Susquehanna River

Chester River

ChoptankRiver

PatuxentRiver

Patapsco River

Flood DepositThickness (cm)

0 1 2 3 4

0 9 18

Kilometers

Poole’sIsland

Sediment deposition from Tropical Storm Lee in the upper Chesapeake Bay:�eld observations and model predictions

Cindy Palinkas1, Je� Halka2, Ming Li1, Larry Sanford1, Peng Cheng11Horn Point Laboratory

University of Maryland Center for Environmental Science Cambridge, MD, USA

cpalinkas@hpl.umces.eduEP33B-0861

2Maryland Geological SurveyMaryland Department of Natural Resources

Baltimore, MD, USA

Susquehanna River

Chester River

Choptank River

Cruise 2 (24 Oct)Cruise 1 (28 Sept)

0 7 14

Kilometers

87

654

3

210

6.5

0.5

S1

S2

S3

S4

2.5

!!

!

!

!!

!

!

!

!

!!!

"

"

"

"

"

"

"

#

#

#GoosesReefBuoy

ThomasPt

Light

Conowingo Dam

-200

20

Wind (m/s)

0

1

2 x 104

SusquehannaRiver �ow (m3/s)

0

5

10

15 x 105

Susquehannasediment load (t)

0

1

2

3Signi�cant

wave height (m)

0

2

4

6 Mean waveperiod (s)

08/21 08/28 09/04 09/11 09/180

20

40

6080

Turbidity(NTU)

Day in 2011

(top to bottom):Winds at Thomas Pt Light: Irene: as high as 25 m/s directly down the axis of the Bay for several hours Lee: indistinguishable from non-storm conditions.

Susquehanna River: Flow: elevated during Hurricane Irene, but dwarfed by �ow during and after TS Lee Sediment loads (estimated with the model): peaked at about 5.5 t/day on 9 Sept 2011 in response to TS Lee, corresponding to an estimated suspended-sediment concentration (SSC) of almost 3 g/l.

Surface waves at the Gooses Reef CBIBS buoy in mid-Bay: Heights: up to 3 m during Hurricane Irene; about 0.5 m during TS Lee Wave periods: up to 6 sec during Hurricane Irene, about 2.5 sec during Lee

Surface turbidity at Gooses Reef (no local SSC data were available) Irene: spiked sharply, in phase with the highest wave forcing Lee: lagged behind the peak in Susquehanna River loads by almost 4 days, but was approximately twice as high as during Irene and remained elevated for several days, corresponding to the sediment plume passing the Gooses Reef site.

Field Methods and Observations

Two rapid-response cruises: 28 September 2011, ~3 weeks after peak discharge of the Susquehanna River, focused on the upper Bay24 October 2011, extending observations northward and southward of the initial stations.

16 gravity cores, analyzed for �ood-sediment signatures: Physical strati�cation in x-radiographs Finer grain size Relatively uniform 7Be (half-life 53.3 d) activities

Compare to previous observations of �ood sedimentation following TS Agnes in 1972 (Zabawa and Schubel, 1974) and grain-size measurements of Kerhin et al. (1988)

Lee7Kherin etal., 1988

0

1

2

3

4

5

6

0.1 1.0 10.0Total 7Be Activity (dpm/g)

Dep

th (c

m)

Median Diameter (microns)

0

1

2

3

4

5

6

0 10 20 30 40

LeeS3Kherin etal., 1988

0

1

2

3

4

5

6

0 10 20 30 40

ND

0

1

2

3

4

5

6

0.1 1.0 10.0

Dep

th (c

m)

ND

ND5

10

15

20

25

10 cm?

X-radiograph

0

5

10

15

20

25

03 cm

Spatially interpolated map of �ood-deposit thickness, using maximum 7Be penetration depth as a proxy. A depocenter is apparent at Lee6 and Lee7, south of which deposition decreases rapidly, resulting in a thin drape of �ood sediment in most of the upper Bay. Little to no �ood sediment was likely deposited south of LeeS2.

A deposit mass can be estimated using this map by assuming a dry bulk density of 0.5 g/cm3, assuming a grain density of 2.65 g/cm3 and water content of ~60%. The resulting calculated deposit mass is 4.9 x 106 t, which is ~70% of sediment delivery (estimated by the model). Note that recently delivered �ood sediment may have higher water content; this would result in a lower dry bulk density value and a lower deposit-mass estimate.

Coupled Hydrodynamic-Sediment-Transport Model Description (details in Cheng et al., submitted):

Hydrodynamic model: based on ROMS (Regional Ocean Modeling System), with 240 x 160 horizontal grids and 20 vertical layers Forced by freshwater in�ows at river heads, tidal and non-tidal �ows at the o�shore boundary, and winds and heat exchanges across the water surface Run from 1 Sept 2011 to 30 Jun 2012, initiated with outputs from a hindcast simulation from 1 Jan 2010 to 31 Aug 2011

Sediment-transport model: based on Warner et al. (2008) and simulates sediment erosion, suspension, transport, and deposition. Fluvial sediment: 40% clay, 50% silt, 10% sand Seabed: simpli�ed and initialized with uniformly distributed silt (grain size of 0.022 mm). At high SSC, the contribution of suspended sediment to water density is included by treating the water as a water-sediment mixture. Surface waves: not included in the model. The model is only used for TS Lee; wave-induced stress estimates (see above) derived from observations are used for Hurricane Irene. Susquehanna River discharge: estimated with a regression between SSC (mg/l) and river discharge (R, m3/s) at Conowingo, MD, using USGS data collected between Jan 1978 and Jul 2011

High SSC reached from the Susquehanna River to the mouth of the Potomac River on 13 Sept, corresponding to the sediment plume observed in satellite images.

