Oceanography of the Pacific Northwest Coastal Ocean and Estuaries with Application to Coastal Ecosystems

Barbara M. Hickey1 and Neil S. Banas

School of Oceanography Box 355351 University of Washington, Seattle, WA 98195-7940 Tel.: 206 543 4737 e-mail: [email protected]

Submitted to Estuaries May 2002

1Corresponding author. Hickey and Banas 2

Abstract

This paper reviews and synthesizes recent results on both the coastal zone of the U.S. Pacific Northwest

(PNW) and several of its estuaries, as well as presenting new data from the PNCERS program on links between the inner shelf and the estuaries, and smaller-scale estuarine processes. In general ocean processes are large-scale on this coast: this is true of both seasonal variations and event-scale upwelling-downwelling fluctuations, which are highly energetic. Upwelling supplies most of the nutrients available for production, although the intensity of upwelling increases southward while primary production is higher in the north, off the Washington coast. This discrepancy is attributable to mesoscale features: variations in shelf width and shape, submarine canyons, and the Columbia River plume. These and other mesoscale features (banks, the Juan de Fuca eddy) are important as well in transport and retention of planktonic larvae and harmful algae blooms.

The coastal-plain estuaries, with the exception of the Columbia River, are relatively small, with large tidal forcing and highly seasonal direct river inputs that are low-to-negligible during the growing season. As a result primary production in the estuaries is controlled principally not by river-driven stratification but by coastal upwelling and bulk exchange with the ocean. Both baroclinic mechanisms (the gravitational circulation) and barotropic ones

(lateral stirring by tide and wind) contribute to this bulk exchange, though tidal circulations appear to dominate during the low-riverflow growing season on ~monthly scales. Because estuarine hydrography and ecology are so dominated by ocean signals, the coast estuaries, like the coastal ocean, are largely synchronous on seasonal and event time scales, though intrusions of the Columbia River plume can cause strong asymmetries between Washington and

Oregon estuaries during spring downwelling conditions. Property coherence increases between spring and summer as wind forcing becomes more spatially coherent along the coast. Estuarine habitat is structured not only by large scale forcing but also by fine scale processes in the extensive intertidal zone, such as differential solar heating or differential advection by tidal currents. Hickey and Banas 3

Introduction

Recent results on the physical oceanography of the U.S. Pacific Northwest (PNW) coastal region are integrated in this paper to provide a framework for understanding ecosystem variability. The coastal region important to the regional ecosystem includes both the nearshore zone and the coastal estuaries. Many species utilize both these regions at different life stages. For example, Dungeness crab frequently utilize coastal estuaries for the first year of their life, re-entering the ocean to become part of the fishery as juveniles (Gunderson et al. 1990). Salmon, on the other hand, utilize the estuary at both the beginning and end of their life cycles and the coastal zone as adults.

As we will demonstrate, ocean variability in nearshore regions of the U.S. West Coast and, in particular, its coastal estuaries, is distinctly different from that in estuaries and nearshore regions of the U.S. East Coast. The West

Coast is embedded within an Eastern Boundary Current System; the East Coast is embedded within a Western

Boundary System. Thus, whereas the West Coast is dominated by upwelling, the East Coast is not; whereas upwelling provides plentiful nutrients to the West Coast and its estuaries, on the East Coast nutrients are more commonly supplied by river outflow.

Ocean variability along the West Coast is generally very large scale (> 500 km), a result of large-scale atmospheric systems (Halliwell and Allen 1987). Nevertheless, we will show that significant alongshore gradients occur in coastal productivity in the PNW. Moreover, we will demonstrate that mesoscale features such as banks and submarine canyons play important and even critical roles in ecosystem function.

In many ways, at least during the summer growing season, coastal estuaries in the PNW may be considered as extensions of the coastal ocean: as we will discuss, flushing rates are on the order of a few days and property variability is controlled by changes at the ocean end of the estuary rather than by riverflow at its head. Thus, like the ocean processes, both the several-day and seasonal fluctuations that occur over the growing season occur nearly simultaneously across the PNW coast estuaries. However, actual values of water properties such as temperature, salinity and velocity will differ depending on the specific estuary configuration.

In the following, the large-scale processes acting on the PNW coastal zone are first described. This is followed by a discussion of nutrient variability (Section 2). With this setting the effects of important mesoscale features such as submarine banks, canyons and river plumes are presented (Section 3). Following the description of coastal processes, processes and variability within the coastal estuaries is described (Section 4), with particular focus on the estuaries studied in the PNCERS program. The interaction of the coastal ocean with these estuaries and the Hickey and Banas 4

similarity of water property variation in the several estuaries are demonstrated using time series and survey data collected in the PNCERS program.

1. Large scale processes in the Pacific Northwest coastal ocean a) Seasonal variability

The U.S. Pacific Northwest coastal zone is embedded within the California Current System (CCS), a system of currents with strong interannual, seasonal and several-day (event) scale variability (Fig. 1) (Hickey 1998). The

California Current System includes the southward California Current, the wintertime northward Davidson Current, the northward California Undercurrent, which flows over the continental slope beneath the southward upper layers, as well as "nameless" shelf and slope currents with primarily shorter-than-seasonal time scales. The PNW includes one major river plume (the Columbia), several smaller estuaries, and (primarily in the north) numerous submarine canyons. The dominant scales and dynamics of the circulation over much of the CCS are set by several characteristics of the physical environment; namely, 1) strong alongshore winds; 2) large alongshore scales for both the winds and the bottom topography (Halliwell and Allen 1987); and 3) a relatively narrow and deep continental shelf. Because of these characteristics, coastal-trapped waves (disturbances that travel northward along the shelf and slope) are efficiently generated and propagate long distances along the continental margins of much of western North

America. Thus, much of the variability in the PNW is caused by processes occurring southward of the region (i.e.,

“remote forcing”). Because of the generally southward alongshore wind stress in spring and summer, coastal upwelling is the dominant process controlling water property variability (see review in Smith 1995).

The California Current flows southward year-round offshore of the U.S. West Coast from the shelf break to a distance of 1000 km from the coast (Hickey 1979, Hickey 1998) (Fig. 1). The current is strongest at the sea surface, and generally extends over the upper 500 m of the water column. Seasonal mean speeds are ~10 cm s-1. The

California Undercurrent is a relatively narrow feature (~10-40 km) flowing northward over the continental slope of the CCS at depths of about 100-400 m as a nearly continuous feature, transporting warmer, saltier Southern water northward along the coast. The Undercurrent has a jet-like structure, with the core of the jet located just seaward of and just below the shelf break and with peak speeds of ~30-50 cm s-1. The Undercurrent provides a possible northward transport route for larvae, larval fish and even phytoplankton seed stock. Because of its proximity to the Hickey and Banas 5

shelf break, the Undercurrent is the source of much of the water supplied to the shelf during coastal upwelling. The onshore transport of this water during upwelling offers a mechanism for onshore transport of plankton entrained in the Undercurrent.

A southward undercurrent (the “Washington Undercurrent”) occurs over the continental slope in the winter season in the PNW (Werner and Hickey 1983). This undercurrent occurs at deeper depths than the northward undercurrent (~300-500 m). The existence of this undercurrent, like that of the northward undercurrent, likely depends on the co-occurrence of opposing wind stress and alongshore pressure gradient forces. The Davidson

Current flows northward in fall and winter north of Point Conception. This northward flow is generally broader

(~100 km in width) and sometimes stronger than the corresponding subsurface northward flow in other seasons (the

"Undercurrent") and extends seaward of the slope.

Currents and water properties of the CCS both over the shelf and in the region offshore of the shelf undergo large seasonal fluctuations. The California Current and Undercurrent are strongest in summer to early fall and weakest in winter. The Davidson Current is strongest in winter. Seasonal mean shelf currents are generally southward in the upper water column from early spring to summer and northward the rest of the year. Over the shelf, the seasonal duration of spring-summer southward flow usually increases with distance offshore and with proximity to the sea surface (Strub et al. 1987b). A northward undercurrent is commonly observed on shelves during the summer and early fall. Off the coast of Vancouver Island a northward flowing buoyancy driven current exists year-round from the coast to at least mid shelf (the Vancouver Island Coastal Current) (Thomson 1981, Hickey et al. 1991). This current opposes the southward shelf break jet current that connects to southward flow off the outer Washington shelf.

Seasonal fluctuations are continuous with similar fluctuations in the Alaskan gyre and the majority of seasonal change in the sea surface height and hence geostrophic currents have been shown to occur within a few tens of kilometers of the coast (Strub and James 2002). Seasonal currents are largely driven by alongshore wind stress

(see review included in Batteen 1997). Satellite altimetry data illustrate that seasonal features gradually migrate offshore and out into the main California Current, so that alternating seasonal bands of northward and southward flow (superimposed on the long term mean California Current) are observed as far as several hundred kilometers from the coast (Strub and James, 2002).

The transition of currents and water properties over the shelf and slope between winter and spring, the

"Spring Transition," is a sudden and dramatic event in the CCS (Strub et al. 1987a). Along much of the coast, during the transition, sea level drops at least 10 cm, currents reverse from northward to southward within a period of several Hickey and Banas 6

days and isopycnals slope upward toward the coast in response to coastal upwelling (Smith 1995). The transition is driven by changes in the large scale wind field and these changes are a result of changes in the large scale atmospheric pressure field over the CCS. A similar rapid transition between summer upwelling and fall downwelling oceanic characteristics does not occur (Strub and James 1988). b) Several-day time scales

In spite of strong seasonal variability in the PNW, the dominant variability occurs at several-day time scales

(Hickey, 1989). Thus, on the shelf, seasonal conditions as described above are often reversed for shorter periods of time. Because of bottom friction, reversals occur more frequently nearshore (Brink et al. 1987). Fluctuations in currents, water properties and sea level over the shelf at most locations are dominated by wind forcing, with typical scales of 3-10 d. A schematic of the locally wind-driven ocean surface circulation in the PNW for winds toward the south (“fair weather”) and winds toward the north (“poor weather”) is shown in Fig. 2. During periods of fair weather the stress of the southward winds on the sea surface accelerates the coastal currents, producing offshore- and alongshore-directed currents in the surface Ekman layer, and alongshore currents elsewhere in the water column

(geostrophically balanced to first order, Allen et al. 1995). Under these conditions, plumes of fresher water originating at coastal estuaries tend to spread offshore and to the south (Garcia-Berdeal et al. 2002). Upwelling occurs within a few kilometers of the coast (typically, within one Rossby radius, about 10 km). During periods of poor weather circulation patterns reverse and freshwater plumes move back onshore (Hickey et al. 1998).

