s c i e n c e in a c t i o n

n e w s f r o m t h e C A L F E D b a y - d e l ta s c i e n c e p r o g r a m

Three years ago, “shallow waterhabitat” was the buzzword of theCALFED ecosystem restoration program,a triplet that conjured up visions ofrecreating the labyrinth of thrivingDelta marshes now barricaded behindlevees, and of saving the silverysalmon and smelt that would streaminto them to eat, sleep and spawn.

“CALFED started with the conceptthat creating various kinds of shallowwater habitats would make the ecosys-tem healthier and lead to a return ofnative species, but it’s much morecomplicated than that,” says SamLuoma, chief scientist for the federal-state CALFED Bay-Delta Program, whichis working to balance competingdemand for the Estuary’s fresh waterwith one of the most ambitious ecosys-tem restoration programs on the conti-nent. “We need to slow down ourexpectations, manage the Delta for avariety of species, as well as theendangered ones, and try to focusrestoration in places and ways thatimprove conditions and excludeinvaders.”

CALFED’s Ecosystem Restoration Plan(ERP) calls for the creation of up to9,800 acres of a mixture of shallowwater habitats (tidal perennial aquatic,shoals, sloughs and mid-channelislands), and another 30,000-45,000acres of fresh emergent wetlands inthe Delta alone by the year 2030 (italso calls for large acreages in SuisunMarsh and the North Bay). CALFED’s2000 Record of Decision is more gener-al, but sets ERP milestones in its bio-

logical opinions adding up to 13,600acres of shallow water type habitatsand wetlands within the first sevenyears of implementation. BeyondCALFED, U.S. Fish & Wildlife’s 1996 DeltaNative Fishes Recovery Plan – aimed athelping Delta smelt and splittail,among other natives — also recom-mends the restoration and protectionof shallow water habitat less than threemeters deep. And down in the Bay, the1999 Habitat Goals call for the recre-ation of over 50,000 acres of tidalmarsh in the baylands along theEstuary’s shores – to benefit not onlyendangered clapper rails, harvest miceand other wetland species, but alsothe ecosystem – within the next fewdecades.

“There’s lots of people on the band-wagon saying we need more habitat,more marsh, more shallows, but littlescientific justification yet for what theoutcome of creating it will be,” says theU.S. Geological Survey’s Jim Cloern.

At the most basic level, scientists allagree that the Estuary’s shallows, andtheir fringe of muddy flats and marshyshores, are indeed ecologically impor-tant – places where plants and fishgrow, eggs hatch, worms burrow, sed-iment settles and ducks dive and rest.But just which kind of shallows makethe best habitat for fish has beenharder to map out than say, whatmakes a good treetop for a spottedowl, or a big enough berry patch for agrizzly bear or the right height creekbank for a riparian brush rabbit.Learning about habitat hip deep inmuddy water or thick with tules ordown on a river bottom is a muchmore challenging endeavor involvingboats, rubber pants, diving gear, netsand meticulous timekeeping to keeptrack of tides, seasons and watermovements.

As CALFED plans and funds some ofthe biggest restoration projects todate, it has asked local scientists tocome up with a clearer picture of whatexactly goes on in the shallows andwhich flora and fauna, even whichbasic ecological processes, stand tobenefit from its plans to break openthe levees of numerous Delta islands

and let the rivers and tides reclaimtheir former domain.

“We forget that the processes thatcreated the marshes over geologic timeare not the same processes that willrestore them,” says researcher PhilipWilliams, pointing out that sea level riseover the last 6,000 years drowned theSacramento and San Joaquin river val-leys and created the over 500 squaremiles of marshes and shallows of the1800s Delta. Even the word “Delta” con-jures up a procedural misconception –a vision of the mighty Mississippi

Puzzling Over the Shallows

Continued page 2

DECEMBER 2001

DEFINITIONS

WHAT IS SHALLOW WATER HABITAT?Shallow water habitat comes in all

kinds of flavors, including shoals arounddeeper bays, “dead-end” sloughs andtidal, permanent and seasonal wetlands.To humans, perceptions of shallow anddeep are dependent on a person’s height,according to a recent paper by the U.S.Geological Survey’s Larry Brown. Mostofficial definitions run in the less thantwo-to-three meters deep category, ornot so deep that an NBA basketball playerin up to the eyeballs couldn’t still jumpup for air. Human height aside, manydefinitions better relate to the life historyof fish – Delta smelt need shallower areasto spawn than the open bays and riverswhere they spend most their adult lives,for example – or to the ability of light topenetrate so that tiny plants can grow (adepth a little over Sam Luoma's head).And don’t forget that in a tidal system,what is shallow one minute is deep thenext. A national U.S. Fish & WildlifeService wetlands classification system(1979) says that shallow water habitat inareas isolated from the tides is water twometers or less in depth, and in tidallyinfluenced areas is water two meters orless in depth at lowest low water. In arecent more Bay-Delta-specific paperdefining shallow water habitat, theInteragency Ecological Program’s estuar-ine ecology work team expanded the Fish& Wildlife definition to include areas upto four meters deep in the large openwaters of San Francisco, San Pablo, Suisunand Honker bays (see iep.water.ca.gov).

DWR scientist samples larval fish along marsh-edge in Mildred Island.

importing vast loads of sedimentsfrom the interior and spreading it out at its mouth into the sea.

Our own rivers, backed up behinddams and into narrow channels, nowhave little sediment to import intothe new shallows and marshes every-one is gung-ho to create – sedimentthese former wetlands may now needto bring them back up to intertidal levels after years of slowly subsidingbehind levees. “Building marshestoday is a much more sediment-reliant process than it was in the first place,” says Williams.

The sediment may be a long timecoming in some places, according tonew science (see The BreachingBusiness, p.3). As it dribs and drabsinto some of our restoration sites andfloods into others, as more invasiveplants than natives move in to colo-nize newly created shallows, as foodsupplies for fish decline and mercuryspreads into the estuarine food web(see Where’s the Food? p.8 andMercury in the Mix, p.6), many scientists are saying that perhaps thetime has come to rethink our goals.

“We’ve learned that for mostrestoration projects, it’s extremelydifficult to get back to a predistur-bance habitat,” says CharlesSimenstad of the University ofWashington. “In fact, the best we can hope for may be some alternative ecosystem.”

That catchy triplet “shallow waterhabitat” may be a part of this newecosystem, but many scientists aresaying the term is just too vague tobe useful anymore. It’s not so muchthe shallows we want, they say, butsomething more specific, like a shoalor a slough or a salt marsh, or some-thing more complex, like a mosaic ofthese habitats.

In the meantime, the puzzles in theshallows abound. Just breaching alevee and getting an instant greatmarsh, as happened at the famedCarl’s Marsh off San Pablo Bay (seep.7), may not be the norm up anddown the Estuary. So where are thebest places to breach? How much canwe rely on sediment and how much onvegetation to build the marsh? If newmarshes are filled with invasive plantsand fish, will they still have some habi-tat benefits to the natives? Will flood-ing all these old farmfields make morefood for fish or not, and in which sea-

sons? Or will it just release more mer-cury into the food web? Is it a betterbet to build a gravel bed in a creek,tear down a dam, supply more springflows or allow more seasonal floodingthan to try to bring back the Delta’sonce-vast shallows? Many of theseunresolved issues were no surprise toCALFED and figure prominently in itsEcosystem Restoration Plan. And as aresult of its investment in a whole newwave of interdisciplinary restorationscience, answers to some of thesequestions are slowly emerging.

“This messy circumstance in the Deltais not unique; there’s nowhere on theglobe where we’ve got this puzzlesolved,” sums up Cloern. “What we’redoing here is the leading edge,brand new environmental science. Itwill be we that provide models forthe rest of the globe about how toapply science to making environ-mental decisions.”

DECEMBER 2001 2

OUTLOOKA YOUNG SCIENCE:LEARNING AS WE GOSAM LUOMALEAD SCIENTIST CALFED

“Restoration is a very new science,perhaps no more than two decades old.It started with intuition, with the ideathat we could just define what we want-ed and then go out and restore it. Butthe thought that we could just turn theDelta into all this thriving shallow waterhabitat was too simplistic. The new sci-entific findings provide a reality check.

