The Future of Fish in Response to Large-Scale Change in the San Francisco Estuary, California

Peter B. Moyle* Center for Watershed Sciences and Department of Wildlife, Fish, and Conservation Biology, University of California 1 Shields Avenue, Davis, California 95616, USA

Abstract.—The San Francisco Estuary is the largest estuary on the west coast of North America and one of the most altered. It supports a diverse fish fauna in which key species are in severe decline. The estuary is faced with catastrophic structural and ecological changes, especially in the Sacramento–San Joaquin Delta and Suisun Marsh, as the result of levee failure caused by the combination of earthquakes, land subsidence, sea level rise, and increased high outflows events (from climate change). The resulting flooding of Sacramento–San Joaquin Delta islands and Suisun Marsh is predicted to disrupt California’s water supply system and, consequently, the state’s economy. From a fish perspective, the changes seem likely to create conditions in which key native species and striped bass can persist at least at present low levels, after a period of possible high mortality created by the initial flooding events. Taking actions to regulate ecological changes in the estuary before a disaster occurs could actually improve conditions for desirable fishes while being highly compatible with delivering services the Sacramento–San Joaquin Delta provides, especially water supply. Specific actions include improving habitat for fish in Suisun Marsh, Cache Slough, the Yolo Bypass, and the San Joaquin River, while creating islands in which flooding can be managed. The key is increasing and maintaining habitat heterogeneity. The present situation in the estuary represents an unprecedented opportunity to reverse the impacts of more than 150 years of undesirable ecological change.

Introduction changes to the SFE are imminent (Mount and Twiss 2005; Lund et al. 2007). The changes The San Francisco Estuary (SFE) is the are likely to be most dramatic in the upper largest estuary on the west coast of North part of the estuary, the Sacramento–San Joa- America and one of the most altered (Nich- quin Delta (the Delta), where large-scale le- ols et al. 1986). It is highly urbanized but vee failure can seriously disrupt local, region- contains extensively diked agricultural lands al, and state economies. A principal concern and managed wetland habitats. It is also is disruption of California’s water distribution highly invaded by alien species, especially the system. Much of the freshwater used by San aquatic habitats. Not surprisingly, the native Joaquin Valley farms and the vast urban areas species of plants and animals have declined of Southern California originates directly or in abundance; several are extinct and others indirectly from the estuary’s inflowing rivers. are listed as threatened or endangered under This water is removed from the Delta by the state and federal laws (Herbold et al. 1992). massive pumps of the State Water Project and Human-caused changes to the SFE are still the federal Central Valley Project. Additional taking place at an accelerated rate and there water is removed to supply aqueducts to cit- are strong indications that major, catastrophic ies around San Francisco Bay and to irrigate farms in the Delta. Large scale levee failure * Corresponding author: [email protected] could also eliminate farming on thousands of moyle acres of Delta islands, threaten urban areas, nez Strait (Suisun Bay), and the river conflu- and disrupt railroads, pipelines, and other ence at Chipps Island (Delta), respectively. infrastructure (Lund et al. 2007). Likewise, a These narrows have allowed the three regions sudden catastrophic change to the Delta and to have distinct identities, emphasized by hu- SFE will have substantial ecological impacts. man modifications to them. The Delta is per- A major question being asked by manage- ceived as a region where freshwater from the ment agencies and regional stakeholders is rivers tidally sloshes back and forth in leveed “how can we prevent large-scale change from channels, flowing between islands of agricul- taking place in the SFE, especially the Delta?” tural fields. The islands are highly subsided Answering precursors to this question (mainly, (many are 5+ m below sea level; Figure 2), sur- how do we protect endangered fish and fish- rounded by 1,800 km of fragile levees made eries?) was one of the reasons for the estab- of local materials, often peat, with some rein- lishment of CALFED in 1996, a massive joint forcement with rip rap (Brown and Michniuk state-federal management and research effort 2007). Historically, the Delta was a vast marsh- (http://calwater.ca.gov/), which has been land that was flooded annually by undammed criticized for not quickly resolving problems rivers (Lund et al. 2007). of the SFE (Little Hoover Commission 2005). In contrast, Suisun Bay was, and still is, a A report produced by the University of Cali- large area of open water that is transitional fornia, Davis and the Public Policy Institute of between the fresh waters of the Delta and the California (Lund et al. 2007) turned the origi- salt waters of San Francisco Bay; it is a shallow nal question on its head, asking instead “How region of wind-stirred, brackish water, lined can the Delta be managed to accommodate with tidal marshes. The largest of these marsh- large-scale change before undesirable changes, nearly as large as Suisun Bay itself, is Su- es are forced by catastrophic events?” In this isun Marsh. This 30,000+ ha marsh is largely paper, key findings of Lund et al. (2007) are managed today as freshwater wetlands to sup- summarized in relation to aquatic organisms, port waterfowl hunting in both private clubs especially fish. I first describe the SFE, provide and public wildlife areas. The key for main- a brief introduction to its fish fauna, and then taining its freshwater character is inflow from discuss the major drivers of change. I then de- the Sacramento River via Montezuma Slough. scribe what is likely to happen to key fish spe- Montezuma Slough has large tidal gates on its cies if present management trends continue, upper end that control salinity in the marsh followed by suggestions for major actions that by allowing freshwater to flow in but prevent- could be taken to improve the SFE for fishes ing the tides from pushing it back out again. even in the face of large-scale change. More than 360 km of levees separate marsh islands from the tidal channels, in which water The San Francisco Estuary is still seasonally brackish. The channels are highly productive of fish, which are a mixture The SFE is the outlet of the Sacramento of euryhaline freshwater and marine species and San Joaquin rivers, which in turn drain (Matern et al. 2002). much of central California. The primary The marine species in Suisun Marsh source of water for the rivers is the Sierra Ne- come from San Francisco Bay, which is largely vada, which intercepts moisture-laden clouds a marine system, with variable but high salini- coming off the Pacific Ocean. The estuary has ties; the actual salinity depends on location three distinct segments, San Francisco Bay and season. San Francisco Bay is ringed by cit- (including San Pablo Bay), Suisun Bay and ies, and its fringe marshes are fragments of its Marsh, and the Sacramento–San Joaquin Del- original tidal marsh system. ta (Figure 1). Each segment has a confined All three parts of the SFE were once more outlet through which the tides surge back and variable in their salinities and river-driven forth, creating complex hydrodynamics: the hydrodynamics than they have been for the Golden Gate (San Francisco Bay), the Carqui- past 50–60 years (Bay Institute 1998). Dur- san francisco estuary fish