The action of the alongshore wind stress on the sea surface results in an alongshore, baroclinically and frictionally sheared coastal jet in the direction of the wind stress (see model studies in Allen et al. 1995 and Allen and Newberger 1996). The location of maximum speed in the coastal jet moves progressively farther offshore as long as the wind stress continues to act. Typical cross-shelf velocity profiles for Washington and Oregon are shown in Hickey (1989) and Huyer and Smith (1974). The speed maximum most typically occurs near mid shelf (Hickey,

1989). Velocity can decrease by a factor of more than two from top to bottom (or even reverse sign) and by a factor of more than two from the inner shelf to the mid shelf. The cross shelf and vertical structure of the velocity field is important when considering transport of larvae by the coastal current system (Rooper et al. 2002).

The cartoon of shelf circulation shown in Fig. 2 does not include the important effects of remote forcing.

Alongshore gradients in alongshore coastal wind stress are significant, with stronger winds (typically upwelling- favorable) south of the PNW in the spring and summer (Hickey, 1979). Because the West Coast north of Point

Conception has no promontories sufficient to significantly disrupt the coastal wave guide, coastal-trapped waves are Hickey and Banas 7

generated and these waves to first order add to the local wind-generated alongshore currents. Waves off central

Washington have been shown to originate primarily from northern California (Battisti and Hickey 1984). At any given time and location, the ratio of remote and local forcing varies and their relative importance has significant interannual variability due to the dependence on alongshore wind stress gradients (Battisti and Hickey 1984). In summer, free waves are usually important in the PNW, particularly at more northern latitudes such as the British

Columbia coast (Hickey et al. 1991). In winter, local wind forcing dominates in the PNW, especially in regions where winter storms are accompanied by strong northward winds whose strength increases in the direction of propagating waves.

Fluctuations in cross-shelf velocity are not as well understood as those in alongshelf velocity. Although model results show onshore and offshore flow in the surface and bottom boundary layers after sufficient spin-up of the system to an applied wind stress (e.g., Allen et al. 1995, Allen and Newberger 1996), observed velocities are frequently much more complex than those predicted by model studies (Brink et al. 1994). The relatively short alongshore coherence scales (~10-20 km) appear to belie the large-scale nature of the atmospheric forcing mentioned above. In general, the cross-shelf flows appear to be highly three-dimensional, thus including effects of smaller-scale features in the bottom topography and the coastline as well as in the wind field.

Meandering jets and an energetic eddy field carry much of the variance in the California Current off northern and central California (Strub et al. 1991). These jets, which extend over at least the upper 200 m of the water column carry recently upwelled coastal water and associated biological production seaward of the shelf to distances of several hundred kilometers. The strongest jets are generated near coastal promontories where flow separates from the coast, the resulting jet becoming unstable (see model studies in, e.g., Batteen 1997). The meandering jet that separates from the coast near southern Oregon can be traced southward along the whole

California coast (Barth et al. 2000). In contrast, most of the PNW coastal region is not dominated by meandering jets. Satellite-derived patterns of sea surface temperature show only one region where upwelling appears to be enhanced off the Washington coast, and the colder water upwelled in this area flows southward down the shelf rather than across the shelf and coastal margin (Fig. 3).

2. Nutrient supply in the Pacific Northwest coastal zone

Hickey and Banas 8

The CCS contains water of three types: Pacific Subarctic, North Pacific Central and Southern (sometimes termed "Equatorial"). Pacific Subarctic water, characterized by low salinity and temperature and high oxygen and nutrients, is advected southward in the CCS (Hickey 1979, 1998). North Pacific central water, characterized by high salinity and temperature and low oxygen and nutrients, enters the CCS from the west. Southern water, characterized by high salinity, temperature and nutrients, and low oxygen, enters the CCS from the south with the northward undercurrent. In general, salinity and temperature increase southward in the CCS and salinity also increases with depth.

Upwelling along the coast brings colder, saltier and nutrient richer water to the surface adjacent to the coast all along the U.S. West Coast (Huyer 1983). In general, the strength and duration of upwelling (as seen at the sea surface) increases to the south in the PNW. Maximum upwelling occurs in spring and summer. With the exception of regions affected by the Columbia plume, stratification in the CCS is remarkably similar at most locations and is largely controlled by the large-scale advection and upwelling of water masses as described above (Huyer 1983).

In contrast to most U.S. East Coast environments, the shelf is relatively narrow and the nutricline is fortuitously positioned so that nutrient-rich deeper water can be effectively brought to the surface by the wind-driven upwelling. In contrast to most East Coast coastal areas, nitrate input to the ocean from coastal rivers is negligible even from the Columbia, which accounts for the majority of the drainage in this region (Barnes et al. 1972). Both seasonal and event-scale patterns of all macronutrients on the continental shelf are dominated by seasonal and event- scale patterns in upwelling processes (Fig. 4) (Landry et al. 1989, Hickey 1989). Wind-driven upwelling of nutrients from deeper layers fuels coastal productivity, resulting in both a strong seasonal cycle and several-day fluctuations in productivity that follow changes in the wind direction and, hence, upwelling. During an upwelling event, phytoplankton respond to the infusion of nutrients near the coast and this "bloom" is moved offshore, continuing to grow while depleting the nutrient supply. When winds reverse (as occurs during storms), the bloom moves back toward shore where it can contact the coast or enter coastal estuaries (Roegner et al. 2002).

3. Mesoscale features and along-coast gradients

Hickey and Banas 9

The large-scale nature of oceanographic processes in the PNW has been described in the preceding sections. In the section below we describe several mesoscale features that cause local variation in flow patterns and response to forcing. These features may play a role in ecosystem variability that is as significant or even more significant than the large-scale processes. Important mesoscale features include river plumes, submarine canyons, banks and coastal promontories. Such features can modulate the local upwelling response, they can alter flow patterns and they can change environmental characteristics such as turbidity, mixed layer depth, stratification and mixing rates. For these reasons, such features are likely of particular importance to phytoplankton/zooplankton production, growth and retention as well as larval transport and retention.

a) Along-coast gradients in productivity and forcing

Time series of vertically integrated chlorophyll for the Washington and Oregon shelves suggest that chlorophyll is greater on the Washington shelf (Fig. 5). This result, derived from averaging a number of unrelated surveys in the 1970s and 1980s (hence, data are both temporally and spatially aliased), is confirmed by satellite- derived images of ocean color (Strub et al. 1990) and also by recent surveys of the Columbia plume region (Peterson pers. comm.). Off Oregon, only over Heceta Bank does chlorophyll approach values seen off the Washington coast.

Limited studies also suggest higher primary productivity off the Washington coast (Anderson 1972), suggesting that the alongshore difference is not simply due to greater retention on the Washington shelf. The greater productivity is also observed higher in the food chain; e.g., in euphausiids and copepods (Landry and Lorenzen 1989). Moreover, juvenile salmon are observed more frequently off the Washington coast (Pearcy 1992).

The apparently greater richness of the Washington coast is particularly surprising because the gradients in the primary forcing, alongshore wind stress, increase in the opposite direction; i.e., the amplitude of upwelling- favorable coastal winds decreases northward in the PNW (Hickey, 1979). Southward stress frequently differs by almost a factor of two between southern Oregon and northern Washington. In addition, the duration of coastal upwelling also decreases seasonally towards the north. In spite of the alongshore decrease in wind stress we note that the seasonal variation in macronutrient supply to the mid-shelf does not differ substantially between the two regions

(Fig. 4). This result can be attributed to several processes: differences in circulation patterns due to differences in shelf structure; upwelling enhancement by canyons (see subsection c below); and influences of the Columbia plume

(see subsection d below). Hickey and Banas 10

The width and shape of the continental shelf varies substantially in the PNW (Fig. 6). Most important, the width of the shallow nearshore region (arbitrarily defined as the area shallower than 100 m) is greater by more than a factor of two (~50 km) off Washington than off Oregon, with the exception of Heceta/Stonewall Bank in southern

Oregon. The shallow nearshore region is favored by the juveniles and/or returning larvae of many species (e.g.,

Rooper et al. 2002). Model studies in Allen et al. (1995) show that a wider, gently sloping shelf results in slower circulation (i.e., possibly greater retention). Also, on such a shelf the upwelling flow tends to be more concentrated in the bottom boundary layer than in steeper regions. This might explain the apparently similar levels of macronutrients on the Washington and Oregon shelves in spite of the substantially weaker wind stress to the north.

b) Banks

On a West Coast-wide survey of domoic acid in surface waters in 1998, high values of this toxin were measured only in the vicinity of known topographic features such as banks or offshore islands (Fig. 7) (Trainer et al.

2000). Domoic acid frequently results in closures of razor clam beaches along the Washington coast, it has been measured in Dungeness crabs, and it has been responsible for a number of mortalities in seabirds and marine mammals in California (Trainer et al. 2002). Domoic acid is associated with the diatom Pseudo-nitzschia. It seems likely that in regions where large coastal promontories occur, such as off southern Oregon and northern and central

California, plankton and larvae can be swept offshore and southward by the meandering jets and/or eddies that form where coastal jets detach from the shelf. These plankton and larvae likely return to the coast rarely, if at all.

Available nutrients would have been depleted well before the meander could return to the coast (meander scales are several weeks to months). On the other hand, in regions where banks and more complex mesoscale topography occur, such as offshore of the Strait of Juan de Fuca (the Juan de Fuca eddy) or Heceta/Stonewall Bank off the central Oregon coast, flow patterns favoring retention, and perhaps more continuous macronutrient supply as well, are more likely. Maps of ocean pigment clearly show that chlorophyll is greater and located father offshore in the vicinity of both of these features (Strub et al. 1990). Under weak-southward-wind conditions or during periods of northward winds associated with storms, plankton and larvae in these retention areas can move inshore to settle on the coast or enter coastal estuaries (e.g., Trainer et al. 2002). c) Submarine canyons

Hickey and Banas 11

The Washington coast is indented by a number of submarine canyons (Fig. 6). Upwelling of nutrient-rich water is enhanced several-fold in the presence of such canyons (see model study in Allen 1996 and observations in

Hickey 1989). The upwelling from canyons may at least partially compensate for the generally weaker upwelling winds that occur off the Washington coast relative to Oregon.

Canyons also alter regional circulation patterns in a manner that increases local retention (Hickey 1995,

1997). In particular, counterclockwise circulation patterns are generally observed both within and over submarine canyons (although not necessarily extending to the sea surface) (Fig. 8). Such eddies provide an effective mechanism for trapping particles such as suspended sediment or food for organic detritus (Hickey 1995). Zooplankton densities are frequently denser over the submarine canyons off the Washington coast (Swartzman and Hickey 2001).

d) The Columbia River plume

The Columbia River provides over 77% of the drainage between the Strait of Juan de Fuca and San

Francisco Bay (Barnes et al. 1972). The plume from the Columbia River likely has major ecological effects in the

PNW. On a seasonal basis, the plume from the Columbia flows northward over the shelf and slope in fall and winter, and southward well offshore of the shelf in spring and summer. In winter, the plume has a dramatic effect on the

Washington coast, producing time-variable currents as large as the wind-driven currents (Hickey et al. 1998). In summer, fresh water from the Columbia gives rise to the low-salinity signal and associated front used to trace the meandering jet that separates from the shelf at Cape Blanco (Huyer 1983). Both observational and modeling studies show that the plume is a "moving target," changing direction, thickness and width with every change in local wind strength or direction (Fig. 9) (Hickey et al. 1998, Garcia-Berdeal et al. 2002).