“One of CALFED’s priorities is to makesure restoration uses the best availablescience. We’ve learned several keythings in the past three years. First,we’ve done the math on sediment sup-ply and demand, and we have learnedthat supply will be a challenge inreturning vast reaches of the Delta tomarshes. This is the first time we’ve putreal numbers into the sediment equa-tion. People are now working to solidifythose numbers. Second, we’ve learned atremendous amount from looking at allthe ‘restoration’ of shallow water habi-tat that has already happened as aresult of natural or accidental leveebreaches (see p.3). Looking at how the

breach sites have evolved historicallywill help us more clearly see the rangeof possible outcomes of our ownrestoration efforts. Third, we’ve learnedthat it will be challenging to predictwhat will happen, at least at our presentstate of knowledge. This Delta system ismore complicated than anyone origi-nally thought. We must learn all we canfrom our successes, and even more sofrom the restoration projects we may bedisappointed with.

“Looking ahead, I see several keytasks before us. First, we need to contin-ue our new multi-disciplinary approachto restoration science so we can betterunderstand the links between hydrody-namics, ecology, chemistry, geomor-phology and biology. Taking this kindof approach in our study of places likeFranks Tract and Mildred Island hasreally paid off (see p.10). Second, weneed to learn more about the linksbetween exotic and native species, andhow exotics impact different kinds oforganisms. Third, we need to develop amuch more detailed understanding ofthe life history of the fish we are tryingsave, what their habitat needs really areand what truly threatens them. Part ofthis will mean investing in analysis ofdata we already have. Creative solutionsto ecosystem restoration will come frombetter knowledge in these areas.”

Mildred Island

Puzzling Over the Shallows

s c i e n c e in a c t i o n

RESEARCH

The Breaching BusinessScientists have been wading out

into shallows all over the Delta andBay seeking insight into what reallyhappens when we breach a levee torestore a former marsh: how muchsediment settles in or flushes through,what kinds of plants grow and howfast, what kinds of fish use the newhabitat and how much can we rely onnature to do the restoration work?CALFED has funded two “BREACH”studies – one comparing breachedlevee sites in the Delta, and a newerone now underway examining sitesfurther downstream in Suisun and SanPablo bays. Results to date providenew clues about what to expect fromshallow water habitat restoration.

The first set of BREACH studies, con-ducted between 1998 and 2000 in theDelta, analyzed six shallow water siteswith historical levee breaches in orderto predict the feasibility, patterns andrates of restoration to natural ecologi-cal function. The six sites ranged inage from 13 to 67 years since breach,and in subsidence levels from 1.1 to 2.3meters below sea level.

Some of the sites resembled largelakes (like Mildred Island, an expanseof open water ringed by levees); somefeatured tule marshes (such as Donlanand Venice Cut islands, whose eleva-tions were already intertidal at thetime of breaching or had been artifi-cially raised with dredged material);and some were in the more intermedi-ate stages of wetland development(Old Prospect Island). Researchers alsohunted down four remnants of Deltawetlands that had never been com-pletely enclosed by levees to serve asreference points for comparison tohistoric conditions.

“This was a whole new approach for the Bay and Delta region,” saysBREACH studies team leader CharlesSimenstad of the University ofWashington. “Nobody had ever reallylooked at naturally breached sites forinsight into where we were headed.”

To evaluate restoration trajectoriesat these sites between February 1998and December 1999, the BREACH Iresearch team undertook an integrat-ed approach – combining the analysis

of historical data and photographswith field work to see what was actu-ally happening down in the shallowsand oozes. Methods included applyingGIS to aerial photographs; followingup with ground truthing and shorelinetransects to assess marshplain devel-opment; conducting high-resolutionGPS surveys; sinking sediment andbenthic cores; and trapping insectsand fish.

First up was an attempt to createconceptual model of how thebreached levee study sites evolvedover time, an analysis by PhilipWilliams and Michelle Orr of PhilipWilliams & Associates and theUniversity of New Orleans’ DeniseReed. The model suggested that sedi-ment accumulation in deep, openwater areas initially appears to out-pace rates of sea level rise, then slows.The slow-down occurs because afterthe first gush of sediments and waterinto the site from the breach, theamount of sand and mud floating inthe water column diminishes andwhat’s left, and what’s on the bottom,keeps getting stirred up by waves.

So depending on local conditions,sediment build up at some sites maycontinue to outpace sea level rise(especially if vegetation gets estab-lished), while at others it may not.

Mildred provides a good case in point. To figure out how much sediment the island had accumulatedsince its 1983 breach, Orr sunk a coreto the point of resistance (whichresearchers assume is the level of the original pre-breach farmfield,where the soils are harder and morecompacted), then measured thedepth, which turned out to be O.64meters. Averaged out over the years,Mildred’s annual rate of sedimentbuild up came out to 47-51 mm,though Orr thinks a sizable part of thisresulted from the first post-breachgush. A look at historical data fromRhode Island confirmed a similar rate.

“Even though we’ve seen about twofeet of accreted sediments at Mildred,this data is telling us that the systemis heavily subsided, that a return totidal elevations may take a century or more, and might never happen,

Continued page 4

3

Antioch

Sacra

mento

Rive

r

ProspectIsland

Lindsay Slough

Sand Mound

Slough BaconIsland

HollandTract

MandevilleIsland

McDonaldIsland

EmpireTract

Venice Island

FranksTract

WebbTract

BouldinIsland

StatenIsland

TylerIsland

GrandIsland

SutterIsland

MedfordIsland

MildredIsland

BethelIsland

JerseyIsland

BradfordIsland

TwitchellIsland

ShermanIsland

BrannanIsland

RyerIsland

AndrusIsland

HotchkissTract

Brown’sIsland

San JoaquinRiver

MandevilleTip

VeniceCut

DonlanIsland

BREACH I STUDY SITES

SITES Years Meters Since

North Subsided BreachOld Prospect 1.5 35

Lindsay Slough - -

WestLower Sherman 2.3 73

Donlan Island 2.0 61 (13)*

Browns Island - -

Central & EastL. Mandeville Tip 1.1 65

Venice Cut 2.0 67 (14)*

Mildred \ 4.5 15

U. Mandeville Tip - -

Sand Mound Slough - -* dredged material added Source: Williams & Orr

LEGENDBreached Levee SiteNatural Reference Site

The Breaching Business

and that there’s no guarantee the site will ever become a marsh,” says Simenstad.

Mildred’s neighbor Franks Tract isin even worse shape, in terms of itslong-term potential to grow tulesand cattails. Though Franks wasbreached in 1938, nearly half a centu-ry before Mildred, and though it wasmuch less subsided than Mildred inthe first place, there is no evidencethat sediment has built up relative to sea level rise since the breach. Theproblem at Franks is partly related to the waves rolling across its longfetches, which keep resuspending,rather than settling, the sediment.Indeed researchers now believeFranks Tract may never fill in at all,and may have reached what Williamscalls an “open water equilibrium.”Big Break and Sherman Lake mayhave also reached this state.

Elevation plays a critical role inwhat kinds of plants grow on thebreach sites. The sites need to besomewhere near tidal levels beforemarsh plants can begin to take root.The BREACH I conceptual model sug-gests that pioneer marsh vegetationestablishes rapidly (within fouryears) at elevations around 1 meterrelative to mean lower low water(MLLW). Once established, vegetationspreads at lower elevations by lateralexpansion from initial patches. Thisoutward expansion proceeds at aslow rate (maximum of 1.5-3.0 metersper year) and requires sheltered conditions with few waves.

“The marsh builds laterally, as well as from the bottom up,” saysSimenstad, explaining how seedlingstake root in the shallows at the edgesof a subsided site, then trap moresediments with their roots. “It’s apositive feedback loop. Once it’sshallow enough for plants to colonize,they build more of their own habitat.”

