force (IPET) was commissioned by Congress following Hurricane Katrina to evaluate the following Hurricane Katrina to evaluate the reasons and consequences of failed reasons and consequences of failed protection of New Orleansprotection of New Orleans

• IPET Task 9 focused on socio-economic and IPET Task 9 focused on socio-economic and environmental consequencesenvironmental consequences

• A major environmental questions was, what A major environmental questions was, what were the environmental impacts to Lake were the environmental impacts to Lake Pontchartrain as a result of pumping out Pontchartrain as a result of pumping out contaminated flood water from New Orleans contaminated flood water from New Orleans following Hurricanes Katrina and Rita.following Hurricanes Katrina and Rita.

Page 3: Mark S. Dortch, PhD, PE, D.WRE

3

Study Objective

As part of IPET Task 9, model the fate and As part of IPET Task 9, model the fate and transport of contaminants pumped into transport of contaminants pumped into Lake Pontchartrain following the actual Lake Pontchartrain following the actual events to evaluate the environmental events to evaluate the environmental impacts as contrasted against conditions impacts as contrasted against conditions that would have occurred without levee that would have occurred without levee failures and overtoppingfailures and overtopping

Constraint: study had to be done in a few monthsConstraint: study had to be done in a few months

Funding: USACE/IWR to USACE/ERDC-EL

Page 4: Mark S. Dortch, PhD, PE, D.WRE

4

Approach

• A 3D numerical hydrodynamic (CH3D-WES/Z) and A 3D numerical hydrodynamic (CH3D-WES/Z) and water quality model (CE-QUAL-ICM) were appliedwater quality model (CE-QUAL-ICM) were applied

• Modeled constituents included: Modeled constituents included: water column and bottom sediment concentrations of:– Arsenic (As)– Lead (Pb)– Benzo(a)pyrene (BaP)– DDE– Fecal Coliform bacteria (FCB), water concentration only

• Compared actual versus base– Actual – as occurred with levee failures and overtopping– Base – without levee failures and overtopping

• Compared water and sediment concentrations with protective eco and human health criteria

Page 5: Mark S. Dortch, PhD, PE, D.WRE

5

CH3D-WES• 3D, time-varying, baroclinic, free surface 3D, time-varying, baroclinic, free surface

Hydro model, sigma and Z-plane versions Hydro model, sigma and Z-plane versions (used Z-plane version)(used Z-plane version)

• Transports salinity and temperature for Transports salinity and temperature for density coupling and baroclinic forcing (not density coupling and baroclinic forcing (not used in this study due to minor salinity used in this study due to minor salinity differences resulting from the storm)differences resulting from the storm)

• Structured, curvilinear, non-orthogonal, Structured, curvilinear, non-orthogonal, boundary-fitted coordinates, finite difference boundary-fitted coordinates, finite difference methodmethod

• Block grid structure for sub-grid resolution Block grid structure for sub-grid resolution and parallelization (not used in this study)and parallelization (not used in this study)

Page 6: Mark S. Dortch, PhD, PE, D.WRE

6

CE-QUAL-ICM• Multi-dimensional, unstructured, finite volume, Multi-dimensional, unstructured, finite volume,

surface water quality, eutrophication, and surface water quality, eutrophication, and contaminant modelcontaminant model

• Must provide Hydro to ICMMust provide Hydro to ICM

• Developed on Chesapeake Bay (circa 1990), but Developed on Chesapeake Bay (circa 1990), but has been applied to many other systemshas been applied to many other systems

• Over 35 state variables ranging from temperature, Over 35 state variables ranging from temperature, salinity, and nutrients to sea grass and lower food salinity, and nutrients to sea grass and lower food chain, plus fate/transport of trace toxic substanceschain, plus fate/transport of trace toxic substances

• Benthic diagenesis sub-model for predicting bed-Benthic diagenesis sub-model for predicting bed-water column nutrient and DO fluxeswater column nutrient and DO fluxes