Figure 1. Map of San Francisco Estuary. (Moyle 2002) ing wet years, the spring snowmelt from the and the State Water Project, however, allowed Sierra Nevada could temporarily make fresh freshwater inflows to be regulated, reducing the surface waters of San Francisco Bay, while variability in water quality so that, for the pur- during late summer of drought years, ocean poses of policy and public perception (if not salt could be detected at the upper ends of always in reality), the Delta and Suisun Marsh the Delta, especially after upstream diversions became permanent freshwater systems, Su- became established (DWR 1993). The devel- isun Bay became a brackish water system, and opment of the federal Central Valley Project San Francisco Bay became an exclusively ma- moyle

Figure 2. Map of the Sacramento–San Joaquin Delta showing land subsidence. Courtesy of Public Policy Institute of California, after DWR (1993). Recent surveys indicate that subsidence in most areas is 1–5 ft greater than shown. (Department of Water Resources, unpublished data) rine system. The two water projects (and other from the southern edge of the Delta, for agri- related projects) constructed huge dams on cultural and urban use. the Sacramento and San Joaquin rivers and all For further information on the environ- their major tributaries in the 1930s–1960s, with mental and ecological history of the SFE, see consequences such as increasing summer flows Herbold et al. (1992), Hollibaugh (1996), of the Sacramento River and drying up part of Bay Institute (1998), Lund et al. (2007), or the San Joaquin River. The dams also allowed http://www.deltavision.ca.gov. The rest of for the regulation of salinity in the upper SFE. this essay focuses primarily on the Delta and By releasing large quantities of freshwater, espe- Suisun Bay and Marsh because the lowermost cially in the summer, the dam operators could part of the SFE, San Francisco Bay, has a whole both keep saltwater out of the Delta and Suisun additional set of issues related to its intense Marsh and permit the pumping of freshwater urbanization. san francisco estuary fish The Fishes Drivers of Change The San Francisco Estuary has a high di- The major drivers of change in the SFE versity of fishes, representing marine, fresh- that are likely to result in major shifts in envi- water, anadromous, and estuarine species, ronmental conditions, including catastrophic both native and alien (Matern et al. 2002; shifts, are earthquakes, island subsidence, sea Moyle 2002). The total fish fauna is about level rise, climate change, and invasions of 120 species that can be found in one envi- new alien species (Lund et al. 2007). Human ronment or another on a fairly regular basis, land and water use could arguably be listed of which about 30 (25%) are aliens, mostly as another driver of change, but these uses in freshwater and brackish water. About 75 are strongly affected by the first five drivers so species, largely marine, have been recorded will not be treated further as drivers here. The from San Francisco Bay in recent years, of major catastrophic consequence of the five which only 5 are alien species (Schroeter drivers is extensive levee failure in the Delta and Moyle 2006). In Suisun Marsh and Bay, and Suisun Marsh. From an ecological per- 53 species are known, a mixture of euryha- spective, the consequence is sudden change line marine, freshwater, and anadromous in the hydrodynamics of the two regions as species, plus a few more stenohaline species the islands fill with water, creating new habitat that appear during periods of high or low sa- conditions, followed by invasions of both de- linity. They represent 28 native species and sirable and undesirable species into the new 25 aliens (Matern et al. 2002). In the Delta, habitat space. there are about 46 regularly occurring species, a mixture of freshwater and anadro- Earthquakes mous fishes, of which 27 are aliens. There are at least five major faults in the The introduction of alien species has ac- SFE region, but there have been no major companied large-scale environmental change earthquakes in the region since the great 1906 and the explosive invasion of many of the San Francisco earthquake. This means that alien species has been a driver of declines of pressure is building on local faults, steadily native species, including extinctions of na- increasing the probability that one will move tive fishes such as thicktail chub Gila crassi- as time goes by (Mount and Twiss 2005). The cauda and Sacramento perch Archoplites in- major impact likely from earthquakes is col- terruptus (Moyle 2002; Light and Marchetti lapse of levees in the Delta and Suisun Marsh 2007). Changes over the past 50 years, since because of their poor foundation soils and the advent of the major water projects, con- weak construction. tributed to severe declines of most native species, including four runs of Chinook salmon Island Subsidence Oncorhynchus tshawytscha and the delta smelt Hypomesus transpacificus. The result has been The islands of the Delta were originally that five fishes, including delta smelt and two marshlands on a thick base of peaty soils. runs of Chinook salmon, are currently listed More than 182,000 ha of islands were diked as threatened or endangered by state and fed- and drained for farming in the 19th and early eral governments (Moyle 2002). In more re- 20th centuries, and soils were typically burned cent years, declines in fisheries have also been annually to release nutrients from the peat, a major concern, especially fall-run Chinook causing the interiors of the islands to subside salmon, white sturgeon Acipenser transmon- rapidly. Even after the practice of burning tanus, and alien striped bass Morone saxatilis. stopped, subsidence continued through oxi- Some of the fishes most likely to be affected dation of plowed soils and the loss of soil by by future large-scale changes to the SFE and wind erosion. As a result, all islands with peat also likely to drive policy decisions are listed soils used for farming have subsided, with in Table 1. subsidence greatest (3–7 m below sea level) moyle

Table 1. Fish species of major importance for management in the San Francisco Estuary. (Moyle 2002) Species Native? Why important Comments Delta smelt yes Threatened species, Open water pelagic, near Hypomesus transpacificus endemic extinction Longfin smelt yes Special concern species Open water pelagic, Spirinchus thaleichthys anadromous Splittail yes Special concern species, Spawns on flood plains, Pogonichthys macrolepidotus endemic rears in brackish water Tule perch yes Native, declining Represents complex of Hysterocarpus traskii native fishes Sacramento perch yes Extirpated Reintroduction program Archoplites interruptus proposed Striped bass no Estuarine sport fish Anadromous, uses entire Morone saxatilis declining estuary White sturgeon yes Sport fish, declining Anadromous, rears in Acipenser transmontanus estuary Green sturgeon yes Endemic, threatened Anadromous, rears in A. medirostris ocean Chinook salmon yes Four runs Anadromous Oncorhynchus tshawytscha Chinook salmon, yes threatened Appears in estuary spring run mainly as smolts Chinook salmon, yes endangered Appears in estuary winter run mainly as smolts Chinook salmon, yes Sport and commercial Large numbers of fry in fall run fisheries, hatchery- estuary but survival low driven Largemouth bass no Alien predator, sport fish, Represents complex of Micropterus salmoides increasing alien pond fishes in Delta Pacific herring yes Commercial fish Represents euryhaline Clupea pallasii marine fishes

in west and central Delta (Figure 2). Despite 50 years is plausible (Mount and Twiss 2005). improved soil conservation practices, subsid- Higher mean sea levels result in higher high ence increases as long as farming continues. tide levels, increasing the probability that le- The effect of subsidence is to create a series vees in the Delta and Suisun Marsh will over- of bowl-like depressions, into which water will top and then collapse, especially if combined pour in if it breaks through the levees. with flood flows coming down the rivers.