Although the presence of vegeta-tion reduces the potential for scourand promotes additional accumulation,the rate of elevation change canremain very slow. Breached sitemarshplains that have been vegetat-ed for decades remain well belownatural marshplain elevations,according to the BREACH I research.Sherman Island, for example, has had emergent vegetation for over 60 years and remains 0.38 of a meterbelow Mean Higher High Water

(MHHW), or O.48 meters below thereference site at nearby Browns Island.

Going beyond the historical andconceptual perspective, Reed took acloser look at current sedimentationrates, measuring build up and ero-sion at the 10 study sites betweenMarch 1998 and June 1999 with vari-ous types of cores and markers laidout in controlled plots.

Reed found at least 10 mm, andsometimes more than 20 mm, of accre-tion during the 13-month period. Moreaccretion occurred between March andAugust than between August andDecember, apparently because most of the river runoff and accompanyingsediment inputs occur during the latespring and early summer.

The northern sites showed thehighest levels of accumulation (up to40 mm), confirming hypotheses thatthe amount of tidal energy and sedi-ment supply varies between theNorth, Central and Western regions,and affects build up. There was littledifference in sediment findings forthe latter two regions, but a big dif-ference between their levels and thenorthern site at Old Prospect Island.The key geomorphic factor here, saysReed, is the location of these sitesnear the base of the Yolo Bypass andthe accompanying direct connectionto the Sacramento River. The researchtook place during a time of majorflooding in the bypass, mimicking ahistoric process of big sedimentinputs with high storm flows.

Sediment accretion doesn’t alwaysequal elevation change. Restorationsites with high rates of accretion rarelyshowed similarly high rates of eleva-tion change, or rates as high as refer-ence sites, says Reed. “The youngdeveloping marsh soils compact morethan the more mature older sub-strates,” she explains. One key factor islocation on the edges, versus the inte-rior, of the sites. At plots at the edge ofthe Prospect Island’s marsh, almost 40mm of material accumulated over themarker horizon, while elevation changewas effectively zero. But in the interiorof the island, a similar amount ofaccretion resulted in a elevationchange of almost 30 mm over the 13months. The difference may be dueto different plants producing differ-ing root structures and properties, orto the supply of heavier sandiermaterial to the edge of marsh.

Another key factor is type and volume of the material laid down onthe restoration floor. Reed found verylittle variation across the Delta interms of the accumulation of organicmaterial, but a big difference in theaccumulation of mineral material,such as that pouring into ProspectIsland from the Yolo Bypass. “What we really want, if we want a nice high marsh, is high weight for firm soil and high volume for more soil,and that comes from a high supply of mineral material,” she says.

Based on all the research, theBREACH I team projected that restora-tion projects could expect sedimentaccumulation rates of 4 cm per yearfor subtidal habitats and 1 cm per yearfor intertidal habitats, depending onwave and current energy.

“It was a big thing when we real-ized that these restoration sites werenot going to fill in by themselves,”says Reed. “Clearly, restoring Deltadynamics alone will not get us thearray of shallow water habitat wethink we need. Some structural meas-ures will also need to be taken tobuild substrates back; otherwise,instead of shallow water, we’ll getdeep water.”

If the water is shallow enough,then the BREACH team fieldwork indicated that tules will grow upwithin several years, but if it remainssubtidal, it will get colonized by sub-merged or floating aquatic plants,

DECEMBER 2001 4

CONCEPTUAL MODEL OF BREACHED WETLANDAGRICULTURAL PHASE

~1 YEAR POST-FLOODING

2-3 YEARS POST-FLOODING

20 YEARS POST-FLOODING

LEGENDArtificial LeveeAgricultural SurfacePost-breach Alluvial Fill

Natural LeveeMarsh Soil

Tules

MHHW

MLLW

MHHW

MLLW

MHHW

MLLW

MHHW

MLLW

One of BREACH I's conceptual models, showing a restoration tra-jectory of rapid development of shallow water, emergent marsh.Source: Williams, Orr & ReedDrafting: Nina de Luca

like pennywort, or invaders, likewater hyacinth, Brazilian waterweed(Egeria densa) and parrot’s feather.Vegetation surveys of BREACH I sitesindicated that all the less subsided(or artificially elevated) areas of OldProspect, Donlon and Venice Cutislands showed initial rapid tulemarsh establishment after breachingor dredged material placement.

“The marshes are only therebecause the dredged material wasplaced there, not because they camenaturally,” says Reed. “It shows usthat we don’t need to raise sites allthe way back to the level of naturalmarshes to get tules, just kickstartthem up to intertidal elevations.”

BREACH I sites with subtidal eleva-tions were dominated by invasiveaquatic plants, according to the study,raising new questions about whetherfloating or submerged vegetationhelps or hinders the transition fromsubtidal to intertidal elevations.

Researchers then moved on beyondthe mud and plants to see what kindsof aquatic critters inhabit these breachsites. In some cases, they found morebenthic invertebrates (clams, worms,etc.) in the oozes of the mature refer-ence sites, but the breached sites hadsimilar species compositions, indicat-ing their potential to contribute to theemerging wetland food web. However,insects falling out of the emergentvegetation were often more dense inthe restoring sites than in the maturereference sites. There were also distinct differences in densities andtypes of insects and invertebrates living in the emerging tule marshesfrom those in exotic and native float-ing vegetation. A comparison betweenwater hyacinth and native pennywortby the University of Washington’sJason Toft found higher densities ofthe amphipods and macroinvertebratesfavored by hungry fish in the penny-wort, largely due to differences inhabitat architecture – the hyacinth

was taller and covered more surfacearea, reducing dissolved oxygen levelsin the water for fish.

All of the information on elevation,vegetation and food offers buildingblocks in answering the final, andperhaps central, question of theBREACH I studies – are these floodedtracts good for fish? The Departmentof Water Resources’ Lenny Grimaldoand his colleagues set out to deter-mine if fish species composition inthree of the sites — Mildred, LowerMandeville Tip and Venice Cut – wasin any way related to the age of thewetland, its status as a breach or reference site or to physical attributeswithin the sites (e.g., vegetationstructure, water temperature, salinity, water clarity, etc.).

To find out, he set up fish enclo-sures 30 by 49 square meters in sizeand used block nets and beach seinesto catch juvenile fish in intertidal andnearshore areas, and purse seining tosample deep subtidal areas inaccessi-ble by wading. Researchers measuredenvironmental variables before, dur-ing and after each sampling event,and surveyed vegetation. During the16 months of the study, they collecteda total of 47,138 fish representing 32species. The five most abundantspecies, comprising 90% of the totalcatch, were threadfin shad, inlandsilverside, redear sunfish, bluegilland largemouth bass. The mostabundant natives, only 2% of the

catch, were tule perch, splittail, chinook salmon and prickly sculpin.

According to Grimaldo, the resultsconfirm what other studies havebefore them – that introduced fisheasily colonize and dominate thesenew habitats. But he also discoveredthat native fish turned up in all habitat types – from those with open waters to those with dense submerged aquatic vegetation. This suggests that restoring just onehabitat type cannot be expected torecover all native fish, he says. It alsosuggests that there may be somehabitat trade-offs between breachedwetland restorations of different elevations.

“Introduced species are so estab-lished it’s hard to tease out what’sactually good for the natives,” saysGrimaldo. In the BREACH I studies,Grimaldo did find some clues in theseasons and water temperatures:native fish spawned and reared at thestudy sites during an early window inthe spring under a cool temperatureregime, ranging from 10-18°C. Incontrast, introduced fish spawnedand reared from late spring to earlyfall, when temperatures were warmer,ranging from 15-25°C. “We may needto take advantage of opportunities torestore wetlands at certain elevationsthat flood in the spring when thenatives spawn but not during thesummer when the non-natives

s c i e n c e in a c t i o n

5

Continued page 6

30• C

25

20

15

10

5

0

1.6

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0JAN FEB MAR APR MAY JUNE JULY

NATIVE FISH - QUANTITY/METER3

INTRODUCED FISH - QUANTITY/METER3

WATER TEMPERATURE - C•

RESIDENT LARVAL FISH DENSITIES COLLECTED IN CENTRAL DELTA DURING 1999

Source: Grimaldo

DWR crew explores fish-habitat linkages usingdepletion seining inside a block-net enclosure.