• Domain decomposition with MPI for parallelization Domain decomposition with MPI for parallelization (not used for this study)(not used for this study)

Page 7: Mark S. Dortch, PhD, PE, D.WRE

7

Lake Pontchartrain Computational Grid

•6,038 surface cells

•Max of 6 layers, 1.52 m thick except surface

•21,018 total cells

Rigolets Inlet

Page 8: Mark S. Dortch, PhD, PE, D.WRE

8

Pump Stations Included in Model

Orleans Metro Orleans East

Page 9: Mark S. Dortch, PhD, PE, D.WRE

9

Pump Hydrographs for Actual Conditions

5-10% of lake volume pumped out

(2005)

Flows estimated from another IPET task

Started 9/11 and stopped Oct 17, 37 days

Page 10: Mark S. Dortch, PhD, PE, D.WRE

10

Pumped Flow for Base Condition

• Assumed no levee failures or overtoppingAssumed no levee failures or overtopping• Used measured rainfalls at Slidell (Katrina had 20 Used measured rainfalls at Slidell (Katrina had 20

cm) and NO A/P (Rita had 5.84 cm) and NO metro & cm) and NO A/P (Rita had 5.84 cm) and NO metro & east basin area (2.8E8 meast basin area (2.8E8 m22), yielding 5.7E7 and 1.7E7 ), yielding 5.7E7 and 1.7E7 mm33 (Katrina and Rita, respectively) (Katrina and Rita, respectively)

• Base pumped volume Base pumped volume (collected rainfall)(collected rainfall) ≈ 10% ≈ 10% Actual pumped volume for KatrinaActual pumped volume for Katrina

• These volumes can be pumped out in a day or less These volumes can be pumped out in a day or less (0.6 days for Katrina at total pump capacities)(0.6 days for Katrina at total pump capacities)

• Flow per pump = 0.6 * pump capacityFlow per pump = 0.6 * pump capacity• Duration = 1 day due to model requirement of daily Duration = 1 day due to model requirement of daily

flowsflows

Page 11: Mark S. Dortch, PhD, PE, D.WRE

11

Other Hydro Model Inputs

• 90 day simulation starting on Sep 1, 200590 day simulation starting on Sep 1, 2005

• Winds from NO Int. Airport (no data from 9/1 to Winds from NO Int. Airport (no data from 9/1 to 9/7, so linearly ramped up from 0, “spin-up”)9/7, so linearly ramped up from 0, “spin-up”)

• Seaward water level boundary at Rigolets Seaward water level boundary at Rigolets (Waveland, MS gage was blown out) – used sum (Waveland, MS gage was blown out) – used sum of:of:– Astronomical based on NOAA predictions at Waveland, Astronomical based on NOAA predictions at Waveland,

MSMS

– Meteorological based on Norco, Bayou LaBranche, LA Meteorological based on Norco, Bayou LaBranche, LA gage with 48 hour filter and 24 hour shift forwardgage with 48 hour filter and 24 hour shift forward

• Tributary flows were not includedTributary flows were not included

Page 12: Mark S. Dortch, PhD, PE, D.WRE

12

Hydrodynamic Model Calibration

At the Norco tide gage

Hurricane Rita

Model spin-up

Katrina hit on August 29

Page 13: Mark S. Dortch, PhD, PE, D.WRE

13

Hydrodynamic Animation

Sep 9-13, before and during 2 days of pump-out

Page 14: Mark S. Dortch, PhD, PE, D.WRE

14

Contaminant Model Fate Processes

• Equilibrium sorption to suspended solidsEquilibrium sorption to suspended solids

• Settling of adsorbed particulate contam.Settling of adsorbed particulate contam.

• Volatilization of dissolved organic contam.Volatilization of dissolved organic contam.