River plumes are generally turbid, thereby providing less light for plankton growth, while at the same time providing better cover from grazing for higher trophic levels. Plumes also provide retention areas; eddy-like features are generated within a plume under both steady (Garcia-Berdeal et al. 2002) and unsteady (Yankovsky et al. 2001) outflow conditions. Inshore of the Columbia plume on the Washington coast in winter, a retentive circulation pattern occurs during periods of upwelling-favorable winds (Hickey et al. 1998). Deep mixing is inhibited by high stratification at the base of the plume, thus tending to keep plankton within the euphotic zone. Plumes alter regional current patterns in the upper layers, providing along-plume jets for rapid transport, and convergences and trapping at frontal boundaries on the edges. The model example shown in Figure 10 illustrates these retentive circulation Hickey and Banas 12

patterns and also the along-plume jets. Information on stratification enhancement is given in Garcia-Berdeal et al.

(2002) using a numerical model and Hickey et al. (1998) from observations of the Columbia plume. Recent studies suggest that plume edges are preferred feeding sites for zooplankton. The fact that juvenile salmonids are frequently found near the Columbia plume (Pearcy 1992) may be due to the local retention patterns or to frontal convergences, either of which might enhance food availability in this region.

Other than the Columbia, river plumes on the PNW coast are relatively small, and satellite imagery suggests that their traceable effects are confined to within one or two tidal excursions of the mouth of the river or estuary (not shown). Other river or estuarine plumes include those from Grays Harbor and Willapa Bay, Washington and Coos

Bay, Oregon.

Both the structure and magnitude of the Columbia River plume have significant interannual variability.

During years of high snowpack in the Pacific Northwest (such as 1999), very fresh water from the plume can flood the major coastal estuaries north of the Columbia estuary for prolonged periods, reversing the normal estuarine density and salinity gradients over much of the estuaries. Because such plume intrusions would not occur in estuaries off the Oregon coast, the presence or absence of the plume may provide an important environmental distinction between these estuaries as well as between nearshore coastal regions (see Section 4d below).

e) The Strait of Juan de Fuca

The counterclockwise cold eddy off the Strait of Juan de Fuca (also called the "Tully" eddy; Tully 1942) is situated southwest of Vancouver Island and offshore of northern Washington. The eddy, which has a diameter of about 50 km, forms in spring and declines in fall (Freeland and Denman 1982). The eddy is a dominant feature of circulation patterns off the northern Washington coast and is visible in summertime satellite imagery as a relative minimum in sea surface temperature (Fig. 3) and, generally, a relative maximum in chlorophyll a. The seasonal eddy is a result of the interaction between effluent from the Strait, southward wind-driven currents along the continental slope and the underlying topography, a spur of the Juan de Fuca submarine canyon. A connection between the eddy and the Washington coast was demonstrated in July 1991, when oil that spilled in the eddy was found on the

Washington coast 6 days later (Venkatesh and Crawford 1993). Recent preliminary studies with drifters introduced into a diagnostic numerical model for a summer period in 1998 suggest that drifting particles can escape from the eddy to flow southeastward along the Washington shelf (MacFadyen et al. 2002). During storms, onshore flow in the Hickey and Banas 13

surface Ekman layer moves drifter pathways closer to the coast and even reverses the path to a northward direction.

Pathways of drifters deployed in 2001 in the field were consistent with these modeled pathways; a drifter that had moved southward from its deployment site in the eddy reversed direction during a storm and moved back up the coast, approaching as close as 1 km from the beach. Thus, it seems likely that marine organisms residing in the Juan de Fuca eddy can, under certain ocean conditions, impact the Washington coast.

The photic zone in the Juan de Fuca eddy region is characterized by high ambient macronutrients supplied by wind mixing, episodic wind-driven upwelling, topographically controlled upwelling (Freeland and Denman 1982) and the outflow from Juan de Fuca Strait where deep, nutrient-rich water is brought to the surface by estuarine circulation and tidal mixing (Mackas et al. 1980). Thus, although the ultimate source of nutrients for the eddy is the same as that in a nearshore coastal upwelling region (California Undercurrent water), infusion of upwelled nutrients into the eddy likely occurs on different time scales and with different rates than in regions adjacent to the coast.

Repeated surveys on the northern Washington coast have demonstrated that when domoic acid is present off the coast it is usually within or near the Juan de Fuca eddy (Trainer et al. 2002). The diatom Pseudo-nitzschia is always present in significant numbers when the acid is present and these diatoms are known toxin producers. A relationship between toxification of clams and onshore water movement in storms has been demonstrated in a time series (Trainer et al. 2002) for at least one toxic event. Growing conditions in this mesoscale feature must differ from the large scale conditions along the coast where toxin is not usually produced; and thus, in this case, the mesoscale dynamics are as important as the large scale dynamics in determining the nature of the ecosystem.

4. Pacific Northwest estuaries

We have shown that the oceanic environment of the Northwest coast is broadly coherent, although mesoscale features may be important to local circulation as well as to the ecosystem. In the following section we describe the physical characteristics and dynamics of estuaries linked to that coastal ocean. The geomorphology, freshwater forcing, and tidal regime of these estuaries is first described, with comparison to the more commonly studied coastal-plain estuaries of the East Coast (a). This is followed by a brief description of the PNCERS estuary dataset (b). Coupling between processes in the coastal ocean and the estuaries is next described including a discussion of alongshore coherence between the estuaries and the important effects of Columbia plume intrusions (c).

Last, significant differences between intertidal bank and channel water properties are illustrated (d). Hickey and Banas 14

a) Physical characteristics and forcing

The four estuaries studied in PNCERS (Grays Harbor, Willapa Bay, Yaquina Bay, and Coos Bay, see detailed map in Figure 11) are members of a chain of small estuaries that begins along the Washington coast, spans the coast of Oregon, and continues into northern California. Most of these estuaries are drowned river valleys, formed from sea level rise during the last 10,000 y. Some have also been shaped by ocean-built bars, either partially

(e.g., Willapa Bay, Washington) or entirely (e.g., Netarts Bay, Oregon). Emmett et al. (2000) reviews the geography of this system in detail.

Indices of geomorphology and tide and river forcing for the PNCERS estuaries are given in Table 1. For comparison, the same parameters are included for the Columbia River estuary; San Francisco Bay and South San

Francisco Bay alone; Naragansett Bay; Chesapeake Bay and its tributary the James River; and Plum Island Sound, a small embayment on the Massachusetts coast. Except where otherwise marked, data are from the NOAA National

Estuarine Inventory Data Atlas (NOAA 1985). Volume parameters, which are particularly difficult to define and measure (e.g., Malamud-Roam 2000), are here calculated by simple, approximate methods for the sake of uniformity, and thus only gross patterns among the area and volume parameters are significant. Volume is calculated as the product of mean depth and surface area at mean sea level (MSL), a method which gives errors up to ~20% in comparison with other published figures (NOAA/EPA 1991). Mean tidal prism volume is reported as a percentage of volume at high water, which is calculated as MSL volume plus half the tidal prism itself.

Coos Bay is only a few times larger than tiny Plum Island Sound, but is nevertheless the largest of the

Oregon estuaries. Grays Harbor and Willapa Bay, the two coastal-plain estuaries north of the Columbia, are an order of magnitude larger, comparable in volume and morphology to South San Francisco Bay. Both Washington estuaries consist of multiply-connected channels 10-20 m deep surrounded by wide mud and sand flats. Half or more of the surface area of these estuaries lies in the intertidal zone. Significantly, even the smaller, narrower estuaries of Oregon have similar percentages of intertidal area (Table 1, Percy et al. 1974).

Tides on this coast are mixed-semidiurnal, with spring-neap amplitude variation on the order of 50%

(Emmett et al. 2000). Mean tidal ranges, as shown in Table 1, are generally twice as large as on the outer Atlantic

Coast. The combination of large tidal range with broad, open intertidal surface area yields tidal prisms that are large fractions (30-50%) of total volume. This result holds very generally for Northwest coast estuaries, and is a marked Hickey and Banas 15

difference between these systems and all but the smallest of their counterparts on other North American coasts.

These large tidal prisms suggest that flushing by tidal action is probably important in all these estuaries, even those that receive significant riverflow (Dyer 1973). Tidal excursions, as estimated from current measurements in Willapa,

Grays, and Coos, are 12-15 km, significant fractions (25-50%) of the length of the estuaries.

Table 1 includes long-term mean flows for the lowest- and highest-flow months of the year, and, as a measure of the strength of river forcing relative to estuary size, the "river-filling time," volume divided by flow rate.

The output from the Columbia River is two orders of magnitude larger than riverflow into the other coastal estuaries.

With the exception of the Columbia, these estuaries receive freshwater input from local rainfall only, not from snowmelt, and their riverflows show strong seasonality in concert with the winter storm and summer dry- and fair- weather seasons. Local riverflow peaks during winter storms and is negligible during late summer. The seasonal variation is generally several times greater than in East Coast estuaries, though flood and drought events beyond the mean seasonal cycle have not been considered here. As a result we might expect the hydrodynamic classification of

Northwest estuaries to change dramatically between seasons, or even—where flushing and adjustment times are short—between individual wind events.