DECEMBER 2001 6

spawn,” says Grimaldo. Other researchhas already shown that seasonalflooding of large floodplain areas likethe Yolo Bypass favors native species,like splittail, better adapted to winterflood extremes.

Grimaldo says there’s a lot of uncer-tainty about increasing native fishpopulations through habitat creation –because the Estuary’s hydrology is so permanently altered, and naturalvariability in flow conditions – whichstrongly influence natives — cannot bereplaced. For example, though therewere more native fish at the referencesites, they were much more influencedby the water temperature and vegeta-tion than by the age of the wetland.“Restoring marshes to intertidal levelsis not the only endpoint; anotherimportant endpoint is increasing natural variability,” he says.

Getting a better grip on the impactsof exotic plants might help too. Thepresence and density of submergedaquatic vegetation does influence thelives of fish, according to Grimaldo.These underwater plants provide bothbenefits (food web support) and prob-lems (obstacles) for fish, depending onthe size of the fish and the density ofthe plants. Many of the breach sites areso subsided and so deep that theymake perfect habitat for invasive Egeriadensa, which can become so thick thatit creates a wall between deepwaterand intertidal habitats that growingfish cannot maneuver in. This may con-strain salmon, for example, into deeperchannels where there’s less to eat andhigher risk of being eaten – as exploredin a tandem predation study.

“The proliferation of Egeria densa insubtidal habitats in the Delta has prob-ably had one of the biggest effects onfish habitat and assemblage structurein the region,” says Grimaldo.

In sum, the BREACH I research sug-gests that more opportunistic fisheswill benefit from early restorationphases, while fish with more restrictedhabitat and food preferences may ben-efit from later-stage intertidal habitats.

We may have to face the fact, sums up Simenstad, that in these Delta breach projects, remain deep for so long, “we’re ending up with anearly stage habitat we don’t want or a habitat for introduced species. Sothere’s an implicit ecological trade-offto strategies that promote large areasand long periods of the subtidal phaseof habitat restoration.”

The Breaching Business CONTAMINANTS

MERCURY IN THE MIXThe fact that flooding big island

tracts and restoring tidal wetlandscould be boosting mercury inputs intothe estuarine food web has notescaped CALFED. In its effort to take anintegrated approach to restoration sci-ence, CALFED has supported a multi-agency, multi-year mercury researcheffort, the first years of which havealready pinpointed mercury hot spotsin the Delta (Cache Creek, Cosumnesand Yolo basins), and the more recentyears of which have been examiningthe effects of wetland restoration onmercury methylation. Mercury is atrace metal once widely used in Sierragold mining, and widely mined itselfin the Coast Ranges. As a result, mer-cury is now widespread in Bay andDelta sediments. Wetland processescan promote the methylation of thismercury, transforming it into a chemi-cal form more easily taken up byclams, fish, birds and otherestuarine life, with potentialsub-lethal effects for them andhealth risks for any humansconsuming contaminated fish.

As part of CALFED’s mercuryresearch, U.C. Davis’ DarellSlotton has generated a numberof new findings. Though fishand clams in restoration sitesfrequently indicate similar levelsof mercury as those in off-sitechannels and rivers, the reasonmay not be that methyl mercuryproduction is the same every-where, but that vigorous tidalaction is doing a good job ofspreading it around within Deltasubregions. Preliminary data onthe amount of methyl mercurycoming in and going out on thetides suggests that restorationtracts, particularly the moremature, highly vegetated ones,may be important localizedsources of the metal in its mostbioavailable form.

“As far as methyl mercuryexport goes, the worst kind ofhabitat we sampled in the Deltawas the kind you find atMandeville Tip, where the tulemarsh is just solid. The secondworst was the kind at LittleFranks Tract, where the Egeria isso thick it’s like a fish tankchock-full of weeds. Both

showed a big response in producingdissolved methyl mercury on the tideand then passing it out into the sys-tem,” says Slotton. By contrast, Libertyand Little Holland tracts, with their bigmudflats, produced a lot less methylmercury, and the wide-open depths ofMildred actually passed less methylmercury out into the channels thancame in on the tides, he says. One thingstill puzzling Slotton is why some of themost extreme methyl mercury exportsites among the dozen or so habitat types examined were also in theregion where mercury bioaccumulationin silversides, a small fish, were lowest.

Slotton has also sought to respond tothose who believe that the raw methylmercury is just being dragged in andout with the mud, and has nothing todo with the restoration. He experiment-ed with factoring this issue into hisdata (see map). “If you normalize forresuspended sediment in the water, themercury levels look even higher,” he says.

Sacramento

Antioch

Sacra

mento

River

LibertyIsland

FranksTract

MildredIsland

ShermanIsland

LowerCacheSlough

San JoaquinRiver

MandevilleTip

Headreach

LittleFranks

LittleHolland

Tract

ToeDrain

(northerntule marshes)

(southernsand flats)

NORTH DELTA SITES

CENTRAL DELTA SITES

AQUEOUS METHYL MERCURY IN PAIRED INFLOWING VS. OUTFLOWING TIDAL WATER IN DELTA FLOODED TRACTS

LEGENDInflowing TideOutflowing Tide

10.08.06.04.02.00.0

MeHg ConcentrationNormalized to TSS(aqueous ng MeHg/gram of TSS)

Source: Slotton

s c i e n c e in a c t i o n

7

RESEARCH

Bay Breach RollsIn 2000, the BREACH team moved

down the Estuary to study tidal marshrestoration and processes at 10 sites inSuisun and San Pablo bays. Very earlyresults suggest that North Bay breachsites develop much faster than Deltasites, and that more native species maybe using these new habitats.

As in the Delta, the BREACH II studyis comparing restored sites with rela-tively ancient marshes (several thou-sand years old) and “centennial”marshes (around a hundred years old,and formed with the flood of sedimentfrom the Gold Rush). The main differ-ences with the Delta study sites arestarting elevation — Delta sites aresubsided up to six meters below sealevel, whereas most Bay sites are onlya few meters – and rates of sedimentdeposition.

“Getting the Delta sites back up totule marsh elevations takes forever, butdown at Pond 2A and Carl’s Marsh, forexample, we’ve found there’s stillenough sediments moving around theBay to see that the elevation debt hasmeasurably decreased over time,” saysthe University of Washington’s CharlesSimenstad.

No data is in yet on sediment accu-mulation or vegetation trends, butthere are some early samples of fishand birds at the Bay breach sites thatsuggest they may stack up to theancient marshes in some ways. SanFrancisco State University graduate stu-dent Tammie Visintainer has sampledfish at the sites during three seasonsthis year. Based on a back-of-the-envelope look at the April results only,native species made up the majority ofthe composition at all sites, rangingfrom 82-100%. Splittail, a threatenednative species, were common in somesites. At Suisun Bay’s Ryer Island, a 17-year-old restored site, splittail comprised 22% of the total catch, atSonoma’s Carl’s Marsh 13%, and atNapa’s Pond 2A, a former salt pond, 9%.

“In general, there was a higher taxarichness at the breach sites compared tothe reference sites,” says Visintainer.Looking at some of the later samplings,she notes that species compositionchanges greatly with season. Forinstance, juvenile Pacific herring madeup 61% of the total catch at Toy

Property (breach site), 99% of the catchat Greenpoint (reference site), 64% ofthe catch at Petaluma Ancient (ref),and 20% of the catch at Carl's Marsh(breach), but after April, counts greatlydiminished at all sites as the herringheaded out into the Bay.

BREACH II researchers also surveyedbirds at the study sites, and noted afew interesting things. Based on a pre-liminary look at data from sunrise pointcount surveys conducted by JulianWood and Hildie Spautz from the PointReyes Bird Observatory, the earlyrestoration sites, which still haveexposed mud, are more attractive toshorebirds and waterfowl, especiallyduring the migration period, than moremature sites with more vegetationcover. The more mature breached sites,and the reference ancient and centen-nial marsh sites with tall dense vegeta-tion, support more passerines and othermarsh-breeding sparrows, wrens andrails.

