American Fisheries Society Symposium 39:81–96, 2004© Copyright by the American Fisheries Society 2004

81

Spatial and Temporal Distribution of Native and AlienIchthyoplankton in Three Habitat Types of the

Sacramento–San Joaquin Delta

LENNY F. GRIMALDO*, ROBERT E. MILLER1,CHRISTOPHER M. PEREGRIN2, AND ZACHARY P. HYMANSON1

Aquatic Ecology Section, California Department of Water Resources3251 S Street, Sacramento, California 95816, USA

Abstract.—We examined the spatial and temporal variability of native and alienichthyoplankton in three habitat types (marsh edge, shallow open-water, and riverchannel) in one reference and three restored marshes in the Sacramento–San JoaquinDelta, California, during 1998 and 1999. More than 6,700 fish embryos and 25,000 larvaerepresented by 10 families were collected in 240 tows during the 2-year study. Overall,the assemblage was dominated by alien fishes, but natives were more abundant duringwinter and spring, whereas aliens were more abundant during summer. Overall abun-dance was highest in marsh edge habitats, suggesting that this habitat provides favor-able larval rearing habitats for many fishes. The reference marsh was dominated byalien species making it difficult to assess whether it had attributes that promoted use bynative fish. Ichthyoplankton abundance varied comparably at restored sites of similarconfiguration. The restored site, with minimal tidal exchange and greater lower trophicproductivity, supported the highest densities of alien fish. We conclude that restorationprojects in this region of the estuary must consider the potential impacts of alien fisheson natives and evaluate strategies designed to improve recruitment success of nativefishes. Specifically, we suggest that restored wetlands that offer only winter and springinundation periods may provide maximum benefits to natives while limiting access bymany alien fishes regardless of specific habitat-use requirements.

Introduction

Many researchers have described how ichthy-oplankton populations can vary in composi-tion and abundance along geographic axes inestuaries, mostly with respect to salinity(Houde and Alpern-Lovdal 1985; Dodson etal. 1989; Meng et al. 1994; Rakocinski et al.1996) or distance from embayment openings(Laprise and Dodson 1989a). Though these

studies have yielded valuable information onthe range of various taxa within estuaries,microscale distributions of fish larvae havebeen overlooked (Houde and Rutherford 1993).Even within localized regions of estuaries,there are frequent strong physical and biologi-cal gradients (Jassby et al. 1995; Lucas et al.2002). These gradients are often responsiblefor the inherent patchiness of larval fish popu-lations (Cushing 1983; Houde and Alpern-Lovdal 1985). Because larval survival canhinge upon the microhabitat conditions withina system (Houde 1987; Houde and Rutherford1993), identifying spatial distributions of lar-vae among habitats could yield valuable in-sights into processes influencing recruitmentvariability.

Many estuarine organisms exhibit dis-

* Corresponding author: lgrimald@water.ca.gov1 Present address: California Bay-Delta Authority,650 Capitol Mall, Sacramento, California, 95814,USA2 Present address: California Parks and Recreation,8885 Rio San Diego Drive, Suite 270, San Diego,California 92108, USA

82 GRIMALDO ET AL.

tinct vertical distributions to aid migrationor position maintenance in specific regionsof estuaries (Laprise and Dodson 1989b;Kimmerer et al. 1998; Bennett et al. 2002). Todate, there has been little research to deter-mine if ichthyoplankton abundance variesacross gradients from nearshore to offshorehabitats (Paller 1987; Dewey and Jennings1992; Cardinale et al. 1998; Gadomski andBarfoot 1998). The purpose of our study wasto examine the temporal and spatial distri-bution of ichthyoplankton along threesubtidal habitat zones parallel to shore inone natural (reference) and three restoredmarshes. The three habitat zones were marshedge, shallow open-water, and main riverchannel. We simultaneously measured physi-cal variables and zooplankton densities todetermine how characteristics of each habi-tat type may have influenced ichthyoplank-ton abundance and distribution.