Sea Level Rise Climate Change Sea level is rising in the SFE and has been The climate of California is becoming signifi- for at least 6,000 years (Lund et al. 2007). cantly warmer, a trend that is likely to contin- Because of global warming, the rate of rise is ue for some time (Dettinger 2005). While av- accelerating. There is scientific debate about erage precipitation is not expected to change how rapidly and how much sea level will con- much, more will fall as rain and less as snow in tinue to rise, but a 30–50 cm rise in the next the high mountains, resulting in more periods san francisco estuary fish of severe high run-off. Year-to-year variability a Bay–Delta Conservation Plan process, and in rainfall is also expected to increase, increas- (4) numerous other actions and processes ing the frequency of extended droughts and by agencies at all levels of government. The big floods (Dettinger 2005). One result of this scenario of most concern is simultaneous and change is increased hydrostatic pressure on cascading failures of levees throughout the levees during storms and floods and thus in- Delta because of the impact of such failures creased likelihood of failure. on Southern California’s water supply, on agriculture and other uses of Delta islands, Invasive Species and on urban areas in and around the Delta The SFE has the reputation of being the (Lund et al. 2007). Here, however, I discuss most invaded estuary in the world and new in- mainly the impacts on the ecosystems of the vasions continue at a high rate (Cohen and Delta and Suisun Marsh, especially with re- Carlton 1998). Recent invaders (e.g., overbite spect to fish. clam Corbula amurensis, Brazilian waterweed Egeria densa) have already had major impacts Delta on ecosystem structure and function (Kim- For the central and western Delta, the ba- merer 2002; Nobriga et al. 2005). New invad- sic disaster scenario is that following multiple ers or expanding populations of existing in- levee failures, water would rush in, filling as vaders are likely to take advantage of the new much as 2.5 3 109 m3 of space in the island habitats created by large-scale levee failure basins. If the levee collapses occurred as the (Light and Marchetti 2007), further exacer- result of the combination of high outflows bating the effects of levee failure and increas- and high tides, the islands would likely fill ing the difficulty of protecting native species. mostly with freshwater. If the levee collapses occurred as the result of an earthquake dur- Ecological Effects of Large-Scale ing a low-flow period, much of the water fill- Change ing the islands would be drawn up from Su- isun Bay and even San Francisco Bay, creating The likelihood is high that two or more of lagoons with varying degrees of salinity. the above drivers of change will act together The open-water habitat thus created would to create catastrophic levee failure within the be up to 10 m deep and subject to strong tidal next 50 years (Mount and Twiss 2005; Lund currents through the new levee breaches, as et al. 2007), assuming the SFE continues to well as to mixing from frequent winds. The be managed as it is today. The probability of hydrodynamic and salinity regimes of each such an event is high enough that it is pre- flooded island would depend on the- num sumably more a matter of when and how ber and location of levee breaches, closeness much rather than if. In recognition of this, of the area to Suisun Bay (source of saltwa- the Delta Risk Management Strategy (DRMS) ter) and to inflowing rivers, and relationship team of the California Department of Water to infrastructure such as the ship canal that Resources has modeled the effects of up to 50 goes through the system. As levees continued simultaneous levee breaches on Delta islands to erode, many of the flooded islands would (http://www.drms.water.ca.gov). Other signs presumably come to resemble Frank’s Tract, a of high levels of interest include (1) the ap- large island in the central Delta that flooded pointment by Governor Arnold Schwarzeneg- in the 1930s and was left that way. It is current- ger in February 2007 of the Delta Vision Com- ly a freshwater lagoon with complex hydrody- mittee with a Blue Ribbon Task Force to find namics that is dominated in summer by dense ways to prevent or reduce the impacts of the growths of alien Brazilian waterweed. likely disaster, (2) the recent passage of bond Presumably, the flood water would initially issues to fix levees and other infrastructure af- be highly turbid from the disturbance of peat fecting urban areas, (3) the establishment of and sediment on the islands (DWR 2007), moyle but once the suspended material had settled, on high tides and high river flows. A scenario there would likely be blooms of algae, espe- of wide-scale levee failure in the Delta is likely cially when temperatures are warm, because of to include levees protecting Van Sickle Island, the release of nutrients from the soils and the located at the entrance to Montezuma Slough, increased water transparency. Depending on the main artery of the marsh. Van Sickle Island the species making up the algal blooms (dia- is already subject to frequent levee failures, al- toms versus green algae versus cyanobacteria), though failures are quickly repaired. The rapid a bloom of zooplankton should quickly follow. repair response (by public agencies) occurs in Within a year, brackish island lagoons would be good part because if the island floods, the en- heavily colonized by the overbite clam, an alien tire freshwater distribution system of the cen- species that currently dominates the benthos tral marsh (Roaring River Slough) ceases to of Suisun Bay. Presumably, the clams would work efficiently and the southern third of the then consume much of the primary and sec- marsh (between Montezuma Slough and Su- ondary production of the new lagoons, as they isun Bay) becomes tidal and brackish. Future do in Suisun Bay, reducing zooplankton popu- failures will likely make much of the central lations. Island lagoons that contained fresh or marsh tidally brackish. Less dramatically, in- low salinity water, would likely be colonized in creasingly high tides (from sea level rise) and 1–2 years with the species that dominate similar increasingly large flood events (from climate areas in the Delta: Brazilian waterweed and, in change) are likely to cause levee overtopping areas with sufficient flow, Asian clamsCorbicula and failures within the Marsh at increasing fluminea. The combination would result in la- rates. Thus, broadly, the ultimate fate of much goons choked with weeds and low zooplankton of Suisun Marsh is to become a tidal brackish populations (similar to Frank’s Tract today). It marsh, with seasonally higher salinities in many is possible that flooded islands located close to areas than were historically present. Much of sources of both freshwater inflow (Sacramen- it is likely to be permanently inundated. The to River) and tidal sources of saltwater could details of how the future marsh will actually maintain a pool of water that would fluctuate look will depend on the interactions of a num- enough in salinity on either an annual or in- ber of factors: (1) the rapidity and extent of terannual basis to keep the overbite clam or sea level rise, (2) the depth of tidal and other Brazilian waterweed-Asian clam from becom- flooding, (3) the residence time of the water in ing dominant. Biomass production in such la- different areas (i.e., the hydrological relation- goons could revert to a pelagic system of phyto- ship between the flooded areas and the deeper plankton, zooplankton, shrimp, and fish, such channels and sloughs that drain them), (4) the as was the case of Suisun Bay before the inva- response of the natural vegetation to the inun- sion of overbite clams in the 1980s. Obviously, dation and salinity gradients, and (5) the influ- these scenarios can all be strongly affected by ence of existing artificial dikes and channels, local conditions of wind, tide, and river, as well including railroad and road beds. This future as the diverse configurations of the lagoons, marsh, however, will certainly have a mosaic of which will change constantly as levees deterio- habitats centered around tidal brackish water rate after the initial breaches. marsh. Much of the new marsh will be drained by channels that should gradually recover their Suisun Marsh historic dendritic nature and be kept open by strong tidal action. Suisun Marsh, for the most part, has not subsided as much as the Delta, although much of it is 0–2.3 m below sea level and parts of it Effects on Fishes of Large-Scale (ca. 16%) are more than 2.3 m below sea level Change (C. Enright, Department of Water Resources, personal communication). Before European In the broadest sense, the creation of settlement, it was high marsh, mostly flooding more aquatic habitat in the Delta and Suisun san francisco estuary fish