“You can see the value of having amosaic of marsh habitat, with a vari-ety of ages and marsh cover types,including vegetation, ponds andchannel systems, from these find-ings,” says Spautz.

According to Simenstad, the emerg-ing differences between Bay and Deltabreach study findings suggest some-what different conclusions about therate and fate of shallow water habitatrestoration. For instance, the role of lowsuspended sediment sources and wind-generated sediment resuspension thatappears to be so prominent in the Deltamay not stall the development rates asmuch lower in the Estuary becausemany more of these marshes are smallsites adjacent to considerable riverineand Bay sediment sources.

While there still isn’t any data onthe sediment accumulation andrestoration trajectories of the BREACHII sites, some sites overlap with workalready done by Philip Williams andMichelle Orr, and other lessons learnedmay weigh in. In a review of the histo-ry of 15 re-flooded sites around theBay, ranging in size from 18 to 220hectares, and in age from two to 29years old, Williams and Orr found that50% cover established at nine of thesites within 20 years, some within justfour years, a faster rate of cover thanon the Delta sites.

For the Bay, Williams and Orr concluded that three factors retard the

timeframe for vegetation establish-ment: limited sediment supply, erosionof deposited estuarine mud via inter-nally generated wind waves andrestricted tidal exchange. They pointout that the shorter timeframe for vege-tation colonization and marshplainevolution experienced in earlier, smallerand less subsided sites may not neces-sarily be replicable by simple leveebreaching on larger subsided restora-tion sites now planned.

“We have a rather rosy view of howeasy it is to restore tidal marshes basedon small projects next to high sedimentsources with no wind waves, like Carl’sMarsh,” says Williams. Carl’s Marsh is a45-acre site with a starting elevation ofnear mean lower low water that wasbreached in 1994 by Fish & Game’s CarlWilcox dynamiting the levee.

“Initial elevation is a key determi-nant, but also the bigger sites takelonger, and a lot of sites proposedthese days are larger than anythingdone in the past,” says Orr, referring tothree projects of the thousand-acre-scale now in the works for the North Bay(Cullinan, Montezuma and Hamilton).“I’m not saying we shouldn’t do large-

Rate of vegetative colonization as a function of initial elevation. Shaded bar represents the approximate Spartinafoliosa colonization elevation. Error bars represent therange of uncertainty. * Subsequent lateral colonizationoccurs down to approximately mean tide level (MTL).

2.5

2.0

1.5

1.0

0.5

0.0

-0.5

-1.00 20 40 60 80 natural

site

Upper Mandeville TipProspectWest

Donlon/Venice

Donlon/Venice

25

20

15

10

5

0-4.8 -1.5 -1.0 0.0 0.5

Skaggs Island

Carl’s Marsh

Warm Springs

-0.5 1.0 1.5

Napa Pond 3

Alameda( Pond 3)

Faber

Outer Bair

Cogswell

InnerMuzzi

Slaughterhouse

White Slough

Greenpoint

Nevada

NapaPonds 4 & 5

CullinanRanch

Napa Pond 2A

Outer Muzzi

BAY-DELTA VEGETATION THRESHOLDS

Age when 50% vegetatedCurrent age if not yet vegetated (as of Year 2000)Planned Napa-Sonoma restorations

INITIAL ELEVATION (m MTL)

YEARS TO 50%

VEG

ETATED

AGE OF SITE (YEARS)

MEDIAN

ELEVATION

& RAN

GE (m

MLLW

)

Continued page 8

DELTA

BAY*

Source: Williams & Orr

RESEARCH

Where’s the Food?The shallows grow food, providing

places where the warmth, light andslow-moving conditions in the waterspur the growth of tiny driftingplants and animals (plankton) at thebase of the estuarine food web. Justwhat kind of shallows produce food,and under what conditions is a keyscience question for CALFED as it setsgoals and buys land for the creationof new shallows to sustain smelt,salmon and other native fish on thewane. A recent CALFED-funded inves-tigation into the food question —overseen by the U.S. GeologicalSurvey and conducted by a multidis-ciplinary team including scientistsfrom the Survey, three universitiesand the Department of WaterResources — suggests that the bestfood isn’t coming from where wethought it was, that the supply islimited and that the Delta’s ability to produce food, in general, is notonly low, but in decline.

“You have to figure out what’slimiting living resources before youcan conceive of programs to restorethem,” says team leader Jim Cloernof the Geological Survey.

Food in the estuarine ecosystemderives from organic matter – nutri-ents, broken-up bits of plants, algaeand phytoplankton (the plant formof plankton) in the water and soil.And one key task for the researchteam was to assess the sources andquality of organic matter (measuredas dissolved and particulate organiccarbon, or DOC and POC, in scienceacronymese) in the Delta. Organiccarbon is the source of the energyand calories that fuel biological pro-duction at the base of the food web.To trace the carbon in Delta watersback to its sources, Cloern’s teamused stable isotopes, biomarkers andmass-balance models.

“If you’re trying to find out wherethe food comes from, you need asmany different clues from as manydifferent sources as possible,” saysCloern. “There’s no one tool, markeror method that can tell us if a parti-cle of organic matter is 2000-year-old stuff from the soils around Red

Bluff, or phytoplankton produced inthe reservoir behind Shasta Dam ornutrients coming out of theSacramento sewage outfall.”

Cloern’s team found that one bio-geochemical tool (stable isotopes ofcarbon and nitrogen) commonlyused to identify sources of organiccarbon in other estuaries doesn’twork in the Delta. “We analyzed anunprecedented 868 plant samples forC-N isotope ratios and could find nounique signatures for individualplant groups,” he says. Instead, theparticulate organic matter in theDelta often has isotopic compositionsthat “don’t look like any living plantsources,” says Cloern, suggestingthat the bulk of it might be very oldand low-quality matter originallyproduced by terrestrial plants.

Most of the organic matter in theDelta, according to mass-balancemodels constructed by Alan Jassby ofU.C. Davis, is imported by its rivers,or what scientists call outside (or“exogenous”) sources. But most ofwhat’s delivered is what Cloern calls“trash organic carbon.” This low-quality dissolved organic matter hasto be converted into living particles(a process facilitated by bacteria)before it can be consumed by clams,copepods and other critters at thenext level of the food web, a conver-sion resulting in an inefficient respi-ratory loss of a large fraction of itsenergy. Even though the pool sizes ofPOC are smaller than those of DOC inthe Delta, the former is more biologi-cally available and energy efficientas a food pathway because the POC isthe pool containing phytoplanktoncells. As Cloern likes to say,“Phytoplankton rules.”

While some phytoplankton comesfrom rivers, most is home grown inthe Delta’s shallows and marshes.“The bottom line is that three yearsago, we thought outside sourceswere the biggest contributors toDelta food supply, but now we knowthat inside sources are most impor-tant,” says Cloern.

DECEMBER 2001 8

The Geological Survey’s AndyArnsberg on the lookout.

scale restoration, just that we shouldhave different expectations,” she says.

The review of the 15 sites also suggeststhat the formation of tidal channelswithin the marshes is greatly dependenton whether and how high the site is filledprior to breaching. Filled sites at marsh-plain elevations (above 0.3 below meanhigher high water) can vegetate quickly,but after several decades show littledevelopment of tidal channels.

Tidal channels are good habitat for fish,perhaps some of the best shallow waterhabitat there is. Looking at three artificiallyfilled sites (Faber, Pond 3 and Muzzi),Williams and Orr found that they had beenfilled too high to allow tidal channels todevelop quickly.

“If you fill all the way up to the natu-ral marshplain level, then the only timeyou get appreciable water on the plain isduring the spring tides – just a fewtimes per year,” says Williams. “That’snot enough water draining on and offthe marsh daily to scour out a tidalchannel. “ Williams recommends a fillthreshold of about 1 foot below MHHW topromote channel development. “It usedto be that success was defined by vege-tation, but we now realize channels areimportant for fish and detritus movingin and out.”

The message that both BREACH stud-ies, and the related research, seem to behammering home is that the key to suc-cess is the restoration of processes, notjust landscapes. “Natural evolution ofmarsh processes is going to take time,far longer than anyone thought,” saysWilliams. “Our highest priority shouldnow be to develop restoration strategiesand objectives that are compatible withlong-term estuarine processes.”