• Surficial benthic sediment mass balance Surficial benthic sediment mass balance due to settling, resuspension, and burialdue to settling, resuspension, and burial

• FCB die offFCB die off

Page 15: Mark S. Dortch, PhD, PE, D.WRE

15

Contaminant Model Inputs• Lake total suspended solids, TSS = 19.2 mg/L Lake total suspended solids, TSS = 19.2 mg/L

(based on NTU regression and observed lake-wide (based on NTU regression and observed lake-wide median NTU following Katrina)median NTU following Katrina)

• TSS settling rate = 1 m/d based on Corps IHNC TSS settling rate = 1 m/d based on Corps IHNC DMMU studyDMMU study

• Sediment fraction TOC for water column and bed, Sediment fraction TOC for water column and bed, foc = 0.02 based on DMMU studyfoc = 0.02 based on DMMU study

• Surficial benthic sediment porosity = 0.9Surficial benthic sediment porosity = 0.9• Sediment burial rate = 0.026 m/yr based on steady-Sediment burial rate = 0.026 m/yr based on steady-

state solids balance with 0 resuspensionstate solids balance with 0 resuspension• FCB die off rate = 1 per day (conservative value)FCB die off rate = 1 per day (conservative value)• Degradation rate of metals and organics = 0.0Degradation rate of metals and organics = 0.0

Page 16: Mark S. Dortch, PhD, PE, D.WRE

16

Model Inputs (cont.)Kd Input Values for Model, L/kg

• Pb: 4,000 Pb: 4,000 • As: 500As: 500• BaP: 0.3 E6BaP: 0.3 E6• DDE: 0.6 E6DDE: 0.6 E6

Values are close to literature (metals) and Kow Values are close to literature (metals) and Kow computed values (organics), but they were computed values (organics), but they were adjusted slightly using observed benthic adjusted slightly using observed benthic sediment and water column data from NO flood sediment and water column data from NO flood waters to back out Kdwaters to back out Kd

Page 17: Mark S. Dortch, PhD, PE, D.WRE

17

Model Inputs (cont.)Volatilization Rates

• BaP: 0.005 m/dBaP: 0.005 m/d

• DDE: 0.19 m/dDDE: 0.19 m/d

Computed using wind speed of 5 mph, Computed using wind speed of 5 mph, Henry’s constants, and molecular weightsHenry’s constants, and molecular weights

Page 18: Mark S. Dortch, PhD, PE, D.WRE

18

Contaminant Model Loadings (Q*C)

• Actual conditionsActual conditions– Based on measurements in NO floodwaters for Based on measurements in NO floodwaters for

total concentrations (EPA database of values total concentrations (EPA database of values monitored by EPA, USGS, LSU, LaDEQ)monitored by EPA, USGS, LSU, LaDEQ)

– Computed median and 95UCL of concentrations Computed median and 95UCL of concentrations for NO East and NO Metrofor NO East and NO Metro

• Base conditionsBase conditions– Assumed to be same concentrations as Actual Assumed to be same concentrations as Actual

conditions; based on literature – concentrations conditions; based on literature – concentrations in Katrina flood waters were typical for storm in Katrina flood waters were typical for storm waterwater

Page 19: Mark S. Dortch, PhD, PE, D.WRE

Contaminant Loading Concentrations

Constituent Median, ug/L 95UCL, ug/L

Orleans Metro

Arsenic 20 20

BaP 5 5

DDE 0.05 0.05

Lead 5 44

Fecal coliform bacteria 2,200* 70,041*

Orleans East

Arsenic 20 26

BaP 5 5

DDE 0.05 0.38

Lead 2.5 12

Fecal coliform bacteria 200* 32,869*

* MPN/100 ml

Page 20: Mark S. Dortch, PhD, PE, D.WRE

20

Contaminant Model Validation

• There were some observations for FCB and lake There were some observations for FCB and lake benthic sediment contamination following Katrinabenthic sediment contamination following Katrina

• Computed and observed FCB were same order of Computed and observed FCB were same order of magnitude, but observations were spottymagnitude, but observations were spotty

• Sediment contamination values not really Sediment contamination values not really comparable since there is a long-term sediment comparable since there is a long-term sediment memory and model only had a brief loading history memory and model only had a brief loading history compared with many years in prototypecompared with many years in prototype