Marsh will result in a net increase in numbers would result from filling process, given the and biomass of fish, once the initial flooding volumes of water involved and the nature of period is past. The important question is what the matter (organic matter, mainly peat par- will happen to the species that people care ticles) most likely to be suspended. The fill- most about (Table 1). These are native species ing could be most disastrous for delta smelt that are listed as threatened or endangered or if they were spawning because it would draw are in severe decline or fishes that support them away from suitable spawning areas and fisheries. The effects suggested in the follow- would likely create hydrodynamic conditions ing accounts are highly speculative but based (e.g., diminished tidal range in channels) that on extensive knowledge of the fishes, which would make return difficult. Likewise, surviv- are well studied (Moyle 2002; see also Moyle ing larval smelt would likely find unfavorable et al. 2004; Bennett 2005; Nobriga et al. 2005; conditions for feeding in the newly filled is- Williams 2006). lands and could starve before large populations of microzooplankton (especially rotifers Delta Smelt and copepods) developed. Assuming massive blooms of toxic algae The single most important species from (e.g., Microcystis spp.) do not occur and that the viewpoint of affecting management of stockpiles of toxic materials have been re- water in the Delta is the delta smelt, which is moved before flooding, a month or so after listed as threatened by both state and federal island filling and hydrodynamic stabilization, governments and is on the verge of extinc- conditions for plankton feeding fish such as tion (Bennett 2005). It has a 1-year life cycle, smelt should start becoming favorable, with is a pelagic planktivore, and is endemic to the development of blooms of one or more the SFE; it spawns in the Delta and rears in species of copepods and/or cladocerans. Del- Suisun Bay and Marsh (Moyle 2002; Bennett ta smelt that survived up to this period in the 2005). Under multiple levee breach scenari- islands should then find conditions extremely os, most delta smelt will likely be sucked into favorable for growth and survival, especially in the rapidly filling islands, whether they were islands that maintained salinities of less than upstream spawning in the Delta, downstream 2 mg/L and temperatures of less than 208C. rearing in Suisun Bay, or moving between Thus, the impact of a large-scale levee breach these habitats. A few delta smelt might be able event on delta smelt depends in good part on to avoid displacement if they were located in the timing of the event. Presumably, a higher the distant peripheral habitats such as the proportion of the population would be able mouth of the Napa River, Montezuma Slough, to survive an event in July–November (when or Cache Slough in the north Delta (Hobbs they are rearing in Suisun Bay) than in De- et al. 2007). The DRMS study (DWR 2007) cember–June (when they are spawning in the predicts that many, if not most, fish sucked Delta). into the flooding islands will die of stress, es- In the long run, however, permanently pecially that created by particulate matter in flooded islands in the right place could in- the water, which can abrade gills, and by high crease the amount of favorable habitat for turbidity. The delta smelt, as a small (<9 cm delta smelt. If a flooded island had conditions total length [TL]) delicate, midwater, visual (such as those that exist in Nurse Slough in feeder, would seem especially vulnerable to Suisun Marsh) that excluded dominant inva- these conditions. However, data are lacking to sive benthic species, it would likely become support the DRMS flood mortality hypothesis. highly productive pelagic habitat, habitat that Previous levee breaches on single islands have is apparently in short supply for smelt at times not been accompanied by reports of fish kills, today (Bennett 2005; Hobbs et al. 2006). Del- but no one was looking in the haste to repair ta smelt would presumably also benefit from the levees and pump out the islands. It seems a flooded Suisun Marsh as rearing habitat, if unlikely, however, that a complete fish kill flooding increased productivity of intersect- 10 moyle ing channels, especially Montezuma and Su- Suisun Marsh would presumably also benefit isun sloughs, and salinity fluctuations at the longfin smelt. right scale of time and space reduced the impacts of invasive species. Striped Bass One indirect positive effect on delta smelt of large-scale island flooding would be that the With high fecundity, iteroparity, large large pumps of the State Water Project and size, and a life span of 20+ years, nonnative the federal Central Valley Project in the south striped bass have a high capacity to survive Delta would be shut down for long periods of environmental disasters. Nevertheless, they time. This would happen because of salt wa- have suffered a long-term decline in the SFE, ter at their intakes as the result of flooding although they still support a valuable fishery islands drawing in salt water from Suisun Bay (Moyle 2002). In the SFE, striped bass migrate and beyond. Because, in some years, pumping 125–200 km upstream to spawn in the Sacra- from the two plants can negatively affect delta mento River in late April–early June. The em- smelt populations through entrainment and bryos drift downstream and hatch about the other effects (Bennett 2005), shutting down time they reach Suisun Bay, where the larvae the pumps will remove one potential major rear at low salinities. Juveniles rear through- source of mortality, perhaps compensating for out the estuary but seem to be most abundant some of the flooding mortality. in Suisun Marsh and Bay, where they feed on zooplankton and small fish. By the time they Longfin Smelt exceed 10 cm TL, they have largely switched to feeding on small fish (Feyrer et al. 2003). Longfin smelt have a 2–3-year life cycle, Adult striped bass are largely piscivorous. much of which is spent in San Francisco Bay Adults will spend their entire life in the SFE, and/or the Gulf of the Farallons, outside especially in San Francisco Bay, but when the Golden Gate (Moyle 2002; J. Rosenfield, ocean conditions are right, some will go out Aquatic Restoration Consulting, unpublished into the ocean as well (Moyle 2002). analysis). They spawn in the western Delta in Overall, striped bass populations seem winter and often spend the first year of their relatively immune to the negative effects of life in Suisun Bay and Marsh. Being anad- large-scale levee breaching. If breaching oc- romous and iteroparous with up to three curs in early summer, then large numbers of spawning age-classes, they are less vulnerable larvae and juveniles could die, but in follow- to extirpation by a large-scale event than del- ing years, they would presumably benefit from ta smelt. Like delta smelt and other plankti- increased pelagic habitat, especially if por- vores, however, longfin smelt have suffered tions of it were highly productive of zooplank- a large decline in their population in recent ton and small fish. Larger juveniles and adults years. Thus large-scale levee collapse in the are strong swimmers and could presumably Delta could initially harm longfin smelt, as quickly leave a submerged island after the ini- indicated above for delta smelt, although at tial event, assuming they survived the flood- least some longfin smelt would have lower ing event itself. They could also remain in the vulnerability because of their distance from vicinity of levee breaches to take advantage of the flooded islands. For more than half the the small fish moving through them, as they year (May–November), most longfin smelt do now on existing flooded islands such as would be beyond the likely reach of a flood- Frank’s Tract. ing event. Permanently flooded islands in the western Delta could ultimately become Sacramento Splittail important rearing habitat for larval and juvenile longfin smelt, depending on whether Splittail are largely confined today to the large zooplankton populations develop. In- SFE, where they rear in Suisun Marsh and oth- creased productivity of sloughs/channels in er places with fresh to brackish water sloughs san francisco estuary fish 11