This riverflow pattern yields a seasonal hydrographic cycle that contrasts strongly with traditional models of temperate partially mixed estuaries, with important ecological implications. Tyler and Seliger (1980), for example, show that primary production in Chesapeake Bay is controlled by "stratification dependent pathways" reminiscent of the seasonal dynamics of the open-ocean mixed layer. In that estuary, in winter, mixing by wind and tide erases stratification and resuspends nutrients, while in spring and summer increased riverflow and solar heating produce strong stratification and reduced vertical exchange. In such a system, stratification controls on vertical mixing are crucial to determining plankton growth rates and the potential for phytoplankton blooms, as in San Francisco Bay

(Lucas et al. 1999). In sharp contrast, in Willapa Bay stratification is in general very low during summer, when riverflows are low, and high during the winter, when riverflow peaks (Banas et al. 2002). Vertical, one-dimensional, stratification-centered models of primary productivity thus would not apply here even at the coarsest level. Rather, during the growing season in Pacific Northwest estuaries, hydrography, nutrient levels, and biomass all appear to be controlled less by in situ processes than by mesoscale processes in the coastal ocean (Hickey et al. 2002, Roegner et al. 2002). Below we consider this ocean-estuary coupling in more detail. b) PNCERS observations Hickey and Banas 16

During PNCERS, arrays of moored sensors were maintained in three coastal estuaries (Grays Harbor,

Willapa Bay and Coos Bay) as well as at two sites in the nearshore costal ocean, one off Washington, the other off southern Oregon. Locations are shown in Figure 11. Moored sensors included S4 current meters or ADCPs, and

Seabird C-T sensors or Aanderaa current meters equipped with conductivity and temperature sensors. Estuarine instruments were set in the lower water column, sustained by a taut wire mooring. Sampling interval was less than 30 minutes in the estuaries and hourly on the coast. Two arrays were maintained in Willapa Bay and Coos Bay and one in Grays Harbor. The longest time series (temperature) spans 3 years. In general, salinity time series are much shorter due to fouling and clogging problems. Hydrographic sections were made with a SeaBird 19 CTD at sporadic intervals when mooring instruments were exchanged or cleaned. At the same time as PNCERS, under the direction of

Dr. Jan Newton, Washington State Department of Ecology maintained sensors at a number of locations in Willapa

Bay and also sampled hydrographic sections. More detailed analysis of these time series can be found in Banas et al.

2002 and Siegel et al. 2002. NCAR NCEP six-hourly winds from the Reanalysis project (Kalnay et al. 1996) were obtained at 2.5 degree intervals and interpolated to Grays Harbor and Coos Bay mid shelf locations. These data are provided by the NOAA-CIRES Climate Diagnostics Center, Boulder, Colorado at http://www.cdc.noaa.gov/. These winds are generated from an atmospheric model that is primarily driven by atmospheric pressure but also include data assimilated from both coastal buoys and land stations. The NCEP winds, being a spatial average, provide a more accurate representation of alongshore gradients in wind than in situ buoys, which are frequently biased by cross-shelf wind structure. Data from NDBC Buoy 46029 (the “Columbia River buoy”) were used in Figure 15a

(only).

Data were edited for outliers. Hydrographic section data were used to validate data from the moored arrays.

For subtidal time series, data were filtered with a Butterworth low pass filter and smoothed to hourly intervals.

c) Links to the coastal ocean

Response to upwelling/downwelling

As suggested above, an account of event-scale and seasonal variations in ocean-estuary coupling is critical to a description of the physical dynamics of Northwest coast estuaries as well as their ecosystems. In this section we Hickey and Banas 17

review recent analyses and observations of time-dependent coupling mechanisms, and present new data comparing the responses of the PNCERS estuaries to ocean forcing.

The properties of the ocean water presented at the mouth of the estuary are governed by whether upwelling or downwelling is occurring along the coast at that time (Hickey et al. 2002). During upwelling, surface waters move offshore and cold, saltier, nutrient-rich water is moved upward within a few km of the coast; phytoplankton seed stock are also upwelled into the euphotic zone and, fueled by the high nutrient level, begin to grow (Roegner et al.

2002). The growing phytoplankton move offshore as new seed stock is upwelled so that the highest biomass may be situated some distance from the coastal wall and the mouths of the estuaries. During downwelling, warmer, fresher, nutrient-depleted surface waters move inshore and downward, and offshore phytoplankton blooms likewise are advected back to the coast.

Oceanic phytoplankton can enter a coastal estuary by two routes. During upwelling events, seed stock can be pulled into the estuary, where a local bloom is fueled by the high nutrients brought in with the upwelled water (de-

Angelis and Gordon 1985). During downwelling events, phytoplankton from a prior offshore bloom can be pulled directly into the estuary (Roegner et al. 2002). This biomass, although nutrient-poor and declining rather than growing, may provide a direct food source to secondary production, particularly near the mouth of the estuary. In the

Pacific Northwest, transitions between upwelling and downwelling occur at 2-10 day intervals (Hickey 1989), and so the ocean end-member of estuarine water properties can change significantly over just a few tidal cycles.

In Willapa Bay both upwelling and downwelling water presented at the mouth of the estuary have been observed to travel up-estuary in the lower water column at a rate on the order of 10 km d-1, modifying the gravitational circulation of the estuary as it passes (Hickey et al. 2002). These modulations of circulation and water properties lag local wind stress fluctuations (hence, upwelling or downwelling) by more than a day (Fig. 12). This mode of up-estuary propagation is consistent with the suggestion by Duxbury (1979) that modulation of the gravitational circulation by upwelling and downwelling is responsible for increased mean flushing rates in summer months in Grays Harbor. Such a baroclinic coupling between ocean and estuary is schematized in Fig. 13, with values taken from typical early-summer conditions in Willapa Bay. During upwelling events, high ocean salinities increase the along-channel salinity gradient and hence the magnitude of baroclinic exchange (Hansen and Rattray

1965, Monteiro and Largier 1999); during downwelling events, the salinity contrast between ocean and estuary, and hence the strength of the exchange flow, are reduced. Hickey and Banas 18

At the same time, the propagation of oceanic signals into the estuary appears to have, in addition to this baroclinic, density-driven component, a significant diffusive, primarily barotropic, density-independent component.

In the long-term average in Willapa Bay, oceanic signals appear to propagate upstream in the surface layer (opposing the mean gravitational circulation) at a rate similar to that of their propagation in the lower layer. The strength and along-channel profile of this diffusive process suggest lateral stirring by wide tidal residual eddies tied to bathymetry

(Banas et al. 2002). Lateral wind-driven circulations (e.g., Wang 1979, Geyer 1997) may also be important, but as yet have not been quantified.

The effectiveness of ocean-estuary exchange by tidal stirring depends not only upon processes within the estuary, but also upon net transport and mixing on the shelf within a tidal excursion of the mouth. Shelf processes on the tidal time scale presumably determine, largely, the fraction of an ebb tidal prism that does not simply re-enter the estuary on the following flood tide (the "tidal exchange ratio"). From measurements of horizontal tidal diffusivity

Banas et al. (2002) estimate a tidal exchange ratio ≥ 0.6 for Willapa Bay as a whole, with the possibility of much lower exchange ratios across cross-sections in the landward reaches of the bay. The exchange ratio at the mouth is at the upper limit of the range reported by Dyer (1973), and thus is consistent with the active, highly advective inner- shelf environment on this coast.

The studies reviewed above demonstrate that both baroclinic, density-driven exchange and diffusive, tide- or wind-driven exchange can contribute significantly to the flushing of Northwest estuaries. Whichever of these exchange mechanisms dominates determines the overall rate of ocean-estuary exchange, and the spatial pathways along which oceanic nutrients, phytoplankton, and planktonic larvae enter an estuary on tidal or subtidal time scales.

In general, tide- and wind-driven mechanisms are expected to dominate in small, shallow, well-mixed estuaries, and density-driven exchange is expected to dominate in deeper, partially stratified systems (e.g., Hansen and Rattray

1966). The Northwest coast estuaries span both of these broad categories, and we can expect the relative role of density-driven and density-independent mechanisms to vary significantly between systems, and over time in a single system.

On seasonal time scales, Willapa Bay, for example, appears to vary between river-controlled and tide- controlled flushing. Banas et al. (2002) show that river-controlled exchange dominates in the narrow, landward, southeastern reach of Willapa during all but the lowest late-summer riverflows. At the same time, in the seaward reach of the estuary, flushing by tidal stirring dominates during all but the largest winter storms, even when stratification is sustained at several psu. Thus tidal stirring appears to control the nutrient and biomass budget for Hickey and Banas 19

most of the volume of Willapa during the spring-to-summer growing season. The morphological and forcing properties that place Willapa at this transition point between river- and tide-controlled flushing are not particular to

Willapa, but rather general characteristics of Northwest coast estuaries: large tidal prisms, complex bathymetry, and highly variable riverflow that is frequently large during winter but close to zero during much of the growing season.

Alongshore correlation between estuaries

As discussed in Section 2, wind-driven coastal processes in summer have scales of several hundred kilometers. Thus, if the estuarine-ocean coupling processes described above are uniformly valid, we might expect water properties in many PNW estuaries to vary coherently. The PNCERS data confirm that this is indeed the case: time series of temperature data collected simultaneously in Grays Harbor, Willapa Bay, and Coos Bay demonstrate that in the spring-fall growing season all three estuaries, which span 400 km of the Northwest coast are highly coherent (Fig. 14a). Coherence is highest in late summer-early fall (day 201-271, r = 0.95 between Grays and

Willapa, 0.91 between Coos and Grays). In spring-early summer (day 130-200) a few peaks occur in the northern estuaries that do not occur in the southern estuary, reducing coherence (r = 0.74). The spring-summer period is further analyzed in the next three panels of Figure 14, showing temperature (b), alongshore wind (c) and salinity (d).

Comparison of temperature and wind illustrates the general response to upwelling and downwelling favorable winds, with the wind/property lag (1.5 d at both Grays Harbor and Coos Bay) discussed in the preceding section. Significant differences between the northern and southern estuaries are observed. For example, from day 165-175 temperature decreases in Coos Bay while it increases in Grays Harbor (Fig. 14b). Comparison with alongshore wind (Fig. 14c) demonstrates that this difference is caused by the fact that downwelling favorable winds are stronger during this period near Grays Harbor than near Coos Bay. Several other similar examples of alongshore differences in estuary water properties caused by alongshore wind differences can be seen in the records.

Another example is illustrated in salinity records from northern and southern estuaries from the same period

(Fig. 14d). In the southern estuary (Coos Bay) the upwelling signal is of significantly longer duration than in Grays

Harbor. Note that the greater salinity range and overall lower salinity at the northern estuary is consistent with the generally lower regional salinity due to the proximity of the Columbia plume (see next section).

Columbia River plume intrusions

Hickey and Banas 20

Although the estuaries of Washington and Oregon generally respond to ocean forcing coherently as discussed above, the Columbia River plume can cause major asymmetries between the estuaries. Since the plume moves offshore when it flows southward past Oregon during periods of upwelling-favorable winds, it does not impinge upon most Oregon estuaries directly. When the plume flows north under downwelling winds, however, it fills the nearshore water column north of the river mouth past the depth of the estuary mouths (Garcia-Berdeal et al.

2002, Roegner et al. 2002). The plume may also impact estuaries on the northern Oregon coast during downwelling, when the southwest-tending plume formed under the preceding upwelling conditions and seasonal southward ambient flow moves shoreward. However, mixing during the downwelling event would result in much less density contrast than off the Washington coast, where the plume is relatively new and thus fresher. The effect of the plume on the estuaries is most dramatic and sustained in late spring and early summer, when local riverflow has slackened but the Columbia is still running high with snowmelt.

