Circulation and Residual Flow in the South Basin of Puget Sound, and its Effects on Near Bottom Dissolved Oxygen

Transcription

1 Circulation and Residual Flow in the South Basin of Puget Sound, and its Effects on Near Bottom Dissolved Oxygen - Albertson, S.L, J. Bos, G. Pelletier, M. Roberts Abstract The south basin of Puget Sound is a complex and interconnected system of straits, open reaches, and fjord-like bays. Marine water exchanges with the main basin of Puget Sound over a sill and through Tacoma Narrows. Within the basin, tidally-averaged net estuarine (residual) circulation can be normal (seaward at the surface, landward at depth), inverse (seaward at depth, landward at the surface), and even sideways (in on one shore, out on the other). This is due to the interaction of its complex morphology with processes such as seasonal variations in river flow, snowmelt, evaporation, and wind. Low near-bottom dissolved oxygen (DO) is observed in some years at the north end of Case Inlet. Residual flow patterns around Hartstene Island in the finger inlets could partially explain this variability in DO. The residual flow alternates between an H pattern (Aura Nova 1998), with circulation splitting around Hartstene Island, and a surface clockwise O (counterclockwise at depth) pattern with flow mostly out through Pickering Passage. We deployed two simultaneous current meters in Dana and Pickering Passage for several months and conclude that the residual flow pattern is controlled by variations in the wind. Introduction The South Puget Sound basin is a complex fjord-like estuary within Washington State in the U.S. Pacific Northwest (Fig. 1) dominated by tidal currents. Its open northern boundary is defined by Tacoma Narrows and an entrance sill located just to the south, a reach of shallow bathymetry left behind by retreating glaciers tens of thousands of years ago. These tidal currents are superimposed on a net estuarine (residual) circulation that is caused by outflow of buoyant fresh water at the surface and inflow of dense (salty) Main Basin water at depth to compensate, that ultimately draws its source from the Pacific Ocean. Interactions with tidal mixing, turbulent flow through constrictions like the Tacoma Narrows, and meteorological forcing (e.g., the wind) regulate this residual flow throughout the year and between years. Compared with the Main Basin, residual flow in South Puget Sound is more sluggish and stratified, resulting in lower annual productivity despite an earlier plankton bloom (Strickland 1983). Surface nutrients, typically nitrate, deplete in the summer (Newton 2002) attenuating blooms despite plentiful sunlight. Such surface depletion suggests sensitivity to further Figure 1. The South Sound nitrogen addition (eutrophication). Advection, dispersion, and Basin. residual circulation are fundamental processes that influence productivity because nutrients that enter a basin from one watershed can affect another area at some distance accelerating plankton growth and ultimately reducing oxygen levels at depth. As the population grows along

2 the eastern shore of Puget Sound understanding these connections will become increasingly important. In addition to eutrophication, the inlets residual circulation may be important for favoring marine population differences, as is the case on a larger spatial scale for Puget Sound vs. the Juan de Fuca Strait (Rynearson and Armbrust 2004). The Washington Department of Ecology (Ecology) created a hydrodynamic model of South Puget Sound (Fig. 2) for multiple purposes, but most importantly to determine vulnerability to eutrophication (Albertson 2002a, b). Low-oxygen conditions have been observed in the region from summer to fall (Bos et al. 2001) and often reach an annual minimum during September because of dying biological components, fading sunlight, low winds and lower tidal energy near the equinox. Figure 2. South Puget Sound model grid with current meter (ADCP) locations indicated. In conjunction with the Lacey Olympia Tumwater Thurston (LOTT) wastewater treatment plant (WWTP) study in (Aura Nova 1998), Ecology first re-occupied the historic hydrographic stations (Collias 1974) in September In addition to finding low near-bottom DO in Budd Inlet, which was well-known and expected, there was also low DO in northern Carr and Case Inlet (Fig. 3). This was a motivating factor for locating the prototype of a new vertical profiling mooring (ORCA) in Carr Inlet (Edwards 2007). The low near-bottom DO in Case Inlet was particularly conspicuous in 1997 both in the data (Fig. 3a) and Ecology s initial model (Fig. 3b). a) b)