(Moyle et al. 2004). Adults, which live up to 9 a giant tidal marsh, it was likely a major rear- years, migrate up river to spawn on floodplains ing area for fry before they moved out to sea or edgewater habitat in or just above the Delta as smolts. At the present time, rearing habitat (Moyle et al. 2004). Timing of spawning de- for fry in the SFE is minimal and fry survival pends on timing of natural flooding, usually is low; higher returns of adults from hatchery intermittently between late January and early fry generally occurs when the fry are planted May. Juvenile splittail rear on the floodplain in the SFE below the Delta (Brandes and Mc- and remnant channel habitats for a month Clain 2001; Williams 2006). Highest survival or so and then migrate rapidly downstream of fry and smolts in the SFE occurs in years to brackish rearing areas, where they feed on of high outflow in both the Sacramento and zooplankton and benthos. San Joaquin rivers (Newman and Rice 2002), During a major island flooding event, suggesting that it benefits the fish to move many splittail are likely to be drawn in. How- through the Delta rapidly. Survival of fry and ever, many others would remain in place be- smolts is also highest when the fish are largely cause of living in small sloughs distant from confined to the main river channels and do the Delta and also because they are strong not get moved into the central Delta (Brandes swimmers. If the levee breaching occurred in and McClain 2001). Although juvenile salmon conjunction with natural high flows in Janu- can be captured in SFE at almost any time of ary–May, large numbers of migrating adult or year, most movement is in December through juvenile splittail could be captured. Although April. sudden entrainment on the flooded islands As with other fish, the immediate effect could result in high mortality, the high toler- of a major levee failure in the Delta depends ance of splittail for poor water quality (low on the time of year in which it occurs, with dissolved oxygen, high turbidity, variable the greatest impact likely to be in February– salinities, etc.) suggest adults and large juve- April, assuming migrating juvenile salmon, niles are likely to survive the experience. The especially fry, sucked into the flooded islands flooded islands are not likely to be great habi- would mostly die. The effect would be greater tat for splittail until significant invertebrate for San Joaquin River salmon than for Sacra- fauna develops, especially amphipods and mento River fish because there would more mysid shrimp. However, permanently flooded likely be a continuous river channel for the islands that remain brackish enough to ex- fish to follow on the Sacramento side, due clude Brazilian waterweed should ultimately to location and greater flows. Once island la- become suitable habitat for splittail, especially goons have become established, they would shallower areas. generally be unfavorable habitat for juvenile salmonids because they would contain little Chinook Salmon of the shallow water edge habitat preferred by juvenile salmon. Instead, they would be Four runs of Chinook salmon pass through open-water or weed-choked and contain fair- the SFE on their way upstream to spawn in ly high densities of predators such as striped the Sacramento River: fall run, late-fall run, bass or largemouth bass. The effect would winter run, and spring run (Moyle 2002). All be determined in large part by how easily it runs are depleted from historic numbers, and would be for juvenile salmon to be carried the winter and spring run ESUs are listed as into the lagoons from the rivers and how endangered and threatened, respectively. The easy it would be to escape from them. Once fall run is supported in good part by hatchery the lagoons were established, high outflows production and occurs in tributaries to the down both rivers would minimize their ef- lower San Joaquin River, as well as the Sacra- fects, while low outflows would increase the mento River. Fry and smolts of the salmon are likelihood that juvenile salmon would wind found seasonally in the estuary, on their way up in them, especially on the San Joaquin downstream to the ocean. When the Delta was side of the Delta. It is possible that Suisun Marsh will be heavily used by juvenile salmon yellowfin goby Acanthogobius flavimanus, and once it floods because much of it will be pro- starry flounder Platichthys stellatus. Juveniles ductive shallow brackish-water habitat. of the euryhaline forms frequently appear in Effects on adult salmon would presum- the upper estuary, especially in Suisun Bay ably be small because of their focus on swim- and Marsh, usually during periods of low ming upstream through the Delta, although river flows (Matern et al. 2002). Thus, the ex- there would no doubt be some mortality and panded brackish water habitat in the upper straying if breaches occur during a period of estuary is likely to increase habitat space for significant migration. euryhaline marine species, especially during dry periods. Largemouth Bass Largemouth bass are introduced pisci- Overall Fish Responses vores that have greatly expanded their popu- It should be evident from the above de- lations in the Delta following the invasion of scriptions that responses of fish species to Brazilian waterweed (Brown and Michniuk large-scale island flooding will be highly vari- 2007). The waterweed provides habitat for able, a reflection of the mixed assemblages of the bass by creating cover for juvenile and native and alien fishes interacting in a highly adults, reducing flow rates through chan- altered, complex habitat. Thus, it is likely that nels, and causing sediment to settle from the unanticipated responses to the changed con- water, resulting in clearer water. It is the most ditions will also occur. For example, inland sil- visible species of a complex of alien pond verside Menidia beryllina are now abundant species that thrive in waterweed-dominated throughout the Delta and Suisun Marsh, so freshwater sloughs, including redear sunfish it is likely that they will colonize some of the Lepomis microlophus, bluegill L. macrochirus, newly flooded areas, depressing other fishes white catfish Ameiurus catus, brown bullhead through predation and competition (Bennett A. nebulosus, and common carp Cyprinus car- and Moyle 1996). In addition, new alien in- pio. By and large, these are the same species vaders could cause major shifts in abundance that are dominant in upstream reservoirs of established species. For example, two pisci- (Moyle 2002). Largemouth bass and associat- vores are poised to invade the SFE: northern ed species will expand their populations fur- pike Esox lucius and white bass Morone chrys- ther in flooded freshwater islands where Bra- ops (Moyle 2002; Lund et al. 2007). Despite zilian waterweed becomes established. While these uncertainties, the general patterns of these species can survive in brackish water fish response to sudden large-scale flooding habitat, most of them avoid it and would thus would roughly be the following: be present in only low numbers in brackish 1. Fishes within the suction zone of Delta le- lagoons without dense beds of waterweed. vee breaks (which could be a large, given the capacity of the islands to accept Marine Fish large volumes of water) would be sucked San Francisco Bay supports a diverse fauna into the island with some mortality from of marine fishes that includes most of the sediment in the water column, sudden common species found along the central changes in water quality (salinity, tem- California coast. The abundances of differ- perature, pesticide concentrations, etc.), ent species fluctuate both in response to and/or other factors associated with the ocean conditions and to freshwater flows sudden movement of large volumes of into the bay (Armor and Herrgesell 1985). water. The species affected would depend Not surprisingly, some of the most abundant on time of year of flooding and the loca- species are fairly euryhaline, such as Pacific tion of the flooded islands. herring, northern anchovy Engraulis mordax, 2. Once the waters had settled down, there Pacific staghorn sculpin Leptocottus armatus, would likely be a brief initial period of low plankton densities, followed by blooms this area would increase the amount of first of phytoplankton, then zooplankton, habitat for fishes. Presumably, the in perhaps within a period of 1–3 months. creased habitat would increase popula 3. In the years following the floods (1–5 tions of some of the desirable open-water years), new lagoons formed from island species, although much of it would be flooding would assume the character of dominated by waterweed and alien pond areas in the SFE with similar depths, fishes or by relatively low productivity flows, and salinities. Thus, those in the habitat depleted of nutrients by bivalve more eastern and central parts of the grazing. In this period, Suisun Marsh Delta would likely become dominated by would probably also have become at least Brazilian waterweed and alien freshwater partially flooded, with the potential for fishes. Lagoons in the western Delta that large increases in tidal brackish water maintained low (2–10 parts per thousand) habitat, potentially favorable to desirable salinities most of the time would have con- species such as longfin smelt, delta smelt, ditions similar to those in Suisun Bay. splittail, striped bass, and possibly juvenile Planktonic productivity is greatly reduced Chinook salmon. in Suisun Bay by the filter-feeding over- Thus the long-term overall effect of mas- bite clam, but it still serves as an impor sive flooding would likely be to increase the tant rearing area for pelagic fishes, at populations of at least some desirable species least in some areas (Hobbs et al. 2006). while greatly increasing the abundances of These areas would provide expanded hab- less desirable aliens, such as largemouth bass itat for euryhaline species such as striped and common carp. While there are fisheries bass, longfin smelt, and delta smelt, as for such species, they are deemed less desir- well as additional feeding areas for stur- able because the fish are not native and have geon Acipenser spp., splittail, and other large populations outside the SFE, unlike the benthic feeders that consume clams and species currently deemed more desirable. their associated faunas. A few deep lagoons that were created in interme- diate locations, where salinities and oth- Improving the Estuary for Fish er conditions were highly variable through the interactions of river inflow and tidal The above speculative discussion is based exchange, could become highly pro- on the scenario that California will continue ductive systems supporting dense popula- on its present track of managing the Delta en- tions of plankton and planktivores, vironment through a combination of apply- including delta smelt and striped bass. ing temporary levee repairs, poorly regulating Such areas could become a source for invasive species, removing large quantities of enhanced populations of euryhaline freshwater, managing Suisun Marsh as a low- fishes. salinity system, and monitoring desirable spe- 4. Over the longer term (5+ years), condi cies as they decline. In short, it is likely that tions in the lagoons would change further management of the SFE will continue as usual as levees continued to erode, parts of the until large-scale levee collapse forces large- lagoons fill in with sediment, and islands scale action, much of it likely to be poorly not flooded previously give way to new hy planned and futile in the long run (Lund et draulic forces created by the lagoons al. 2007). As indicated above, the massive col- (waves, changed current patterns, etc.), lapse of levees in the Delta and Suisun Marsh assuming that most levees were not re would not be a long-term disaster for many paired. Essentially, much of the central fishes, and fisheries and could even be a ben- and south Delta could become one efit. Large-scale levee collapse could be a di- large embayment, similar to Suisun Bay saster for the California economy, however, but fresher on its upper end. By size alone, mainly because it would disrupt the state’s water supply system and other infrastructure while favoring desirable (mainly native) pe- (Lund et al. 2007). Thus fixing the Delta and lagic species. The exact extent, frequency, and Suisun Marsh before an inevitable disaster is range of salinity fluctuation needs to be de- highly desirable, and several processes are un- termined by further studies of key organisms derway at the state level to determine options. (both desirable and undesirable species), but Lund et al. (2007) present nine scenarios for present distributional limits of the organisms a future Delta and Suisun Marsh, five of which suggest that fluctuations required are likely to they regard as feasible. Four of the five protect be in the range of 0–12 mg/L over 1–2 years, water supply while allowing some portion of the with high and low values sustained for 4–5 Delta to remain as habitat for native fish and months at a time. other desirable organisms. The five options The following are some general, large- provide suggestions for significantly improving scale actions that could improve habitat het- the habitat in the Delta and Suisun Marsh for erogeneity and create areas with desirable desirable species, provided that action is taken conditions of water quality, including fluctu- before large-scale levee collapse occurs. The ating salinity. This list is neither complete nor options of Lund et al. (2007) are only a tiny inclusive (Lund et al. 2007). fraction of the hundreds of permutations and combinations of actions that could be taken; 1. Suisun Marsh.—This region of the SFE is they are designed to represent examples of the headed inexorably to becoming brackish reasonable alternatives possible to provide a vi- tidal marsh, unless huge amounts of mon- sualization of management options. ey are spent on raising and reinforcing Here I will not discuss the alternatives, levees; such action may not even be pos- but instead discuss actions that will allow fish- sible as a permanent solution, given the friendly habitat to develop while not necessar- compressibility of the marsh soils under- ily reducing most of the services to humans lying the levees. Most of Suisun Marsh is that the SFE provides. These actions could be currently intensely managed in diked part of any scheme that seeks to modify the sections, principally as freshwater habitat Delta to improve or protect its water supply for waterfowl. Even under these condi- functions as well its ecological functions. The tions, the intervening sloughs, especially general approach towards creating an envi- in the few undiked areas, provide good ronment in the SFE that is more friendly to habitat for desirable fish (Matern et al. desirable fish species (and other biota) pre- 2002; R. E. Schroeter, University of Cali- sented here is to increase habitat heterogene- fornia, Davis and P. B. Moyle, unpub- ity. The basic concept is as follows: as much lished data). Improving the Marsh for fish area as possible should support