Spatial distribution of the modeled �ood deposit on 28 Sept 2011: Major deposition of sand near the Susquehanna River mouth A relatively thin deposit of �ne sediment in the rest of the upper Bay, with cross-estuary variability. Most �ne sediment in the upper Bay near the eastern shore; maximum thickness (3.5 cm) above the Chester River mouth Deposition of �ne sediment in the mainstem Bay constricted near the western shore, with a thickness of 1.5 cm.

Predicted change in deposition from 28 Sept-24 Oct: Strong tidal currents just south of the Susquehanna caused bed erosion and net loss of ~1 cm Most sediment was deposited downstream of the narrow channel, with a net gain of <0.5 cm Smaller amounts were transported downstream and deposited in the western half of the Bay

Estimated wave-induced bottom stresses at the latitude of LeeS3 at the peak of Hurricane Irene were much higher than typical tidal bottom stresses (surface waves assumed to be 2.5 m high, with 5.5 s period): Depths <7 m: wave stresses more than an order of magnitude higher than critical At LeeS3 (about 15 m): wave-induced bottom stresses were ~0.6 Pa, still signi�cantly higher than estimated critical stresses Thalweg (26 m): wave-induced bottom stresses equaled the critical stress

These stresses have a strong gradient from extreme erosion in shallow waters to low erosion or deposition in the deepest waters.

Approximately 30% of the cross-section was likely in a surf zone, corresponding closely to the regions identi�ed as sandy in a previous comprehensive bottom-sediment mapping e�ort for northern Chesapeake Bay (Kerhin et al., 1988).

AbstractEpisodic �ood and storm events are important drivers of sediment dynamics in estuarine and marine environments. Event-driven sedimentation has been well-documented by �eld and modeling studies. Yet, few studies have integrated �eld observations and modeling results to overcome the limitations inherent in both techniques. A unique opportunity to integrate �eld observations and model results was provided in late August/early September 2011 with the passage of Hurricane Irene and the remnants of Tropical Storm Lee in the Chesapeake Bay region. These storms di�ered in their timing, track, and impact on the Bay region – Hurricane Irene was primarily a wind/resuspension event, whereas TS Lee was a hydrological/deposition event, with the second largest discharge of the Susquehanna River on record. Because these two storms occurred within a relatively short period of time, both are potentially represented in the sediment record obtained during rapid-response cruises in September and October 2011. The resulting sediment deposit was recognized in cores using classic �ood-sediment signatures (�ne grain size, uniform 7Be activity, physical strati�cation in x-radiographs) and was found to be <4 cm, thickest in the upper Bay. Model runs, conducted only for TS Lee-related sedimentation, generally agreed with these estimates. Integration of the two techniques greatly improves understanding of the transport and fate of �ood sediment in the Chesapeake Bay.

Goals/Objectives

Goal: integrate �eld observations and model simulations of event-driven sedimentation in Chesapeake Bay. A unique opportunity to accomplish this goal was provided in late August/early September 2011 with the passage of Hurricane Irene and the remnants of Tropical Storm (TS hereafter) Lee. The primary focus of this study is on TS Lee, not only because of the large amounts of sediment deposition expected and concern about associated potential ecological damage like that following TS Agnes in 1972, but also because the model does not simulate waves and hence will underestimate bottom stress and resuspension during Hurricane Irene.

Objectives: 1) determine the thickness and extent of sediment deposited in upper Chesapeake Bay following Hurricane Irene and TS Lee with �eld observations and model simulations, and 2) integrate �eld and model techniques to improve understanding of �ood-sediment transport and fate in the Bay.

Hurricane Irene and Tropical Storm LeeHurricane Irene (27-28 August 2011):Tracked east of the Bay, with maximum rainfall on the eastern shore, where tributaries typically have small watersheds and deliver little �ne sediment

River discharge of the Susquehanna River and other tributaries on the Bay’s western shore increased but was well below �ood thresholds. Also, these tributaries (except the Susquehanna) discharge into tributary estuaries before entering the Bay, which serve as e�ective sediment traps, even during high-�ow events

Satellite images taken after the storm show that most turbidity is con�ned within tributaries.

Hurricane IreneTropical Storm Lee (7-11 September 2011):Heavy precipitation focused largely on the Susquehanna River watershed, resulting in the second (after TS Agnes in 1972) highest recorded Susquehanna River discharge.