Fyke net at low tide in Petaluma Marsh

Bay Breach Rolls

So just how productive are theDelta’s shallows? The Survey’s BrianCole incubated phytoplankton inDelta water samples to come up witha general idea of how fast these tinyplants grew. He plugged his findingsinto a model, which projected growthrates, and gave the rates to Jassby.Jassby then examined data on chloro-phyll a and turbidity levels in theDelta between 1975 and 1995, whichwere collected by the InteragencyEcological Program (chlorophyll a andturbidity are two excellent indicatorsof productivity, the former as a meas-ure of phytoplankton abundance, thelatter as a measure of light penetra-tion for photosynthesis).

By combining the IEP data with thephytoplankton model, Jassby calculat-ed that the inherent rate of the produc-tion in the Delta is, and has been fordecades, low. “It really surprised usthat it was such a small number,” saysCloern (70 grams of carbon per squaremeter per year, a rate of primary pro-duction half that of poor producer LakeTahoe and smaller than the most nutri-ent-depleted regions of the ocean). Inaddition, Jassby and Cloern projectedthat between 1975 and 1995, the Delta’sproductivity rate declined by 43%.

Why is the Delta’s phytoplanktonproductivity so low? Partly because ofthe natural turbidity of Delta watersdue to high suspended sediment con-centrations and partly because thesystem has been so altered by dams,diversions and mining – producingdeeper rivers and fewer shallows andfloodplains, which in turn has limitedlight penetration and increasedcloudiness in the water. In addition,

an invasive clam called Corbicula flu-minea has been very likely eating thephytoplankton (if the eating habits ofits fellow invader, Potamocorbula,down in the North Bay is any indica-tion). Cloern says restoration plannershave to accept, as a result, that theDelta is an “inherently low productionsystem,” and work to enhance pro-duction – the opposite, he says, ofefforts in Chesapeake Bay, for exam-ple, where phytoplankton bloomsfrom excessive nutrients are a nui-sance.

Not having very much phytoplank-ton may be limiting the zooplanktonthat feeds on it, according to anotherrelated study. Anke Mueller-Solger, ofU.C. Davis, fed juvenile Daphniamagna (a water flea common amongzooplankton) detritus-rich waters col-lected seasonally from four differentDelta habitats to examine the relativenutritional value of phytoplanktoncompared to the much more abun-dant particulate detritus. Nutritionalmatter in the samples ranged from330 to 3800 parts per billion of POC,with about 15% of the carbon con-tributed by phytoplankton and theremainder by detritus particles.

Growth rates of the water fleasmeasured in the lab differed betweensamples from different habitats: thefleas grew faster in samples fromshallower habitats (largely a functionof how much phytoplankton wasavailable to the water fleas in thesesamples). Mueller-Solger’s resultssuggest that even in a system like theDelta, with high amounts of poten-tially nourishing detrital carbon, zoo-plankton may still require the nutri-

tion associated with fresh phyto-plankton. “These water fleas need toeat their veggies, and enough ofthem,” says Mueller-Solger.

Cloern’s team applied her lab con-clusions about how much chlorophyll azooplankton need to thrive to theirwork out in the field. In over 200samples collected in different seasonsand regions of the Delta, they foundthat 83% didn’t have enough chloro-phyll a to support zooplanktongrowth, such as that measured for thewater fleas. “So there is less algae inthe water than they need to grow atthe optimal rate,” says Cloern. “Andthere’s more and more evidence thatfood is limited at several differenttrophic levels. In terms of restoration,we may now need to accept that it’s aproperty of our system that organismsthat rely on phytoplankton as a foodresource can’t get enough of it togrow well. “

Or they may not be getting the rightkind. Mueller-Solger points out thatnot all algae are created equal, andgenerating the right amounts of theright kinds “so that everybody findstheir favorite food may be rather chal-lenging as a restoration goal.”

So how do the conditions in twoexisting Delta shallow water habitatsaffect these food web relationships?Another member of the U.S. GeologicalSurvey team, Lisa Lucas, directed aclose-up comparison of Mildred Islandand Franks Tract in the Central Delta tofind out. “They look the same from thesky, but they’re functioning com-pletely differently,” says Cloern.

Franks Tract is a 3,000-acre islandfirst flooded in 1938, and nearby

1,000-acre MildredIsland was flooded in1983. Though FranksTract was a lot lesssubsided than Mildredwhen first flooded,neither have devel-oped much marshcover and both con-tain large stretches ofopen water. But themain difference, fromLucas’ point of view, isthat Mildred appearsto be a net producer(or “source”) of phy-toplankton for thepelagic food web

s c i e n c e in a c t i o n

9

Continued page 10

0.8 to 0.90.6 to 0.70.4 to 0.50.30.20.10-0.1 to -0.5-0.5 to -1.0-1.1 to -1.5-1.6 to -2.0-2.1 to -3.0-3.1 to -4.0

1/day

High Slack:T = 25 hr

High Slack:T = 125 hr

FRANKS TRACT

MILDRED ISLANDFRANKS TRACT

MILDRED ISLANDGRAZING AND GROWTH COMPARISON

Here researchers combined light-temperature-dependent photosyn-thesis, respiration and Corbicula grazing to calculate “µ”, or therate of phytoplankton growth available to zooplankton. If µ is pos-itive, then that location is a local net source of phytoplankton(quality food) to the zooplankton; if µ is negative, then that loca-tion is a net sink.

Corbicula grazing rates in meters per daycalculated from measured clam size, densityand biomass in June 1999. Franks Tract was“carpeted” with clams (mean grazing rateof 4.4 m/d), while Mildred had far fewerclams (mean grazing rate of 0.4 m/d).

Source: Jan Thompson & Lisa Lucas

µ/day

(e.g., zooplankton), while Franksappears to be a net importer (or“sink”),based on June 1999 fieldsampling (see below).

In addition, hydrodynamic model-ling by Survey postdoc Nancy Monsensuggests that Franks is more “leaky”while Mildred is more “lakey. “Bothhave levee breaches and tides sloshingin and out, but Franks has a lot morebreaches and a lot more sloshing, saysLucas. In such a leaky system, particleshave multiple opportunities to come inone opening and perhaps go outanother, she says. With all the sloshingaround in Franks Tract, residencetimes for water and particles were low,whereas residence times in the south-ern part of Mildred Island were quitehigh, offering a good incubator spotfor phytoplankton.

One major reason for the source-versus-sink difference between thetwo tracts is the heavy grazing goingon by the invasive clam Corbicula.According to the Geological Survey’sJan Thompson, Franks Tract is chockfull of Corbicula, with benthic grazingrates averaging about four cubicmeters per square meter per day,whereas Mildred had relatively fewclams and low grazing rates averagingone tenth of that seen in Franks.“Right now we don’t have a clue whyCorbicula is paving the bottom of

Franks but absent from some regionsof Mildred,” says Cloern.

“Franks is both leaky and carpetedwith clams, a virtual ‘Roach Motel’, soit ends up being a huge local sink forphytoplankton, perhaps even aregional sink affecting the food sup-ply in the environment around it,”says Lucas. “With so many clams,you’d imagine Franks would havezero chlorophyll a, but we actuallyfound low nonzero levels, which sug-gests it’s importing biomass. “Thismeans that, in order to understandshallow water habitats like floodedislands, restoration scientists have tounderstand the peripheral connectedchannels as well, because tidal cur-rents drive rapid exchanges betweenshallow and deep habitats. “We can’tstudy shallow open habitats as isolat-ed habitats because they’re not iso-lated – they are strongly connectedto surrounding waters,” says Lucas.

To get at this connection, Monsenillustrated patterns for particle trans-port in and out of the two islands anddeeper channels around them, usinga tidal hydrodynamics model calledDelta-TRIM (see next page).