conditions re- will require systematically breaching or sembling those of the historic SFE, especially removing many levees, reconstruction of in the Delta and Suisun Marsh, because these at least part of the original marsh drainage are the conditions to which the native fishes system, and removal of infrastructure. are adapted. However, the improved habitats Models for the creation of new tidal (and are likely to be in different locations than subtidal) marsh areas can be found in they were historically because of changed el- the currently undiked section of Rush evations due to subsidence and sea level rise. Ranch that is part of the San Francisco Thus habitats once present in the deeply sub- Bay National Estuarine Research Reserve sided central Delta will have to be located in (http://rtc.sfsu.edu/nerr/sf_bay_re- smaller patches in the less-subsided peripher- serve) and by an ongoing experimental ies (Figure 2). levee breach at Blacklock (http://www. A key to increasing habitat heterogene- iep.water.ca.gov/suisun/restoration). ity is to have as much area as possible where Even after restructuring the marsh, it may fluctuations in salinity discourage freshwater be desirable to continue to operate the and brackish water invaders (especially clams) large salinity control gates at the upstream end of the Montezuma Slough to man- down the bypass and re-enter the Sacra- age water quality. This slough and Suisun mento River via Cache Slough. The prin- Slough are the deep (2–6 m) main arter- cipal permanent water channel in the ies of the marsh and the principal habi- Yolo Bypass is the Toe Drain, which runs tats of pelagic fishes such as delta and along the levee on the eastern side. About longfin smelt, so it may be possible to op- half the bypass is in the Delta, and the Toe erate the gates to increase the ranges of Drain in this region is a leveed tidal salinity that favor these species and dis- slough, a branch of Cache Slough. The courage undesirable alien species. land in the bypass is a mixture of farm- 2. Cache Slough.—Cache Slough and adjoin- land and wildlife areas, but when it floods, ing areas make up essentially the north- it becomes high quality rearing habitat west corner of the Delta. The region is of for Chinook salmon fry and splittail, as high restoration potential as tidal fresh- well as other fishes. The flood waters may water marsh and slough habitat because also mobilize nutrients from the bypass, (1) island subsidence is low compared helping to support Delta food webs. From to other parts of the Delta, (2) it main- an ecological perspective, a problem with tains much of its original drainage pat- the Yolo Bypass is that is does not flood, tern, even though most of the channels even partially, every year. Construction of are leveed and some artificial cross chan- a gate on the Fremont Weir would per- nels exist, (3) it is a major spawning mit limited controlled flooding from the and rearing region for delta smelt, (4) it Toe Drain every year, improving growth has strong tidal currents that move water and survival of salmon and splittail and from the Sacramento River in and out of improving flows through Cache Slough. its channels, (5) it drains the lower end of 4. San Joaquin floodplain.—The channel of the Yolo Bypass (next section), and (6) the San Joaquin River above and through it contains the large, recently (1998) the Delta is highly channelized and pro- flooded Liberty Island that is being used vides little favorable habitat for desirable as an example of a passive restoration fishes: the water tends to be too deep and project (http://www.delta.dfg.ca.gov/ polluted in some places (e.g., Stockton jfmp/libertyisland.asp). The region can Ship Channel) and dominated by inva- be relatively easily converted into an ex- sive aquatic plants and invertebrates in panded version of the favorable tidal hab- others. One way to improve the habitat itat for desirable fishes (as well as for for fish is to create one or more bypass- waterfowl and other biota) through levee es like the Yolo Bypass. This would involve breaches, elimination of cross channels, removing or breaching levees from islands and other projects that improve circu- (e.g., Stewart Tract) that border the river lation. It is also a region where it should to promote periodic flooding. Such flood- be possible to create favorable habitats plain habitat may be especially beneficial for delta smelt, mainly spawning beaches as rearing habitat for juvenile salmon and productive rearing areas for larvae example of the benefits of restored flood- that also discourage their egg and larval plain in the Delta is provided by the small predators, especially inland silverside. restored floodplain along the lower Co- 3. Yolo Bypass.—To keep the city of Sacramen sumnes River, on the eastern fringe of the to from flooding, an artificial floodplain, Delta, which has proved to be beneficial the Yolo Bypass, was constructed in the to several native fishes and provides expe- 930s. Essentially, when the Sacramento rience in methods of modifying agri- River reaches a certain stage, it spills over cultural lands into fish-friendly flood- a low wall (Fremont Weir) into the plains (Moyle et al. 2007). 24,000-ha, 64-km-long bypass (Sommer at 5. Managed Delta islands.—The previous four al. 2001a, 2001b). The flood waters flow actions focused on areas at the edges of 16 moyle

the upper estuary because of the as- tion to eradicate a new invader before it sumption that subsided islands of the Del- spreads from the site of invasion. ta will fill with water and will become a large open-water system over which little Conclusions control can be exerted, aside from regulating freshwater inflow under some con- The San Francisco Estuary, especially the ditions (Lund et al. 2007). However, with Delta and Suisun Marsh, is predicted to un- some foresight, it may be possible to re- dergo drastic change in the next 50 years, with tain the levee integrity of some islands, by the probability of a major infrastructure disas- making them into islands of regulated ter increasing through time, just on the basis of aquatic habitat. Essentially, this concept earthquakes and land subsidence alone (Mount follows the lead of Delta Wetlands, a pri- and Twiss 2005). When sea level rise and in- vate group that has sought to use Delta creased frequency of flooding due to climate islands for water storage and wildlife habi- change are factored in, major change in the tat (http://www.deltawetlands.com/). coming decades seems inevitable. The disaster The levees of habitat islands would pre- scenario, however, is mainly for human goods sumably be reinforced on the inside by and services, especially water supply to urban being gradually sloped towards the inte- and agricultural areas. From a fish perspective, rior and planted with native vegetation the ecological changes resulting from flooding to stabilize the soils. Gates on the upper of numerous Delta islands and Suisun Marsh and lower ends of the islands could be are likely to create conditions that should be at used to regulate water quality on the is- least as favorable for desirable species at pres- lands, including salinity. This concept ent conditions, after a period of possibly high might be especially useful for islands (e.g., mortality created by the initial flooding events. Twitchell Island) in the western Delta lo- Potentially more favorable habitat will result cated close the Sacramento River, so that from a disaster scenario simply because there salinity could be manipulated by trapping will be increased area of open water and tidal either river or tidal water as needed (as marsh, some of it with enough fluctuation in is currently done with tidal gates in Suisun salinity and other factors to be suitable habitat Marsh). The islands of water could then for delta smelt, striped bass, and other pelagic be managed as nursery areas for desirable species now in decline. There is much uncer- fishes. Ideally, the gates would also al- tainty, however, about how much favorable low an island to be dried out or pumped habitat will be created under disaster scenarios out completely if necessary to control un- because alien fish and invertebrates dominate desirable alien species. so many current habitats. Thus, controlling 6. Prevention of new alien species invasions.