Lower water column salinities from moorings inside the mouths of Willapa Bay and Grays Harbor during

April and May 2000 are shown in Fig. 15. For each station, the along-channel salinity gradient has also been calculated as a subtidal time series, by dividing the difference between high- and low-water salinities by the tidal excursion for each semidiurnal tidal cycle, and then filtering the resulting discrete series. This method, also used by

Banas et al. (2002), takes advantage of the fact that each station effectively samples ~ 15 km of the channel through tidal advection. This allows us to calculate along channel gradients without requiring pairs of stations to obtain differences. An upwelling event, which brings ~32 psu water into the estuaries and produces strong along-channel gradients (on April 19, ~5 psu over one tidal excursion), is followed by a plume intrusion, indicated by a dramatic decrease in salinity and weak along-channel salinity gradients. When downwelling-favorable winds slacken after

~April 27, salinity and the along-channel salinity gradient increase again. The five-month wind time series shown in

Figure 15a suggests that this intermittent alternation of upwelling and plume intrusion continues from late winter through early summer.

During the onset of plume intrusions the along-channel salinity gradient in the estuary can reverse for sustained periods. In Figure 15b, for example, as the plume intrusion intensifies during the period April 20-28, salinity at the Willapa Bay mooring at high slack water (indicated by dots) is generally lower than the subtidal average, indicating that each flood tide is bringing somewhat fresher water into the estuary. This reversal of the expected gradient between mid-estuary and ocean water is illustrated in a CTD transect along the main channel of Hickey and Banas 21

Willapa Bay on May 3, 2000, during the recovery from the plume intrusion (Fig. 16a). Salinity increases downstream from the head to > 21.8 psu, drops to < 21.4 psu, and then increases again within one tidal excursion of the mouth.

Vertical gradients weaken during plume intrusions along with the longitudinal gradients. The vertical salinity difference in the interior of the estuary in the May 3 transect is on the order of 0.1 psu. In comparison, a transect on May 30 during the onset of an upwelling event after a period of intermittent winds (Fig. 16b) shows vertical salinity differences ~2-4 psu within a tidal excursion of the mouth. During a plume intrusion, the reduced salinity contrast between the river and ocean end-members of the estuary presumably weakens baroclinic pressure gradients and thus stratification to the point where vertical shear dispersion can completely homogenize the water column. Thus in contrast to input of freshwater from the local rivers, which tends to increase stratification and gravitational exchange (Banas et al. 2002), input of freshwater from the Columbia River via the coastal ocean tends to produce near-complete mixing in Washington estuaries.

As described in the preceding section, downwelling conditions tend to reduce estuarine salinity gradients even in the absence of plume intrusions (Hickey et al. 2002), though to a much lesser extent. The effect of the

Columbia River plume, then, is to greatly intensify the contrast between spring and summer upwelling and downwelling conditions in the Washington estuaries in comparison with Oregon estuaries. This asymmetry between the two coasts would likely be observed not just on the event scale, but on interannual scales as well. Following wet

(La-Niña-like) winters like 1998-1999, but not following dry (El-Niño-like) winters like 1997-1998, sustained plume intrusions would be expected in the Washington estuaries during May and June. Anecdotal evidence suggests that in years prior to the man-made reduction in spring freshets from the Columbia, the surface of Willapa Bay would freeze sufficiently to support walking, suggesting very low wintertime salinities (Proc. Oyster Grow. Conf. 2002).

d) Spatial variability in the intertidal zone

We have argued that much of the physical forcing important to estuarine productivity is coherent over tens or hundreds of kilometers on this coast, and that the response of the coastal estuaries to this forcing may be coherent and generalizable on this scale as well. At the same time, pervasive, significant variation in currents and hydrography is possible on much smaller scales—as short as ~ 100 m—in estuaries with complex bathymetry, particularly in very shallow regions, which often are most important biologically. These small-scale variations, which can be thought of Hickey and Banas 22

as creating estuarine microenvironments, easily confound attempts to generalize from measurements that do not integrate over larger scales.

In this section we use new data to describe the two mechanisms of lateral variability best resolved by tidal scale observations in the Washington estuaries: 1) direct solar heating of bank water, and 2) the creation of persistent lateral gradients by tidal advection. A full account of the transverse structure of these estuaries—which must consider competition and interaction between tidal currents, density-driven flows, rotational effects, and wind-driven circulations, all of which are shaped by bathymetry (e.g., Friedrichs et al. 1992, Valle-Levinson and O'Donnell

1996)—is beyond the scope of available data.

Solar heating

Coordinated longitudinal (along-channel) and transverse (bank-to-channel-to-bank) CTD transects were obtained in Willapa Bay and Grays Harbor during the summers of 1999 and 2000. These observations frequently suggest solar heating of water on shallow intertidal flats: either direct heating of the water at high tide, or transfer to the water of heat stored in the mud flats themselves from insolation at low tide. Consider, for example, a late- afternoon, early-flood transect along the main channel of Grays Harbor during a period of fair weather in June 1999

(Fig. 17). The warmest water in the channel is associated with neither the ocean nor the river end-member, but rather appears near the surface over a broad middle reach of the channel. CTD casts along this transect were separated by

~4 km, and therefore the spatial structure of this warm water may be patchier than contouring between casts allows.

We interpret this signal as evidence of water warmed during the midday high tide that has circulated back into the main channel on the following ebb. A temperature-salinity (T-S) diagram of this transect (Fig. 17c) shows clearly that this signal represents warming of water at intermediate salinity, and effectively constitutes a third mixing end- member, toward which the T-S profile of the channel is inflected. Furthermore, transverse, channel-to-shoal surveys on the day of the along-channel transect and over the next four days locate a similar warm water mass in depths < 5 m at higher stages of the tide (dots in Figure 17c).

Surveys in Willapa Bay from June and July 2000 (Fig. 18) show similar results: warmest temperatures on banks in the interior of the estuary, inflection of the main-channel T-S profile that lifts intermediate water above the mixing line between the ocean and river end-members. The warmest points in the June 2000 survey, more than 4°C Hickey and Banas 23

warmer than main-channel water of the same salinity, represent the shallowest water sampled, water < 0.5 m deep sampled by foot with a hand-held meter.

Note that since the fair-weather events which bring increased insolation also bring cold, upwelled water, an estuary's response to direct heating may be masked on the event scale, and better resolved by an integration over many events. In Willapa Bay, where time series of along-channel transects exist, the inflection of the T-S profile tends to increase as the summer proceeds, though not monotonically (not shown).

Differential tidal advection

Not all bank-to-channel hydrographic variations result from solar heating or other transformation of water properties. Consider the along- and cross-channel flood-tide transects from July 1999 in Willapa Bay shown in Fig.

19. The along-channel salinity gradient is ~5 psu over one tidal excursion (15 km); across a shallow, narrow bank adjacent to the main channel during late flood, the salinity gradient is ~ 4 psu over only 1.3 km. Huzzey (1988) likewise found that in the York River, which like Willapa consists of a deep central channel flanked by shoals, the freshest water in a cross-section at high slack water was located on the banks. A T-S diagram of the July 1999 transects (Fig. 19c) shows that the bank and channel water masses, unlike those shown in Figs. 17 and 18, are indistinguishable. The lateral variation in salinity and temperature thus must have arisen from advective rearrangement, not transformation, of main-channel water in the intertidal zone.

Large lateral gradients can arise solely from differential advection by tidal currents (Huzzey and Brubaker

1988, O'Donnell 1993); i.e., the fact that on a shallow bank tidal motion is slowed by friction so that a given flood or ebb moves water parcels farther longitudinally in a channel than on an adjacent shoal. This shearing of the flow effectively transfers the along-channel gradient over one tidal excursion, or some fraction thereof, into a cross- channel gradient. In support of this explanation for the lateral variation seen in Willapa in July 1999, repeated channel-to-bank surveys in the same location have shown that the transverse gradient there at high water follows the along-channel variation. On November 1-2, 1999, for example, the along-channel salinity gradient in the central reach of the estuary was much weaker than that shown in Fig. 19, only ~ 0.5 psu over one tidal excursion (Banas et al. 2002), and the salinity variation over the bank was likewise ~ 0.5 psu.

The differential-advective effect would be expected to be strongest on the shallowest banks (like that shown in Figure 19b) where the effect of friction is presumably greatest, and less important on deeper, subtidal shoals. Such Hickey and Banas 24

lateral structure in tidal advection may have important local, biological consequences. For example, sessile organisms in a shallow region with strong lateral gradients may experience mean temperatures or rates of nutrient or food supply appreciably different—more like conditions a large fraction of a tidal excursion up-estuary—than organisms in deeper water a short distance away. At the same time, differential tidal advection may contribute to overall estuarine flushing if these lateral shears are a lateral-dispersion mechanism similar to the models of "tidal trapping" reviewed by Fischer (1976).

5. Discussion

New information on oceanographic processes has been synthesized to provide a framework for better understanding some aspects of ecosystem variability. In addition new data from three coastal estuaries have been presented to identify dominant processes and scales of variability, and to illustrate how the estuaries interact with the coastal ocean.

From an ecological point of view the study presents the important idea that in spite of the fact that upwelling wind stress is as much as a factor of two weaker off Washington than off Oregon, productivity in general is higher off Washington (with the exception, at times, of Heceta Bank in southern Oregon). Two possible mechanisms were discussed: the difference in shelf structure and the presence of submarine canyons on the

Washington coast. An additional possibility to account for these differences is the role of micronutrients. The few primary productivity measurements in the literature illustrate enhanced productivity in the Columbia plume

(Anderson 1972), a rich source of iron. Moreover, because of the Columbia, iron-rich sediments overlie the

Washington shelf in the mid shelf silt deposit (Nittrouer 1978). These sediments could provide iron to upwelling phytoplankton via contact in the bottom boundary layer. Off Oregon, plume sediments are much reduced.

Furthermore, because of the steeper shelf, upwelling water may cross the shelf more in the interior and less in the bottom boundary layer (Allen et al. 1995). Unfortunately, existing data are not sufficient to separate the several plausible mechanisms that could produce the observed alongshore productivity differences.

Within the coastal estuaries, the physical questions of primary ecological importance involve ocean-estuary exchange. The construction of budgets for nutrients or phytoplankton in these estuaries requires better knowledge of the bulk rate of exchange, and its variation on seasonal and event time scales. Likewise, delineation of the pathways Hickey and Banas 25

of larval recruitment into the estuaries requires that we better understand variations in ocean-estuary matter transfer on event and tidal time scales, as well as the fine-grained spatial structure of residual circulations on these scales.

Bulk coupling between ocean and estuary is necessarily central to the dynamics of any small embayment; but its importance is perhaps amplified in PNW estuaries for two reasons. First, the ocean rather than local rivers is the dominant source of nutrients and biomass along this coast. Second, oceanic water properties are extremely variable on the scale of the residence time or adjustment time of the estuaries themselves (i.e., on the event scale).

This coincidence of time scales makes the estuaries highly variable and unsteady themselves, far more so, during summer, than fluctuations in local river input would force on their own.