Hydrodynamic-biological modellingnow in progress will further test the pro-ducer/importer hypothesis. New studiesunderway in 2001 indicate that many ofthe Mildred/Franks findings from 1999 are

holding up in other seasonsand other years. In September 2001,Geological Survey scientists launched anew field experiment aimed at quantify-ing phytoplankton fluxes at the openingin Mildred’s northeast corner. “This newinformation will help us assess whetherMildred is pumping substantial amountsof phytoplankton out to adjacent envi-ronments versus ‘hoarding’ it locally,”says Lucas.

So what all these investigationssuggest, in sum, is that while shallowwater systems may look similar onthe surface, they can function com-pletely differently, and have highvariability in the biomass of phyto-plankton, and thus high variability intheir potential to support fish andpelagic food webs. “There’s no suchthing as a generic shallow waterhabitat that’s good for food andDelta smelt,” says Cloern.

In terms of restoration outcomes,some aspects are controllable – likethe nature and number of the breachsites, and some are not – like theCorbicula. Clearly, the population

ecology ofCorbicula, as a fac-tor affecting foodsupplies in restoredshallows, needs tobe better under-stood. Scientistsalso want to explorehow physical fea-tures like leveebreaks, islanddepths, tides andfreshwater flows, aswell as the nature ofexchanges betweenthe shallows and thedepths, influencefood production.

“We need tolearn more aboutwhat these systemsturn out to bebefore we go outand make more ofthem,” says Lucas.

10DECEMBER 2001

Where’s the Food?

High Slack: T = 0 hr

High Slack: T = 15 hr

Low Slack: T = 0 hr

Late Flood: T = 6 hr

Late Ebb: T = 14 hr

Late Ebb:T = 25 hr

High Slack:High Slack:T = 25 hrT = 25 hr

High Slack:High Slack:T = 125 hrT = 125 hr

High Slack:T = 0 hr

Low Slack:T = 8 hr

High Slack: T = 8 hr

High Slack: T = 101 hr

PARTICLE TRACKS AND TIDAL SLOSHING

Initially, particles are uniformly distributed insideMildred Island, but tidescause Mildred to lose particlesto adjacent channels. Afterseveral tidal cycles (T=125 hr),many more particles areretained in southern Mildredthan in northern Mildred,suggesting a longer “residencetime” in the south than in the north.

Computer model-generated particlemovements in Franks Tract. Tides causeparticles to slosh in from adjacent chan-nels and back out again. After severaltidal cycles (t=101 hr), numerous particleshave been imported to Franks Tract.Many levee breaks around Franks Tractallow for further interaction betweenthe shallow lake and nearby deep chan-nels, rendering Franks Tract very “leaky.”

Channel-lake sloshing alsooccurs at Mildred Island. Afterseveral tidal cycles, numerousparticles have been imported tothe island. Because there arefewer levee breaks at Mildredthan at Franks, this lake-chan-nel interaction is less extensive.

FRANKS TRACT MILDRED ISLAND MILDRED ISLAND

Source: Nancy Monsen

Zooplankton sampling

APPLICATION

Take-Home Messages

Our recent scientific scrutiny of theshallows suggests a number of newthings for resource managers, policy-makers, activists and the public to thinkabout in terms of managing and restor-ing the Bay-Delta ecosystem. Clearly,we’ve learned that many of the restora-tion plans we originally developed basedon biological priorities – saving this fishor this mouse from extinction, bringingback the tules or the cordgrass – nowneed to be rethought based on geomor-phic possibilities. The existing landscapeand the way water now moves across itlimit our ability to build new habitats.We’ve also clarified some of our morecomplex restoration goals – not justwide shallows but also edges and mar-gins, not just more vegetation but wholeecosystem processes, not just endan-gered species but many native species,maybe even a few non-natives, not justone type but a mosaic of habitats…

“All this new information confirmsthat we have to be more mindful ofprocesses,” says the CaliforniaResources Agency’s Tim Ramirez. Or asthe Geological Survey’s Lisa Lucas putsit, we need a detailed study of process-es (as opposed to mere quantificationof outcomes, such as the amount ofvegetation or a head count of endan-gered fish) to see the underlying mech-anisms that create within-habitat and between-habitat differences in ecosystem function.

“The rationale for creating freshwatertidal marshes has not gone away, it’snow a question of where it’s easiest andbest to do it,” says Philip Williams.

The meager sediment supply won’tbe easy to overcome for restorationmanagers. Clearly, really deep islandsare not the ones to breach if you wantto make marshes or native fish habitatany time soon. In a sediment-starvedsystem, managers may now need todecide which new sediment sinks –gravel beds on creeks, floodplainsalong rivers, Delta islands or old saltponds – are the most important ones tofill first. Or which ones will come out theway we want them to based on the bestavailable science.

“Every last fisheries project on ourcreeks and rivers upstream can affect

the sediment budget downstream,” saysStuart Siegel of Wetlands and WaterResources. “More importantly, by creat-ing new sediment sinks via restorationthroughout the Estuary, along with newfill projects like SFO’s runways, we willfundamentally alter how sedimentmoves through the system and theplaces where it erodes and deposits.”

Williams thinks we need to startfocusing on less subsided areas aroundthe periphery of the Estuary likeMcCormick Tract, places where it may bemore “practical” to achieve marshplainlevels. Since sediment demand couldbecome so high relative to its supply,resource managers should considerphasing restoration over time, particu-larly if large new areas like the South Baysalt ponds (a 100-million-cubic-yardsediment sink) are restored. “We have tocreate sediment sinks slowly, maybe afew a year, rather than opening up allthe ponds at once,” Siegel says. As sedi-ment gets scarcer, accretion of organicmatter through vegetation decomposi-tion may also become an increasinglyimportant component.

Even if some flooded Delta islandsdon’t make good marshes, they maymake good food, another managementconsideration in an ecosystem that’snot only short on sediment, but also onphytoplankton. Mildred Island offers aninteresting case in point – too subsidedto get much sediment accumulation butproducing a good crop of phytoplank-ton in its quieter corners. IndeedMildred’s depth and lack of vegetationare pluses, in terms of producing littleof that lethal byproduct of restoredwetlands – methyl mercury. “It doesn’tmake any sense to restore the food supply if it’s just moving more toxicelements into the food web,” says the Geological Survey’s Jim Cloern.

The mercury side of the restorationequation seems to stymie scientists andrestoration managers alike. The YoloBypass and Cosumnes floodplainrestorations, for example, are great forfood, native fish, and sediment inputsto nearby restoration sites, but couldturn out to be bad for mercury, becausethey have some of the biggest inputs ofmining mercury from upstream. “Wereally need to get a handle on mercurymethylation levels in these two flood-plains, and factor our results into anyhabitat expansion there,” says U.C.Davis’ Darell Slotton.

HYDRODYNAMICSA RIVER RUNS THROUGH ITThe “shallows” conjure up visions of

wide waters placid enough for food togrow and fish to swim whichever waythey please. But two Department ofWater Resources engineers (Enright andGuivetchi) knew that something differ-ent was going on in Sherman Lake whennumerically modeled salinities through-out the western Delta were extraordinar-ily sensitive to how they handled thelake in their models. At their urging, theU.S. Geological Survey’s Rick Oltmannand Jon Burau put drifters and currentand salinity meters in Sherman Lake inthe fall of 1998. Instead of the expectedquiescent conditions, with long waterresidence times and associated phyto-plankton production, he found “a majorconveyance pathway” between theSacramento and San Joaquin riversthrough the lake.

Burau found that the tide arrives aboutan hour earlier on the Sacramento side ofSherman Lake, mainly because the tidewave moves faster on the SacramentoRiver because it is deeper, and has totravel farther on the San Joaquin Riverbefore it reaches the lake. These geomet-ric factors produce “a big water-leveldifference across Sherman Lake, which inturn creates strong, fast-moving tidalcurrents and large net exchanges acrossthe island,” he says. The stronger currentsare limited to the east side of ShermanLake, while the west side exhibits theexpected quieter conditions.

Burau found similar rapid tidalexchanges going on in the northern partof Mildred Island, but not in the south-ern part. “We can’t assume that theshallows are homogeneous,” he says.“We’re finding a wide diversity ofhydrodynamic conditions and habitatswithin large islands.”