— where possible the way habitat changes, as sug- A major stochastic factor that can nega- gested above and in Lund et al. (2007), could tively affect efforts to create a more desir- have major ecosystem benefits while being able ecosystem in the Delta is the invasion highly compatible with changing ways in which of alien species that become agents of eco- services provided by the Delta are delivered, system change, such as overbite clam or especially water supply. The principal basis for Brazilian waterweed. Some known harm- action is to increase habitat heterogeneity over ful invaders are likely to arrive in the near present conditions, as well as to increase the to- future (Lund et al. 2007). Thus, part of tal amount of aquatic habitat. No matter what any program of ecosystem creation must actions are taken, there will be a high degree include vigorous efforts to exclude new of uncertainty in the ecological benefits. Nev- invaders from all sources, including the ertheless, the present situation in the estuary shipping, horticultural, pet, and aquacul- represents an unprecedented opportunity to ture industries. There should also be in reverse the impacts of more than 150 years of place a mechanism that allows quick ac- negative ecological change. san francisco estuary fish 17 Acknowledgments DWR (Department of Water Resources). 1993. Sac- ramento–San Joaquin Delta atlas. California The ideas expressed in the essay come Department of Water Resources, Sacramento. from discussions with my coauthors on the DWR (Department of Water Resources). 2007. Del- UC Davis/Public Policy Institute of California ta risk management strategy report. Available: report: Jay Lund, Ellen Hanak, William Flee- www.drms.water.ca.gov (January 2008). nor, Richard Howitt, and Jeffrey Mount, as Feyrer, F., B. Herbold, S. A. Matern, and P. B. Moyle. 2003. Dietary shifts in a stressed fish as- well as with the many friends and colleagues semblage: consequences of a bivalve invasion who work in the San Francisco Estuary, but es- in the San Francisco Estuary. Environmental pecially William Bennett, John Burau, Chris- Biology of Fishes 67:277–288. topher Enright, and Robert Schroeter. Any Herbold, B., A. D. Jassby, and P. B. Moyle. 1992. Sta- mistakes and misconceptions are my own, tus and trends report on aquatic resources in however. the San Francisco Estuary. U.S. Environmental Protection Agency, San Francisco Estuary Proj- References ect, Oakland, California. Hobbs, J. A., W. A. Bennett, and J. E. Burton. 2006. Armor, C., and P. Herrgesell. 1985. Distribution Assessing nursery habitat quality for native and abundance of fishes in the San Francisco smelts (Osmeridae) in the low-salinity zone of Bay Estuary between 1980 and 1982. Hydro- the San Franscisco Estuary. Journal of Fish Bi- biologia 129:211–227. ology 69:907–922. Bay Institute. 1998. From the Sierra to the sea: the Hobbs, J. A., W. A. Bennett, J. Burton, and M. Gras. ecological history of the San Francisco Bay– 2007. Classification of larval and adult delta Delta Watershed. Bay Institute of San Fran- smelt to nursery areas by use of trace elemen- cisco, Novato, California. tal fingerprinting. Transactions of the Ameri- Bennett, W. A. 2005. Critical assessment of the delta can Fisheries Society 136:518–527. smelt population in the San Francisco Estuary, Hollibaugh, J. T. editor. 1996. San Francisco Bay: California. San Francisco Estuary and Water- the ecosystem. American Association for the shed Science [online serial] 3(2), Article 1. Advancement of Science, Pacific Division, San Bennett, W. A., and P. B. Moyle. 1996. Where have all Francisco. the fishes gone: interactive factors producing Kimmerer, W. J. 2002. Effects of freshwater flow fish declines in the Sacramento–San Joaquin on abundance of estuarine organisms: physi- Estuary. Pages 519–542 in J. T. Hollibaugh, edi- cal effects or trophic linkages. Marine Ecology tor. San Francisco Bay: the ecosystem. Ameri- Progress Series 243:39–55. can Association for the Advancement of Sci- Light, T., and M. P. Marchetti. 2007. Distinguish- ence, Pacific Division, San Francisco. ing between invasions and habitat changes Brandes, P. L., and J. S. McClain. 2001. Juvenile as drivers of diversity loss among California’s Chinook salmon abundance, distribution, freshwater fishes. Conservation Biology and survival in the Sacramento–San Joaquin 21:434–446. Estuary. Pages 39–137 in R. L. Brown, editor. Little Hoover Commission. 2005. Still imperiled, Contributions to the biology of Central Valley still important. The Little Hoover Commis- salmonids, volume 2. California Department sion’s review of the CALFED Bay-Delta Pro- of Fish and Game Fish Bulletin 179. gram. Little Hoover Commission, Sacramento, Brown, L. R., and D. Michniuk. 2007. Littoral fish California. assemblages of the alien-dominated Sacramen- Lund, J., E. Hanak., W. Fleenor, R. Howitt, J. Mount, to–San Joaquin Delta, California, 1980–83, and P. Moyle. 2007. Envisioning futures for the 2001–2003. Estuaries and Coasts 30:186–200. Sacramento–San Joaquin Delta. Public Policy Cohen, A. N., and J. T. Carlton. 1998. Accelerating Institute of California, San Francisco. invasion rate in a highly invaded estuary. Sci- Matern, S. A., P. B. Moyle, and L. C. Pierce. 2002. ence 279:555–557. Native and alien fishes in a California estua- Dettinger, M. D. 2005. From climate-change spa- rine marsh: twenty-one years of changing as- ghetti to climate-change distributions for 21st semblages. Transactions of the American Fish- century California. San Francisco Estuary and eries Society 131:797–816. Watershed Science [online serial] 3(4), Ar- Mount, J. A., and R. Twiss. 2005. Subsidence, sea ticle 4. level rise, and seismicity in the Sacramento– 18 moyle

San Joaquin Delta. San Francisco Estuary and thowski. 2005. Fish community ecology in an Watershed Science [online serial] 3(1), Ar- altered river delta: spatial patterns in species ticle 5. composition, life history strategies, and bio- Moyle, P. B. 2002. Inland fishes of California. Re- mass. Estuaries 28:776–785. vised and expanded. University of California Schroeter, R. E., and P. B. Moyle. 2006. Alien fishes Press, Berkeley. in California’s marine environments. Pages Moyle, P. B., R. D. Baxter, T. Sommer, T. C. Foin, 611–620 in M. H. Horn, L. G. Allen, and D. and S. A. Matern. 2004. Biology and population Pondella, editors. Ecology of California ma- dynamics of Sacramento splittail (Pogonichthys rine fishes. UC Press, Berkeley, California. macrolepidotus) in the San Francisco Estuary: a Sommer, T. R., W. C. Harrell, M. Nobriga, R. Brown, review. San Francisco Estuary and Watershed P. B. Moyle, W. J. Kimmerer, and L. Schemel. Science [online serial] 2(2), Article 1. 2001a. California’s Yolo Bypass: evidence that Moyle P. B., P. K. Crain, and K. Whitener. 2007. Pat- flood control can be compatible with fish, terns in the use of a restored California flood- wetlands, wildlife and agriculture. Fisheries plain by native and alien fishes. San Francisco 58(2):325–333. Estuary and Watershed Science [online serial] Sommer, T. R., M. L. Nobriga, W. C. Harrell, W. 5(3), Article 1. Batham, and W. J. Kimmerer. 2001b. Flood- Newman, K. B., and J. Rice. 2002. Modeling the plain rearing of juvenile Chinook salmon: evi- survival of Chinook salmon smolts outmigrat- dence of enhanced growth and survival. Cana- ing through the lower Sacramento River sys- dian Journal of Fisheries and Aquatic Sciences tem. Journal of American Statistical Associa- 58:325–333. tion 97:983–993. Williams, J. G. 2006. Central Valley salmon: a per- Nichols, F. H., J. E. Cloern, S. N. Luoma, and D. H. spective on Chinook and steelhead in the Cen- Peterson. 1986. The modification of an estu- tral Valley of California. San Francisco Estuary ary. Science 231:567–573. and Watershed Science [online serial] 4(3), Nobriga, M. L., F. Feyrer, R. D. Baxter, and M. Cho- Article 2.