The example of Willapa Bay suggests that we may, in fact, be able to neglect the influence of river-driven circulations on the overall flushing rate of these estuaries during some portion of the year (roughly speaking, the growing season) and on sufficiently long time scales (several events or a few weeks). This is a very approximate model of estuarine dynamics during this time period, but a powerful simplification, if tenable. It should be noted that even if one can parameterize estuary flushing in terms of tide- or wind-driven lateral stirring as we are suggesting, freshwater influences may still be important in determining the distribution of heat, salt, and other tracers within the estuary.

The data collected to date, while providing a framework for beginning to understand the processes in the important estuaries of the PNW, have also demonstrated the complexities that they include. In our ongoing research, a three dimensional numerical model is being used to separate the dominant forcing mechanisms, Lagrangain pathways and estuary-ocean fluxes. The PNW coast estuaries constitute a fruitful set for dynamical or ecological comparison: they have enough commonalities (similar tidal forcing, similar riverflow patterns, event-scale synchrony) to be usefully compared, and at the same time enough diversity (in overall size and cross-sectional shape, relative riverflow magnitude, and relation to the Columbia plume) to form an interesting natural experiment.

Acknowledgments

Data collection was supported by the Pacific Northwest Ecosystem Research Study (PNCERS) (grant #

NA76RG0485 and NA96OP0238 from the Coastal Ocean program of the National Oceanic and Atmospheric

Administration). Analysis was supported by PNCERS, Washington Sea Grant (grant # NA16RG1044-R/ES-42 and Hickey and Banas 26

NA16RG1044-R/F-137) and by a grant (#OCE-0001034) to B. Hickey from the National Science Foundation as part of GLOBEC. This is contribution number XXXX of the U.S. GLOBEC program, jointly funded by the National

Science Foundation and National Oceanic and Atmospheric Administration. Last, we would like to thank Dr. Jan

Newton and Mr. Eric Siegel of the Washington State Department of Ecology for graciously providing additional

CTD data for Willapa Bay and Dr. John Klinck and M. Dinneman for providing modeled circulation to illustrate the canyon effects.

. Hickey and Banas 27

Literature Cited

Allen, S. 1996. Topographically generated, subinertial flows within a finite length canyon. Journal of Physical

Oceanography 26:1608-1632.

Allen, J.S. and P.A. Newberger. 1996. Downwelling circulation on the Oregon continental shelf. Part I: Response to

idealized forcing. Journal of Physical Oceanography 26(10):2011-2035.

Allen, J.S., P.A. Newberger, and J. Federiuk. 1995. Upwelling circulation on the Oregon continental shelf. Part I:

Response to idealized forcing. Journal of Physical Oceanography 25:1843-1866.

Anderson, G.C. 1972. Aspects of marine phytoplankton studies near the Columbia River, with special reference to a

subsurface chlorophyll maximum. In D.L. Alverson and A.L. Pruter (eds.), Bioenvironmental Studies of the

Columbia River Estuary and Adjacent Ocean Regions. University of Washington Press, Seattle, Washington, pp.

219-240.

Andrews, R.S. 1965. Modern sediments of Willapa Bay, Washington: a coastal plain estuary. Technical Report

Number 118, Department of Oceanography, University of Washington.

Barnes, C.A., A.C. Duxbury, and B.A. Morse. 1972. Circulation and selected properties of the Columbia River

effluent at sea. In D.L. Alverson and A.L. Pruter (eds.), Bioenvironmental Studies of the Columbia River Estuary

and Adjacent Ocean Regions. University of Washington Press, Seattle, Washington, pp. 41-80.

Barth, J.A, S.D. Pierce, and R.L. Smith. 2000. A separating coastal upwelling jet at Cape Blanco, Oregon and its

connection to the California Current System. Deep-Sea Research 11(47):783-810.

Batteen, M.L. 1997. Wind-forced modeling studies of currents, meanders and eddies in the California Current system.

Journal of Geophysical Research 102(C1):985-1010.

Hickey and Banas 28

Battisti, D. and B.M. Hickey. 1984. Application of remote wind forced coastal trapped wave theory to the Oregon and

Washington coasts. Journal of Physical Oceanography 14:887-903.

Brink, K.H., D.C. Chapman, and G.R. Halliwell. 1987. A stochastic model for wind-driven currents over the

continental shelf. Journal of Geophysical Research 92(C2):1783-1797.

Brink, K.H., J.H. La Case, and J.D. Irish. 1994. The effect of short-scale wind variations on shelf currents. Journal of

Geophysical Research 99(C2):3305-3315.

de-Angelis, M.A. and L.I. Gordon. 1985. Upwelling and river runoff as sources of dissolved nitrous oxide to the

Alsea estuary, Oregon. Estuarine Coastal and Shelf Science 20(4):375-386.

Duxbury, A.C. 1979. Upwelling and estuary flushing. Limnology and Oceanography 24(4):627-633.

Dyer, K.R. 1973. Estuaries: a physical introduction. Wiley, New York.

Emmett, R., R. Llanso, J. Newton, R. Thom, M. Hornberger, C. Morgan, C. Levings, A. Copping, and P. Fishman.

2000. Geographic signatures of North American West Coast estuaries. Estuaries 23(6):765-792.

Fischer, H.B. 1976. Mixing and dispersion in estuaries. Annual Review of Fluid Mechanics 8:107-133.

Freeland, H.J. and K.L. Denman. 1982. A topographically controlled upwelling center off southern Vancouver Island.

Journal of Marine Research 40:1069-1093.

Friedrichs, C.T., Lynch D.R., and D.G. Aubrey. 1992. Velocity asymmetries in frictionally-dominated tidal

embayments: longitudinal and lateral variability. In D. Prandle (ed.), Dynamics and exchanges in estuaries and

the coastal zone. American Geophysical Union, Washington, DC.

Hickey and Banas 29

Garcia-Berdeal, I., B.M. Hickey, and M. Kawase. 2002. Influence of wind stress and ambient flow on a high

discharge river plume. Journal of Geophysical Research. In press.

Geyer W.R. 1997. Influence of wind on dynamics and flushing of shallow estuaries. Estuarine, Coastal and Shelf

Science 44:713-722.

Gunderson, D.R., D.A. Armstrong, Y.B. Shi and R.A. McConnaughey. 1990. Patterns of estuarine use by juvenile

English sole (parophrys retulus) and Dungeness crab (Cancer magister). Estuarines 13(1):59-71.

Halliwell Jr., G.H. and J.S. Allen 1987. Large-scale coastal wind field along the west coast of North America, 1981-

1982. Journal of Geophysical Research. 92 (C2):1861-1884.

Hansen, D.V. and M. Rattray. Jr. 1965. Gravitational circulation in straits and estuaries. Journal of Marine Research

23:104-122.

Hansen, D.V. and M. Rattray. Jr. 1966. New dimensions in estuary classification. Limnology and Oceanography

11:319-326.

Hickey, B.M. 1979. The California current system - hypotheses and facts. Progress in Oceanography 8:191-279.

Hickey, B.M. 1995. Coastal submarine canyons. In P. Muller and D. Henderson (eds), Proceedings of the University

of Hawaii 'Aha Huliko'a Workshop on Flow Topography Interactions; Honolulu, Hawaii; SOEST Special

Publication, 265 pp.

Hickey, B.M. 1997. Response of a narrow submarine canyon to strong wind forcing. Journal of Physical

Oceanography 27(5):697-726.

Hickey, B.M. 1989. Patterns and processes of shelf and slope circulation. In M.R. Landry and B.M. Hickey (eds.),

Coastal Oceanography of Washington and Oregon, Elsevier Science, Amsterdam, The Netherlands, pp. 41-115, Hickey and Banas 30

Hickey, B.M. 1998. Coastal Oceanography of Western North America from the tip of Baja California to Vancouver

Is. In K.H. Brink and A.R. Robinson (eds.), The Sea, Volume 11, Chapter 12, Wiley and Sons, Inc., pp. 345-393.

Hickey, B.M. and T. Royer. 2001. California and Alaskan Currents. In J.H. Steele, S.A. Thorpe, and K.A. Turekian

(eds.), Encyclopedia of Ocean Sciences. Academic Press.

Hickey, B.M., L. Pietrafesa, D. Jay, and W.C. Boicourt. 1998. The Columbia River Plume Study: subtidal variability

of the velocity and salinity fields. Journal of Geophysical Research 103(C5):10,339-10,368.

Hickey, B.M., R.E. Thomson, H. Yih, and P.H. LeBlond. 1991. Velocity and temperature fluctuations in a

buoyancy-driven current off Vancouver Island. Journal of Geophysical Research 96(C6):10,507-10,538.

Hickey, B.M., M. Zhang, and N. Banas. 2002. Coupling between the California Current System and a coastal plain

estuary in low riverflow conditions. Journal of Geophysical Research. In press.

Huyer, A., 1983. Upwelling in the California Current system. Progress in Oceanography 12:259-284.

Huyer, A. and R.L. Smith, 1974. A subsurface ribbon of cool water over the continental shelf off Oregon. Journal of

Physical Oceanography 4:381-391.

Huzzey, L.M. 1988. The lateral density distribution in a partially mixed estuary. Estuarine, Coastal and Shelf Science

26:351-358.

Huzzey, L.M. and J.M. Brubaker. 1988. The formation of longitudinal fronts in a coastal plain estuary. Journal of

Geophysical Research 93:1329-1334.

Jay, D.A., W.R. Geyer, R.J. Uncles, J. Vallino, J. Largier, and W.R. Boynton. 1997. A review of recent

developments in estuarine scalar flux estimation. Estuaries 20(2):262-280. Hickey and Banas 31

Kalnay, E., M. Kanamitsu, R. Kistler, W. Collins and group. 1996. The NCEP/NCAR 40-year Reanalysis Project.

Bulletin of the American meteorological Society 77(3):437-471.

Landry, M.R., J.R. Postel, W.K. Peterson, and J. Newman. 1989. Broad-scale patterns in the distribution of

hydrographic variables. In M.R. Landry and B.M. Hickey (eds.), Coastal Oceanography of Washington and

Oregon. Elsevier Press, Amsterdam, The Netherlands, pp. 1-41.

Landry, M.R. and C. Lorenzen. 1989. Abundance, distribution and grazing impact of zooplankton on the Washington

shelf. In M.R. Landry and B.M. Hickey (eds.), Coastal Oceanography of Washington and Oregon. Elsevier Press,

Amsterdam, The Netherlands, pp. 175-210.

Lucas, L.V., J.R. Koseff, J.E. Cloern, S.G. Monismith, and J.K. Thompson. 1999. Processes governing

phytoplankton blooms in estuaries. I: The local production-loss balance. Marine Ecology Progress Series 187:1-

15.