3 Figure 3. a) Data, and b) model views of (near-bottom) dissolved oxygen in September, Ecology, in partnership with the University of Washington (UW), established annual September cruises looking for low oxygen for most years beginning in The near-bottom dissolved oxygen levels from (excluding 2002) September cruises were averaged (Fig. 4a) and subtracted from specific years to show anomalies, which exhibit difference phases of a northsouth variation. The most extreme phases observed of this variation occurred in 1999, a southern phase (Fig. 4b) where Budd Inlet DO levels were lower than the mean and in 2004, a northern phase (Fig. 4c) when Carr and Case Inlet DO levels were lower than the mean. The pattern suggests biological oxygen demand moving south (1999) or north (2004) depending on the year, with Hartstene Island in the middle. Figure 4. Synoptic maps of near-bottom DO from cruise data between 1999 and 2006: a) mean values, b) anomaly in 1999, and c) anomaly in Normal residual flow in an estuary goes from the landward side (head) to the seaward side (mouth). The estuarine circulation of South Puget Sound is complicated by its bathymetry, which branches into several finger inlets in its western extreme half. This complex estuary has its head near Shelton during the rainy season and perhaps near the Nisqually Reach (Fig. 2) in the dry season when snow melt is the only option for freshwater input via the Nisqually River. Residual flow around Hartstene Island could also be influenced by the wind. Wind is certainly important elsewhere in Puget Sound as shown at Three Tree Point (Bretschneider 1985), Point Monrow (Matsuura 1997), and summarized recently in (Ebbesmeyer and Cannon 2001). Tens of thousands of surface drift cards were released as part of the LOTT recertification project in (Aura Nova 1998) into Budd Inlet. Recoveries on beaches suggested a bifurcation of net surface flow around Hartstene Island forming an H pattern where water from Budd and Eld Inlets tend to head south around Hartstene Island into Case Inlet via Dana Passage and water

4 from Totten and Hammersley Inlet tends to head north around Hartstene Island into Case Inlet via Pickering Passage (Fig. 5). Figure 5. The H surface flow pattern suggested by the LOTT drift card study. In contrast, Ecology s initial model (Albertson et al 2002) with minimal wind forcing, shows the net flow from a simulated dye release (neutrally buoyant tracer) travels south through Dana Passage (Fig. 6). Details on this model s performance show that residual flow can be among the most difficult results to predict correctly (Albertson 2003). Figure 6. Dye release simulation from Ecology s South Puget Sound initial model. Evidence for residual flow coming out of Pickering Passage has been noted in most years during the September cruises ( ) For example, in 1997 hypersaline water that was denser than

5 the seaward end member was found sinking in Case Inlet at three stations closest to Pickering Passage (Fig. 7). This surprising result indicates that seasonally, evaporation can be important in the finger inlets. Figure 7. Sinking masses of hypersaline water emanating from Pickering Passage in September, Methods To resolve causes for these residual flow patterns around Hartstene Island, twin bottom-mounted RDI Broadband 300-kHz Acoustic Doppler Current Profilers (ADCPs) were deployed midchannel on 21 September 2006 at the two choke points for the flow south (Dana Passage) and north (Pickering Passage) around the island (Fig. 2). Velocity profiles were collected every six minutes for several months by averaging 90-second ensembles of pings taken at a frequency of 1 Hz by Workhorse Sentinel ADCPs, the generic model name for all non-phased array broadband ADCPs currently produced by Teledyne-RDI. A broadband ADCP uses coded pulses to make multiple measurements of phase with a single ping, and thereby greatly increase the precision of the measurement. The deployment period included two snowstorms with strong north wind, a pineapple express with strong SW wind, a major wind event (Hanukah Eve Windstorm of 2006) with hurricane-force SW wind, in addition to the mild late-summer conditions that also have a north-northeast wind pattern. Results The bottom-mounted ADCP measured three components of water velocity in 1-m depth bins with some error estimate from the fourth head; the two most important components of velocity for this study are east-west and north-south. Because of the mounting depth and blanking distance (distance closest to the ADCP where it cannot recover from sending a ping in order to

6 measure reflected pulses), the data taken within 2.5 m of the bottom cannot be recorded. Data within two meters of the surface are lost due to side lobe interference, where reflection of the side lobe signals off the surface interferes with the detection of the returning main acoustic signal. Figure 8. Surface (light colored) and near-bottom (dark colored) flows in Dana (blue) and Pickering Passage (red) during late summer. By calculating a depth of no-net-motion, which was 40m in Dana Passage and 42m in Pickering Passage, a net surface and deep flow are found by averaging across all depth layers in either category (Fig. 8). Results for a week beginning on 29 September show that surface layer flow alternates between two dominant modes. Around 3-4 October the surface flow splits in the H pattern around Hartstene Island with flow out (eastward) in both passages. Around September, however, the surface flow is in at Dana Passage and out at Pickering Passage, indicating a clockwise surface O pattern around the island. During this period the compensatory flow at depth is counterclockwise. During the transition between these modes, a third state exists. Around 1-2 October baroclinic estuarine flow in Dana Passage becomes negligible because all depths are moving in unison during ebb and flood tides. During this time the surface-out bottom-in estuarine flow is most active in Pickering Passage. During this condition and during the O state there is a greater possibility that a biological oxygen demand (e.g., dissolved organic mater) could be advecting along with water masses into northern Case Inlet via Pickering Passage. If these states are being controlled by something that varies between years (e.g., the wind) then variability in that controlling force might explain why dissolved oxygen is lower in some years than in others in northern Case Inlet. Dominant wind directions in the region are from the SW and NNE (Fig. 9).