Discharge exceeded the predicted scour threshold for sediments behind the Conowingo Dam (11,320 m3/s; 400,000 ft3/s) for 3 days, scouring ~4 x 106 t of sediment that was subsequently transported into the upper Bay, along with newly eroded watershed sediment.

Total sediment load delivered to the Bay, estimated by the model, was ~6.6 x 106 t, ~3 times greater than the annual load of the Susquehanna.

Satellite images taken after the storm show a dramatic sediment plume in Chesapeake Bay. Tropical Storm Lee

Response of Chesapeake Bay to Hurricane Irene and TS Lee: Wave-induced bottom-stress estimates (see Sanford, 1994): Near-bottom velocity �uctuations: estimated from signi�cant wave height, wave period, and water depth using linear wave theoryCharacteristic water depths: evaluated along an east-west lateral transect across the Bay at core location LeeS3 Wave-bottom friction coe�cients: estimated using a numerical approximation of the Jonsson wave friction factor diagram (e.g., Madsen, 1976), allowing for the complete range of wave, roughness, and Reynolds number dependence of the drag coe�cient.Estimates assumed an equivalent bottom roughness length of 0.1 mm appropriate for silt-dominated bottom sediments, a signi�cant wave height of 2.5 m, and a wave period of 5.5 s, based on peak waves observed at Gooses Reef. Because signi�cant portions of the lateral transect were likely in a surf zone, wave heights were limited to 0.78 times the water depth for the purposes of stress estimation. Estimates of the critical stresses for erosion of mid-Bay silts range from 0.12 Pa at the sediment surface to 0.3 Pa at ~1 mm depth (Sanford, 2006)

The model estimates that, as of 28 Sept, nearly 6.73x 106 t of sediment deposited inside the Bay, 0.15 x106 t remained in the water column, and only 307 t of �uvial sediment escaped from the Bay.

0

2

4

6

8

0 5 10 15 20 25-30

-20

-10

0

10

Lateral distance (km)

velocity (m/s)stress (Pa)critical stress (Pa)

Depth (m)LeeS3Surf zoneSurface

Comparison of Field Observations and Model Simulations There is a general agreement on the deposition thickness between the 7Be and model estimates, with best agreement near the Susquehanna River, where both show a relatively thick depocenter. They also agree in the southern portion of the study area, with both suggesting little to no sediment deposited south of the Choptank River, lending further support to our interpretation of the physical strati�cation in core LeeS3.

The largest discrepancy occurs near Poole’s Island, where the model predicts ~3 cm of deposition but no 7Be was detected. This may be due to a combination of shallow water depths, a constricted channel, and strong wave forcing following TS Lee. A reasonably strong wind-wave event occurred on 16 Sept 2011, after TS Lee but before the �rst coring cruise. Sanford (1994) showed that even moderate wind events can generate large wave-forced sediment resuspension near Poole’s Island, and the recent storm deposits would have been unconsolidated and easy to erode. The present model might not have predicted as much resuspension at this location because it did not include wave forcing.

The �eld and model observations yield similar deposit geometry and mass. This is somewhat unexpected, given that the �eld-based deposit mass is a rough, order-of-magnitude estimate for several reasons. First, the areas used in the calculation extend to the shoreline in most cases and all boundaries were somewhat arbitrary, chosen to exclude potential transport into tributaries. Also, no sediment was assumed to deposit from S3 southward, based on the absence of 7Be or any other clear �ood-sediment signatures. In reality, it is unlikely that �ood sediment was deposited in shallow (less than a few meters or so) areas and some sediment may have entered tributaries. It is also likely that some sediment was transported southward but cannot be resolved using our analysis techniques. Finally, the 1-cm sampling increments would overestimate a “dusting” of sediment.

The deposition pattern observed following TS Lee is similar to that observed following TS Agnes in 1972 (Zabawa and Schubel, 1974), which discharged 31 x 106 t of suspended sediments into Chesapeake Bay (Schubel, 1976). After TS Agnes, a 20-30-cm thick sediment deposit was observed in x-radiographs in upper Chesapeake Bay, with the thickest deposits located near our Lee7. Thus, the river discharge and sediment-deposit thickness associated with TS Agnes was ~4 times that associated with TS Lee.

Grain size within the top 6 cm of each core was relatively uniform, with some downward coarsening like that observed at Lee7 (upper panel).

7Be penetration depths ranged from 0 cm (not detectable) to 4 cm; these depths were greatest in the northern Bay, decreasing southward. 7Be was not detected north of Lee7 or south of Lee S2.

Flood-deposit thicknesses derived from the x-radiographs generally agreed with the maximum penetration depth of 7Be, although the 7Be depths were often 1 cm greater. The largest discrepancy occurred at Lee S3, where 7Be was not detected but a thick (~10 cm) unit of physically strati�ed sediments was observed in the x-radiograph (lower panel). This is probably due to erosion, transport, and redeposition of sediments during Hurricane Irene, creating physical strati�cation that is not associated with TS Lee-related �ood-sediment deposition. It is likely that this storm layer represents accumulation of muds transported o�shore during or after the storm, though the precise mechanism is unclear.