Mackas, D.L., G.C. Loutitt and M.J. Austin. 1980. Spatial distribution of zooplankton and phytoplankton in British

Columbian coastal waters. Canadian Journal of Fishery and Aquatic Science 37:1476-1487.

Monteiro, P.M.S. and J.L. Largier. 1999. Thermal stratification in Saldhana Bay (South Africa) and subtidal,

density-driven exchange with the coastal waters of the Benguela upwelling system. Estuarine and Coastal Shelf

Science 49:877-890.

National Oceanic and Atmospheric Administration. 1985. National estuarine inventory data atlas. Strategic

Assessment Branch, Ocean Assessments Division, Office of Oceanography and Marine Assessment, National

Ocean Service, Rockville, Maryland.

National Oceanic and Atmospheric Administration, Environmental Protection Agency. 1991. Susceptibility and

status of West Coast estuaries to nutrient discharges: San Diego Bay to Puget Sound. Summary report, Strategic

Assessment of Near Coastal Waters. Hickey and Banas 32

Nittrouer, C.A. 1978. The process of detrital sediment accumulation in a continental shelf environment: an

examination of the Washington shelf. Ph.D. Dissertation., Department of Oceanography, University of

Washington, Seattle, Washington, 243 pp.

O'Donnell, J. 1993. Surface fronts in estuaries: a review. Estuaries 16(1):12-39.

Pearcy , W.G. Ocean ecology of North Pacific salmonids. 1992. University of Washington Press, Seattle, WA, 179

pp.

Percy, K.L., D.A. Bella, C. Sutterlin, and P.C. Klingeman. 1974. Descriptions and information sources for Oregon

estuaries. Sea Grant College Program, Oregon State University, Corvallis, Oregon.

Roegner, G.C., B.M. Hickey, J.A. Newton, A.L. Shanks, and D.A. Armstrong. 2002. Estuarine-nearshore links during

a coastal upwelling-downwelling cycle: plume and bloom intrusions into Willapa Bay, Washington. Limnology

and Oceanography. In press.

Smith, R.L. 1995. The physical processes of coastal upwelling systems. In C.P. Summerhayes, K.-C. Emeis, M.V.

Angel, R.L. Smith, and B. Zeitzschel (eds.), Upwelling in the Ocean: Modern Processes and Ancient Records.

Wiley and Sons, pp. 40-64.

Strub, P.T. and C. James. 1988. Atmospheric conditions during the spring and fall transitions in the coastal ocean off

western United States. Journal of Geophysical Research 93(C12):15,561-15,584.

Strub, P.T. and C. James. 2000. Altimeter-derived variability of surface velocities in the California Current System: 2.

Seasonal circulation and eddy statistics. Deep-Sea Research II 47:831-870.

Strub, T. and C. James. 2002. Altimeter-derived surface circulation in the NE Pacific Gyres: Part I. Seasonal

variability. Progress in Oceanography. In press.

Hickey and Banas 33

Strub, P.T., J.S. Allen, A. Huyer, and R.L. Smith. 1987a. Large-scale structure of the spring transition in the coastal

ocean off western North America. Journal of Geophysical Research 92(C2):1527-1544.

Strub, P.T., J.S. Allen, A. Huyer, R.L. Smith and R.C. Beardsley. 1987b. Seasonal cycles of currents, temperatures,

winds and sea level over the northeast Pacific continental shelf. Journal of Geophysical Research 92(C2):1507-

1526.

Strub, P.T., C. James, A.C. Thomas, and M.R. Abbott. 1990. Seasonal and nonseasonal variability of satellite-derived

surface pigment concentration in the California current. Journal of Geophysical Research 95(C7):1501-1530.

Strub, P.T., P.M. Kosro, and A. Huyer 1991. The nature of cold filaments in the California current system. Journal of

Geophysical Research 96(C8):14743-14768.

Swartzman, G. and B.M. Hickey. 2001. Mesoscale biotic features of the coastal nearshore environment. PICES

Annual meeting, Victoria, Canada.

Thomson, R.E. 1981. Oceanography of the British Columbia Coast. Canadian Special Publications in Fishery and

Aquatic Sciences 86, 291 pp.

Trainer, V.L., N.G. Adams, B.D. Bill, C.M. Stehr, J.C. Wekell, P. Moeller, M. Busman, and D. Woodruff. 2000.

Domoic acid production near California coastal upwelling zones, June 1998. Limnology and Oceanography

45:1818-1833.

Trainer, V.L., B.M. Hickey, and R. Horner. 2002. Biological and physical dynamics of domoic acid production off

the Washington USA coast. Journal of Phycology. In press.

Tully, J.P. 1942. Surface non-tidal currents in the approaches to Juan de Fuca strait. Journal of the Fisheries

Research Board of Canada 5:398-409. Hickey and Banas 34

Tyler, M.A. and H.H. Seliger. 1980. Time scale variations of estuarine stratification parameters and impact on the

food chains of the Chesapeake Bay. In B.J. Neilson, A. Kuo, and J. Brubaker (eds), Estuarine Circulation.

Humana, Clifton, New Jersey.

Valle-Levinson, A. and J. O'Donnell. 1996. Tidal interaction with buoyancy driven flow in a coastal plain estuary. In

D.G. Aubrey and C.T. Friedrichs (eds.), Buoyancy effects on coastal and estuarine dynamics, Vol 53. American

Geophysical Union, Washington, D.C.

Wang, D.-P. 1979. Wind-driven circulation in Chesapeake Bay, winter 1975. Journal of Physical Oceanography

9:564-572.

Venkatesh, S. and W.R. Crawford. 1993. Spread of oil from the Tenyo Maru, off the southwest coast of Vancouver

Island. Natural Hazards 8:75-91.

Werner, F. and B.M. Hickey. 1983. The role of a longshore pressure gradient in Pacific Northwest coastal dynamics.

Journal of Physical Oceanography 13:395-410.

Yankovsky, A.E., B.M. Hickey, and A. Munchow. 2001. The impact of variable inflow on the

dynamics of a coastal buoyant plume. Journal of Geophysical Research 106(C9):19,809-19,824.

Sources of Unpublished Material

MacFayden, A., B.M. Hickey and M. Foreman. 2002. In Prep. The Juan De Fuca Eddy as a source of HABs on the

Washington coast. School of Oceanography Box 355351, University of Washington, Seattle WA 98195-7940.

Rooper, C.N., D. Gunderson, B.M. Hickey, and D.A. Armstrong. 2002. Submitted. Transport of English sole

(Pleuronectes vetulus) pelagic eggs and larvae from coastal spawning grounds to estuarine rearing areas.

Limnology and Oceanography. U.S. National Marine Fisheries Service, Alaska Fisheries Science Center, 7600

Sand Point Way N.E., Seattle WA 98115-6349. Hickey and Banas 35

Siegel, E., J. Newton, B.M. Hickey and N.S. Banas. 2002. In prep. Seasonal variability of water properties in

Willapa Bay, Washington. In Prep. Washington Department of Ecology, Environmental Assessment

Program,300 Desmond Dr. SE, Olympia WA 98504.

Malamud-Roam, K.P. 2000. Muted tidal regimes in marshes of the San Francisco estuary: theory and implications

for ecological restoration. Ph.D. Dissertation, University of California at Berkeley. Berkeley, CA 94720.

Grays Willapa Yaquina Coos Columbia San Fran- South Plum I. Naragansett Chesapeake James Harbora Baya Baya Baya Rivera cisco Baya S.F. Bayb Soundc Baya Baya Rivera area at MSL 150 240 13 34 550 1170 480 7.2–15 430 9900 610 2 AMSL (km ) mean depth 4.3 3.2 2.6 4.0 7.3 6.8 4.4 2.3 10 8.5 5.2 H (m) volume below MSL 0.64 0.76 0.034 0.14 4.0 8.0 2.1 ~0.016 4.3 84 3.2 3 V = H · AMSL (km ) ______mean tidal range at 2.1 1.9 1.8 1.7 1.7 1.3 1.4 2.6 0.9 0.8 0.8 mouth (m) intertidal area — 55d 47e 47e — — — — — — — (% area at MHW) tidal prism volume 46 50 52 31 14 16a–27b 37 ~50 10 2.0 8.6 (% volume at MHW) ______drainage area 7.0 2.9 0.66 1.5 670 120 — 0.58 4.6 180 26 (1000 km2) monthly-mean riverflow R (m3 s-1) lowest mo. flow 56 17 ~0 2.8 4200 330 — ~0 34 950 150 highest mo. flow 880 390 68 190 10000 1800 — 10 170 4200 600 river-filling time V/R (d) lowest mo. flow 100 500 long 600 10 300 — long 2000 1000 200 highest mo. flow 8 20 6 8 5 50 — 20 300 200 60 ______aNOAA 1985; bMalamud-Roam 2000; cJay et al. 1997; dAndrews 1965; ePercy et al. 1974.

Table 1. Indices of morphology, tidal forcing, and river input for the four PNCERS estuaries and seven others on the U.S. East and West Coasts.

Hickey and Banas 36

Figure Captions

Figure 1. Schematic of the California Current System. Adapted from Hickey and Royer 2001.

Figure 2. Schematic of wind-driven coastal circulation in the PNW. The cartoon illustrates the offshore (onshore)- directed surface currents that occur in response to an upwelling (downwelling)-favorable wind stress and upwelling

(downwelling) along the coast. Freshwater flows from coastal estuaries and from the Strait of Juan de Fuca are illustrated with darker shading. The location of a persistent summertime mesoscale feature (the Juan de Fuca Eddy) is also shown.

Figure 3. Satellite-derived sea surface temperature in the PNW. The figure illustrates enhanced upwelling along the

Washington coast and enhanced upwelling in the lee of the northern headland along the Washington coast (shown as white arrow). The Juan de Fuca eddy is readily apparent as a coldwater feature opposite the strait of Juan de Fuca.

Figure 4. Seasonal variation of selected nutrients in the PNWat mid shelf off Washington and off Oregon. Adapted from Landry et al. 1989.

Figure 5. Annual cycle of the vertically-integrated chlorophyll over the Washington and Oregon shelves. Adapted from Landry et al. 1989. The data illustrate the typically higher chlorophyll on the Washington shelf.

Figure 6. Topography of the PNW illustrating important canyons and banks.

Figure 7. Particulate domoic acid in pseudo-nitzschia species on the Pacific West Coast in 1998. Maximum concentrations of domoic acid and toxic species are indicated to the right of each area of high toxin. Each of these areas is associated with relatively retentive circulation patterns. Adapted from Trainer et al. 2000.

Figure 8. Modeled circulation at depths ranging from 50 m to 600 m showing cyclonic eddies over and within two submarine canyons off the Washington coast (Dinneman and Klinck pers. comm.). Note that at 50 m the flow is Hickey and Banas 37

relatively undisturbed by the canyon topography. The circulation was forced by an upwellling-favorable wind stress with a magnitude typical for this region.