Other research by Burau and colleaguesdownstream shows similar rapid tidalexchanges between Honker, Grizzly andSuisun bays and their surrounding shal-lows. Exchanges in certain areas are like“one big mixmaster,” says Burau. “If youwere a little fish, unless you could swimlike crazy, you wouldn’t hang out there.”

In Grizzly and Honker bays, theexchange occurs across a very wide area,whereas in the Delta, exchanges are high-ly constrained, moving through narrowbreaches. “When you’re deciding whatflavor of shallow water habitat you want,the width, depth and location of thebreach is very important,” says Burau.(To view drifter paths, see Burau p.12).

s c i e n c e in a c t i o n

11

Continued page 12

CREDITS

Writer/Research, Ariel Rubissow

-Okam

oto; Designer, Darren Campeau, w

ww

.dcampeau.com

Another theme emerging for manage-ment is to get over the old “control freak”syndrome and find ways to build in vari-ability. The scientists say that “extremeevents” are good for native fish andplants and bad for invaders, and thatthe permanent nature of the subtidalareas created by flooding Delta islandsprobably facilitates invasions. “In a lot ofcases, it’s the disturbances – storms,floods, droughts – that formed the sys-tem, so trying to control them is theworst we can do,” says the University ofWashington’s Charles Simenstad. “It’sactually counter to ecosystem complexity.”

Denise Reed of the University of NewOrleans agrees, saying, for example,that 1998 flood conditions reallyhelped the ecosystem. “Restorationplanners need to be prepared to capi-talize on flood events, to allow rivers tooverflow banks into flood plains anddeliver sediment pulses where they’reneeded,” she says.

Salinity is also an important factor increating the variability that puts adamper on invasions, says WaterResources’ Curt Schmutte. If we let thetides into vast new areas of the Delta,salinity may change across the Delta, hesays, and could, in turn, affect whatkind of species live in the shallows.

It does seem to be clear that non-native fish are thriving in the Delta, andthe flooding of large islands may begiving them a fin up. The Department ofWater Resources’ Lenny Grimaldo thinksit may be time to ask ourselves if thefocus on recovery of endangered fishthrough restoration is even feasible. “Wekeep promising we’re going to recovernative fish, but it may be time to revisitour original idea of desired restorationoutcomes and acknowledge otherrestoration benefits such as more stripedbass for the sport fishery or the simpleaesthetic pleasure of waters and wet-lands,” he says.

Despite the obstacles and complexi-ties, Cloern and other scientists thinkwe should persevere with shallow waterhabitat restoration. “It’s still worth it ifour expectation is to prevent endemicspecies from going extinct in our life-time. I think we can create pockets ofDelta smelt habitat, maybe just 10,000acres. But we’re not going to turn theDelta back into a 300,000-acre tulemarsh,” he says.

“Personally, I’d rather see an area wetthan urbanized, regardless of the non-natives,” throws in Grimaldo.

Questions remain. What shouldCALFED spend its money on first,and should it continue investingheavily in shallow water habitatprojects, or is the better bet sea-sonal floodplains like the YoloBypass or San Pablo baylands thatare not so heavily subsided as theDelta islands? According toSimenstad, the science suggeststhat restoration may be faster andmore feasible the lower (e.g., out-side the Delta) the place is in thesystem. In some places many smallpatches of certain kinds of habitatmay have to suffice. In others,larger tracts may be possible.

Given the within-Delta prob-lems, CALFED should think moreabout preserving the meager,relict wetlands that are left, saysSimenstad. “The Port of Stocktonhas proposed Browns Island fordredging and dredge materialdisposal. Is that appropriate whenit may take decades, if not a cen-tury, to restore anything compa-rable in the Delta?” he asks.

CALFED may also want to thinkabout whether its early ecosystemrestoration vision of making the CentralDelta a better place for fish by creatinga wide, long habitat corridor betweenthe Mokelumne River and Suisun Baystill has merit, says the CALFED ScienceProgram’s Kim Taylor. This visiondirected early ideas of where restora-tion should take place, but some scien-tists are still scratching their heads overits basis. “As far as I can tell this is aconcept borrowed from terrestrial con-servation ecology, where there hasbeen debate about the value of corri-dors connecting patches of habitat, buthow it would work for fishes of open-water habitats, I don't know,” says S.F.State's Wim Kimmerer. “Although Deltahabitat corridors could be a great idea,it has not been supported by any sci-ence I know of. We need to see theunderlying conceptual model with sup-porting data before we do any restora-tion based on the corridor idea.” Therecent years of shallow water researchcertainly provide some relevant dataand may even suggest the need toreconsider the corridor-based under-pinnings of CALFED planning.

Another concept that needs to berevisited is the connectivity betweenthe watersheds, the Delta and the Bay.“One major assumption I personallythink we need to throw out is that

shallow water habitats are totally disconnected from the ‘body’ of San Francisco Bay,” says Simenstad,pointing out that these habitats areconnected not only by tidal waters butalso by major estuarine interactions,and fish and avifauna populationdynamics.

In any case, this young science ofrestoration has discovered a lot in thepast few years, fueled by CALFED’sinvestment in it, which some scientistssay has had the added benefit of mov-ing many academics and graduate stu-dents into Bay-Delta specific research.

The feedback on restoration produced by this research is certainlywelcomed by CALFED management.Before the 2000 Record of Decision, theCALFED mission was to get stuff done,says Ramirez. “People wanted to seethings happen, and we went aheadand did a lot of restoration projectswith Prop. 204 and Category III money,but not as much science and research.Now, with Sam Luoma on board, thescience is coming in. All these thingswe’ve learned should become part ofour criteria for deciding which projects to fund in the future, and will help us become more strategic in our investments.”

Take-Home Messages

12DECEMBER 2001

CONTACTS, AUTHORS & WEBLINKS

BREACH I & II http://depts.washington.edu/calfed/calfed.htm

Tidal Hydrology & GeomorphologyMichelle Orr, Philip Williams & Associates mko@pwa-ltd.comPhilip Williams, Philip Williams & Associates pbw@pwa-ltd.com

SedimentologyDenise Reed, University of New Orleans djreed@uno.edu

Marsh & Shallow Water EcologyCharles Simenstad, University of Washington simenstd@u.washington.eduJason Toft, University of Washington tofty@u.washington.edu

Fish UtilizationSteve Bollens, S.F. State University sbollens@sfsu.eduLenny Grimaldo, Department of Water Resources lgrimald@water.ca.govZach Hymanson, Department of Water Resources zachary@water.ca.govTammie Visintainer, S.F. State University tamarine@sfsu.edu

Bird UseNadav Nur, Point Reyes Bird Observatory nadavnur@prbo.orgHildie Spautz, Point Reyes Bird Observatory hspautz@prbo.orgJulian Wood, Point Reyes Bird Observatory jwood @prbo.org

FOOD http://www.sfbay.wr.usgs.gov/access/wqdata/Brian Cole, U.S. Geological Survey becole@usgs.govJim Cloern, U.S. Geological Survey jecloern@usgs.govAlan Jassby, U.C. Davis adjassby@ucdavis.eduLisa Lucas, U.S. Geological Survey llucas@usgs.govNancy Monsen, U.S. Geological Survey nemonsen@usgs.govAnke Mueller-Solger, U.C. Davis & DWR amueller@water.ca.govJan Thompson, U.S. Geological Survey jthompso@usgs.gov

MERCURY http://loer.tamug.tamu.edu/calfed/Reports.htmDarell Slotton, U.C. Davis dgslotton@ucdavis.edu

HYDRODYNAMICS http://sfaby.wr.usgs.gov/access/flow/drifterstudiesJon Burau, U.S.Geological Survey jrburau@usgs.gov

This special publication was paid for by the CALFED Science Program. Special thanks toEcological Applications (re: Lucas, Cloern, Thompson & Monsen), Restoration Ecology(re:Grimaldo, Orr and Williams) and Limnology & Oceanography (re: Mueller-Solger) for letting us preview some of the data to appear in forthcoming articles. For more info onsources and citations, please contact the authors above or see the IEP Newsletter.