• There was qualitative confirmation in that events There was qualitative confirmation in that events had very little impact on lake sediment metal had very little impact on lake sediment metal concentrations for predicted and observedconcentrations for predicted and observed

Page 21: Mark S. Dortch, PhD, PE, D.WRE

21

Arsenic Animation, Water Surface

90 days starting Sep 1, 2005

Page 22: Mark S. Dortch, PhD, PE, D.WRE

22

Maximum Incremental Arsenic Water Surface Concentrations

Actual Base

Max of about 0.014 mg/L in both cases

Page 23: Mark S. Dortch, PhD, PE, D.WRE

23

Maximum Incremental Arsenic Sediment Concentrations

Actual Base

Max of about 0.05 mg/kg Max of about 0.005 mg/kg

Page 24: Mark S. Dortch, PhD, PE, D.WRE

Computed maximum incremental water (ug/L) and sediment (mg/kg) concentrations (total) for Lake Pontchartrain for Actual and Baseline conditions and median and 95UCL

loading concentrations

Condition As water

As sed BaP water

BaP sed DDE water

DDE sed Pb wat.

Pb sed

FCB water*

Actual 13 0.048 3.7 0.173 0.036 0.0024 3.7 0.05 1,055

Actual95 16 0.066 3.7 0.173 0.209 0.0171 25.4 0.38 42, 214

Base 14 0.0052 3.7 0.014 0.037 0.000172 3.7 .006 1,413

Base95 14 0.0054 3.7 0.014 0.053 0.000598 32.1 0.05 44 ,780

*Units for FCB are MPN/100mlNote: water concentrations about the same for both conditions, but sediment about 10X for Actual compared to Base

Page 25: Mark S. Dortch, PhD, PE, D.WRE

25

Lake Contaminant Incremental Concentration Conclusions

• Water column – increases were about the same Water column – increases were about the same or less for or less for actualactual versus versus basebase conditions conditions

• Surficial sediment – increases for Surficial sediment – increases for actualactual conditions were about 10 times those for conditions were about 10 times those for basebase

• Increases in sediment for metals were small Increases in sediment for metals were small (<1%) compared with pre-Katrina concentrations(<1%) compared with pre-Katrina concentrations

• Increase in sediment BaP concentrations were Increase in sediment BaP concentrations were about the same as post-Katrina measurementsabout the same as post-Katrina measurements

• Sediment increase in DDE was much lower than Sediment increase in DDE was much lower than other contaminantsother contaminants

Page 26: Mark S. Dortch, PhD, PE, D.WRE

26

Exceedence of Protective Water and Sediment Screening Benchmarks

• Water Pb for Actual95 and Base95Water Pb for Actual95 and Base95

• Sediment BaP for Actual and Actual95Sediment BaP for Actual and Actual95

• Sediment DDE for Actual and Actual95Sediment DDE for Actual and Actual95

• Water FCB for both conditions and both Water FCB for both conditions and both loadings (the norm for NO storm water loadings (the norm for NO storm water removal)removal)

Page 27: Mark S. Dortch, PhD, PE, D.WRE

27

Overall Conclusions

• Flood water contamination is thought to be due Flood water contamination is thought to be due primarily to pre-Katrina-contaminated urban soils primarily to pre-Katrina-contaminated urban soils (typical urban storm-water), but a much greater (typical urban storm-water), but a much greater water volume was pumped with levee failures, thus water volume was pumped with levee failures, thus greater loadingsgreater loadings

• Lake water column concentrations about the same Lake water column concentrations about the same with/without levee failureswith/without levee failures

• Sediment concentrations for BaP and DDE under Sediment concentrations for BaP and DDE under actualactual conditions with levee failures exceeded eco conditions with levee failures exceeded eco screening guidelines close to south shore, but both screening guidelines close to south shore, but both degrade over timedegrade over time

• Water Pb and FCB are a concern with/without levee Water Pb and FCB are a concern with/without levee failuresfailures