Figure 9. Modeled response of the Columbia plume in summer to changes in wind direction. The figure illustrates the evolution of surface salinity (psu) for southward ambient flow conditions in response to 6 days of downwelling favorable winds, followed by 6 days of upwelling favorable winds at (a) 13 days, (b) 15 days, (c) 16 days, (d) 19 days, (e) 21 days and (f) 25 days with a southward ambient flow of 10 cm s-1. Winds change direction after 19 days immediately after (d). The distance between tick marks is 20 km. Adapted from Garcia-Berdeal et al. 2002.

Figure 10. Modeled velocity structure of the Columbia plume illustrating potential retentive areas (eddy-like features) and frontal jets associated with the river plume. Surface salinity (psu) contours and surface velocity vectors

(m s-1) at t = 28 days for (a) northward ambient flow of 10 cm s-1 and (b) southward ambient flow of 10 cm s-1.

River discharge for both cases is 7000 m3 s-1. Adapted from Garcia-Berdeal et al. 2002.

Figure 11. Map of the Pacific Northwest coast from Washington to Northern California, showing the location of the four PNCERS estuaries and other estuaries in the region. Maps of Grays Harbor, Willapa Bay, and Coos Bay, with the locations of estuarine and offshore moorings are also shown.

Figure 12. Time series of temperature at selected sites in Willapa estuary illustrating up-estuary propagation of an upwelling-driven signal. Time series of salinity in the estuary and alongshore wind on the nearby coast illustrate the response of the estuary water properties to coastal upwelling events-high salinity during periods of upwelling- favorable winds and low salinity during periods of downwelling-favorable winds, with a~1.5 day lag between wind and estuary salinity.

Figure 13. Schematic illustrating baroclinic coupling between the coastal ocean and a coastal plain estuary in the

Pacific Northwest during upwelling and downwelling events for a low riverflow, summer period in an Eastern

Boundary System. From Hickey et al. 2002.

Figure 14. (a) Time series of temperature in three estuaries in the Pacific Northwest, illustrating simultaneous response to large scale upwelling/downwelling along the open coast during spring-fall, 1999. All data are from the Hickey and Banas 38

lower water column and from stations near the mouth of each estuary (see locations in Figure 11). Solid line along the x-axis indicates interval expanded in (b,c,d). (b) Time series of temperature in Grays Harbor and Coos Bay.

Arrows indicate downwelling events (warmer water) prominent only in the northern estuaries. (c) Time series of alongshore wind at latitudes close to Grays Harbor and Coos Bay. Arrows illustrate wind events that cause downwelling seen in panel c. (d) Time series of salinity for the same period, illustrating effects of more persistent upwelling favorable winds at more southern locations.

Figure 15. (a) Time series of the north-south component of nearshore wind from late winter to early summer, 2000.

The dates of the two CTD transects of Willapa Bay shown in Figure 16 are indicated. (b) Salinity and (c) the local along-channel salinity gradient near the mouths of Willapa Bay and Grays Harbor during a three-week period Apr-

May, 2000, showing a brief upwelling event, an intrusion of the Columbia River plume, and a recovery from that intrusion. In (b), both 30-min and subtidal (48-hr-Butterworth-filtered) data are shown. Dots mark times of high slack water in Willapa Bay. In (c), the difference between high-slack and low-slack salinity divided by the tidal excursion for each semidiurnal tidal cycle has been filtered as above to provide a subtidal, single-station time series of the along-channel salinity gradient.

Figure 16. Salinity from CTD transects along the main channel of Willapa Bay on (a) May 3, 2000, near the end of a

Columbia River plume intrusion, and (b) May 30, 2000, during strong upwelling event that replaces plume water

(~21.5 psu) with much saltier water (≥ 29 psu). A reversal of the along-channel salinity gradient is marked in (a).

Triangles at the top of the salinity sections give the location of CTD casts. Tidal stage and transect route are also given for each section. These surveys were made within a five-month wind time series in Figure 15a.

Figure 17. (a,b) Temperature and salinity from a CTD transect along the main channel of Grays Harbor June 11,

1999 during a time of strong solar heating. Triangles at the top of the sections give the location of CTD casts. (c)

Temperature-salinity profile of the along-channel transect (lines) and CTD casts on shoals adjacent to the channel

June 11-15 (dots). Location and tidal stage of bank and channel surveys are also shown. Dots on inset maps indicate location of bank stations.

Hickey and Banas 39

Figure 18. Temperature-salinity diagrams for surveys of Willapa Bay during (a) June and (b) July 2000, showing the hydrographic signature of direct solar heating. Line segments represent CTD casts within the main channels of the estuary; dots represent water on banks with depths < 5 m.

Figure 19. Salinity from CTD transects on July 14, 1999 (a) along the main channel of Willapa Bay and (b) from the channel to shore across a shallow, narrow bank. Vertical line in (a) near 18 km marks the location of the cross bank transect in (b). Location and tidal stage are shown; triangles at the top of the sections give the location of CTD casts.

The nearly identical temperature-salinity profiles of the along channel and cross channel transects are confirmed with a T-S diagram in (c).

Figure 1. Storms Fair Weather N N

J. de Fuca Eddy Juan de Fuca Strait Juan de Fuca Strait

downwelling water upwelling water ◆warmer ◆colder ◆less saline ◆more saline ◆nutrient- ◆nutrient- reduced enriched fresher Grays Grays Harbor Harbor Near-Surface Near-Surface Shelf Flow Shelf Flow Willapa Bay Willapa Bay winter colder plume Columbia plume Columbia Columbia River summer River warmer plume

Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. Figure 7. 50m, Day 10.0 100m, Day 10.0

46.8° 20 cm s -1 20 cm s -1 N N

46.4°

46.0° N

150m, Day 10.0 250m, Day 10.0

46.8° -1 20 cm s 20 cm s -1 N N

46.4°

46.0° N

400m, Day 10.0 600m, Day 10.0 46.8° -1 -1 20 cm s Guide 20 cm s Canyon

N N

Willapa Canyon 46.4°

Astoria Canyon

46.0° N

125° W 124° 125° W 124°

Figure 8. Figure 9. Figure 10. Grays Harbor

47°00'

GH Puget Sound GHOS Grays Harbor 46°50' Willapa Bay Columbia R.

W3 46°40' Yaquina Bay

Coos Bay W6 46°30' N Willapa Bay

46°20' 124°20' 124°00' 123°40'

43°30'

TD CR 43°20' S. F. Bay Coos Bay

CBOS 43°10'

124°40' 124°20'

Figure 11. Along Estuary Temperature 20

W3 ■ 18 ■

C) W4 W1 ■ O W5 W5 16 W4 ■ W3 14 W1 emperature ( T 12 up-estuary propagation ~12 cm s-1 10

Wind Stress vs. Near Mouth Salinity

34 1.5 )

S at W1 r = -0.64 -2 τ 32 1 0.5 30 0 28 -0.5

26 -1 Wind Stress (dynes cm

Salinity (psu) Adjusted by 1.5 d Salinity (psu) 24 -1.5 150 160170 180 190 200 210 220 Time (calendar day)

Figure 12. Figure 13. 18

16 Willapa Grays C) O 14

12 emperature ( T 10 Coos June Sept 8 100 150 200 250 300

15 14 C) O 13 Grays 12 11 emperature ( T 10 Coos 9 8 130 140 150 160 170 180 190 200

10

) Grays

s-1 5 Downwelling

0

-5 Upwelling -10 Coos Alongshore Wind (m

-15 130 140 150 160 170 180 190 200 Longer upwelling in Coos 35

Coos

30

Upwelling Grays

25 Salinity (psu)

downwelling 20 130 140 150 160 170 180 190 200 Calendar Day 1999

Figure 14. 3 30 -1 AY

AY N-S wind (m s )

from the south (downwelling-favorable) M 10 M

0

from the north (upwelling-favorable) -10 Feb Mar AprMay Jun Jul 2000 (a) Salinity (psu) 32 at high slack Willapa Bay subtidal 28 Grays Harbor

24

20

04/16 04/20 04/24 04/28 05/02 05/06 (b) UPWELLING PLUME INTRUSION WIND RELAXATION

Along-channel salinity gradient (psu/km) 0.4

Willapa Bay

0.2 Grays Harbor

0

-0.2 04/16 04/20 04/24 04/28 05/02 05/06 (c)

Figure 15. May 3, 2000 Willapa Bay 0

21.4 19 18 Tidal height at W6 21

21.5

21.8 21.5 20 (MLLW; m) 21.5 4

21.4

21.8

10 22 2 Salinity (psu) Depth (m) 0 0h 12h 0h gradient reversal 0 16 2024 28 32 Local Time 20 0 10 20 30 40 10 km Distance from mouth (km) (a)

May 30, 2000 0 21.5

23 18

22 20

17 19 4

21

26 21.5

24 25 2

27

10 22

Depth (m) 0 28 0h 12h 0h

29 Salinity (psu) Local Time 20 0 10 20 30 40 Distance from mouth (km) (b)

Figure 16. 0 0 12

6

8 16 10 20 4 15 4 22 18 14 24 16 14 8 8

12 13 26 Depth (m) Depth (m)

12 13 12

Temperature (°C) Salinity (psu) 16 16 0 10 1214 16 18 0 16 2024 28 32 20 20 0 10 20 30 0 10 20 30 (a) Distance from Mouth (km) (b) Distance from Mouth (km)

4 Jun 11 20 Grays Harbor 2 channel 18 channel 0 banks banks 0h 12h 0h (depth < 5 m) 16 4 Jun 13

2 14 0 0h 12h 0h Temperature (°C) 12 10 km 4 Jun 15

10 2 Tidal height at mid-estuary (MLLW; m) 5 10 15 20 25 30 0 0h 12h 0h Salinity (psu) (c)

Figure 17. June 26-28, 2000 July 24-25, 2000 20 20

18 18

16 16

14 14 channel banks Temperature (°C) Temperature (°C) 12 12 (depth < 5 m) 10 km 10 10

22 24 26 28 30 32 22 24 26 28 30 32 (a)Salinity (psu) (b) Salinity (psu)

Figure 18. Jul 14, 1999

0

24 25 Tidal height at W6

26 27 (MLLW; m) 30 4

28

31 29

10 31.5 2

Depth (m) 0 0h 12h 0h

Salinity (psu) 20 10 20 30 10 km (a) Distance from mouth (km)

0 16 20 24 28 32

0 Tidal height at W6 1 km (MLLW; m) 26 25 4 29 28 27 5 29.5 2

Depth (m) Salinity (psu) 10 0 -1600 -1200 -800 -400 0 0h 12h 0h Local time (b) Distance from eastern shore (m)

20

18

16

14

Temperature (°C) 12 channel (a) bank (b) 10

22 24 26 28 30 32 (c) Salinity (psu)

Figure 19.