7 Figure 9. Compass wind rose depicting wind directions from 21 Sept 15 Dec 2006 taken at McChord airforce base in Tacoma, WA. A wind from the SW points to the NE on this graph. By averaging the flows at all depths forward and backward in time by 25 hours, and not splitting them into surface and bottom layers, a de-tided time series for both locations can be derived for the entire deployment period (Fig. 10). The temporal patterns of these flow results align very well with the wind patterns, as the shifts in residual flow direction coincide with distinctive wind events. Depths, recorded by color, are actual distances measured above the instrument so that peak flows are just below the sea surface. The most responsive layers to the wind were near the surface, but slightly deeper in Pickering Passage (34 meters above the ADCP) than in Dana Passage (40 meters above the ADCP). The average depth at both locations was similar at about 44 meters; the tidal range in the area is about 4 meters so that depth bins greater than 40 meters are exposed to the air at low tide. Precipitation provides the buoyancy forcing needed to keep the estuarine circulation active, but the dominant forcing is the wind. Updates to the initial South Sound will verify that buoyancy forcing, when the only source of freshwater is from snowmelt on the Nisqually, is not controlling the residual flow around Hartstene Island as much as wind does.

8 Figure 10. The residual (tidally-averaged) east-west flow at Pickering and Dana Passage; wind and precipitation recorded from McChord air force base. By combining these two analyses, separating flow into surface and bottom components and detiding by forward and backward temporal averaging, a robust wind-driven pattern emerges (Fig. 11a). Further it is also possible to split these flow patterns into two orthogonal components, 1) the H (split) mode and 2) the O clockwise mode (Fig. 11b). The choice of the O mode as clockwise instead of counterclockwise is arbitrary, but it is more appropriate because net surface counterclockwise flow only appears to occur during periods when the split H flow (split around the island) is dominant.

9 Figure 11. The residual (de-tided) surface flow split into two modes, CW around the island O and split H. Potential Implications for DO Wind is affecting the flow around Hartstene Island and this may be important to controlling the intermittent low near-bottom DO noted in the region. Residual flow pattern is alternating between a dominant H pattern, and inactive Dana Passage, and an O pattern with clockwise flow in the surface and a compensatory counterclockwise flow at depth. The effects of buoyancy forcing from snowmelt and the ultimate effect of residual flow on DO in Case Inlet will need to be resolved with Ecology s subsequent model development and dissolved oxygen study in the basin (QAPP In Press). Acknowledgements: NOPP (ONR) grant: N Data collection: R/V Skookum (Brian Grantham, Darrel Anderson, Dale Norton); R/V Barnes (Ray McQuin, Nikki Hix). References Albertson, S.L., K. Erickson, J.A. Newton, G. Pelletier, R.A. Reynolds, and M.L. Roberts, 2002a, South Puget Sound Area Water Quality Study Phase 1, Washington State Department of Ecology Publication No

10 Albertson, S.L., K. Erickson, J.A. Newton, G. Pelletier, R.A. Reynolds, and M.L. Roberts, 2002b, South Puget Sound Area Water Quality Study, Washington State Department of Ecology Publication No Albertson S., N. Larson and J. Newton, A Comparison of Predicted Cross-Channel Flows from an EFDC-based Model to ADCP Data in South Puget Sound. Proc. of the 2003 Georgia Basin/Puget Sound Research Conference, 31 March-3 April 2003, Vancouver, BC, Canada. Aura Nova Consultants, Brown and Caldwell, Evans-Hamilton, J.E. Edinger and Associates, Ecology, and the University of Washington Department of Oceanography, 1998, Budd Inlet Scientific Study Final Report, prepared for the LOTT Partnership, Olympia, WA. Bos, J. K., R.A. Reynolds, J. Newton, and S.A. Albertson Assessing sensitivity to eutrophication of the southern Puget Sound basin: Spatial and seasonal perspectives. Proc. of Puget Sound Research Conference, Seattle, WA. Bretschneider DE, Cannon GA, Holbrook JR, and Pashinski DJ Variability of subtidal current structure in a fjord estuary: Puget Sound, Washington. JGR 90, Collias, Eugene E., N. McGary and C. Barnes, 1974, Atlas of Physical and Chemical Properties of Puget Sound and its Approaches, University of Washington Press, Seattle, 235pp. Ebbesmeyer, C.C. and Cannon, G.A, 2001, Review of Puget Sound Physical Oceanography related to the Triple Junction Region, Report for King County Department of Natural Resources, 34pp. Edwards, K. A., M. Kawase, and C. P. Sarason, 2007: Circulation in Carr Inlet, Puget Sound during spring Submitted to Estuaries and Coasts. Matsuura H and Cannon GA Wind effects on sub-tidal currents in Puget Sound. J. Oceanography (Japan, in English), 53, (Ph.D. thesis at UW) Newton, J., Albertson, S., Van Voorhis, K., Maloy, C. and Siegel, E. "Washington State Marine Water Quality, 1998 through 2000." Ecology Publication: Rynearson, T. A., and E. V. Armbrust Genetic differentiation among populations of the planktonic marine diatom Ditylum brightwellii (Bacillariophyceae). J. Phycology 40, Strickland, R. M The fertile fjord. Puget Sound Books, Washington Sea Grant Publication, Seattle, Washington. 145 pp.