Study Area

The Sacramento–San Joaquin Delta is formedby the Sacramento and San Joaquin rivers,which drain into San Francisco Bay (Figure1). The delta is characterized by a myriad ofchannel and slough networks primarily un-der tidal freshwater influence. The delta hasbeen highly modified by human activities, in-cluding extensive dredging, contaminant pol-lution, and water diversions (Conomos et al.1985). At the turn of the 20th century, manyhistoric delta marshes were drained and iso-lated with rock-reinforced (riprap) levees.These and the remaining natural marshes arepatchily distributed throughout the delta. Be-sides extensive marshland loss, numerousalien plants and animals have successfullyinvaded and colonized the delta (Meng et al.1994; Bennett and Moyle 1996; Meng andMatern 2001). The San Francisco Estuary isconsidered the most highly invaded in theUnited States (Cohen and Carlton 1998).

Our four study sites were located in thecentral delta (Figure 1). We selected upperMandeville Tip (UM) as a natural marsh be-cause it was never leveed or directly alteredby human activities; therefore, we assumedthat habitat features (e.g., vegetation composi-

tion, hydrology) would represent relativelynatural conditions. We selected three restoredmarshes that were previously leveed for vary-ing time periods (15–62 years) but werebreached by natural (i.e., levee failure) or hu-man actions during the last 7 decades. Thesesites are Mildred Island (MI; breached 16years), lower Mandeville Tip (LM; breached65 years), and Venice Cut (VC; breached 67years). Breached marshes in the delta are typi-cally termed flooded islands because they arebelow sea level and often still have remnants,if not all, of the constructed levees aroundthem. Historically, these flooded islandswould have been dominated by marsh plainsdrained by small intertidal sloughs and chan-nels (Atwater et al. 1979). The subsidence isprimarily due to soil decomposition causedby agricultural activity and wind erosionduring the leveed period; subsidence is pro-portional to the number of years the islandwas leveed prior to being breached.

The subtidal interiors of the selected re-stored sites were characterized by differentphysical attributes, including shallow areasdominated with submerged aquatic vegetation(SAV), unvegetated shoals, and large openembayments. Subtidal regions at the naturalsite (UM) were confined to a narrow regionbetween the intertidal and main river channelinterface, and most of these habitats were colo-nized by SAV. The main difference among therestored sites was that MI still had much ofthe constructed levee surrounding its deepinterior (~4 m), whereas the interior portionsof VC and LM were generally shallower (~2m) and more exposed to the main river chan-nel. Tidal exchange at MI was largely restrictedto a large breach opening (~8 m wide, 20 mdeep) located at the northern end of the site(Lucas et al. 2002). The SAV beds at MI werelargely confined to an approximately 5-m-wide band along the inner rim of the site. Incontrast, the SAV beds at LM, VC, and UMwere far more extensive, at times extending asfar as 25 m from the shoreline during summer.The dominant SAV at all study sites was Bra-zilian waterweed Egeria densa and Eurasianmilfoil Myriophyllum spicatum, both alien mac-rophytes.

SPATIAL AND TEMPORAL DISTRIBUTION OF NATIVE AND ALIEN ICHTHYOPLANKTON IN THREE HABITAT TYPES 83

Methods

Field methods

Sampling planktonic organisms in shallowwater can be challenging when using tradi-tional stern-mounted towed nets because ofsubmerged structures that can impair, clog, orfoul the gear during tows (Cole and MacMillan1984). To overcome this problem, we bridled a4-m-long × 0.65-m-mouth-diameter ichthyo-

plankton net (505 µm mesh) to a 2.5-m boom(aluminum rod). We deployed the net off theside of a boat using a bow-anchored rope forleverage (Figure 2). A zooplankton net (110 mmmesh), 1 m long with 1.5-m mouth opening,was attached directly to the outside frame ofthe ichthyoplankton net. Flowmeters weremounted across the opening of each net to de-termine the volume of water sampled per tow.The net was towed just below the water sur-face, and submerged objects were avoided by

FIGURE 1. San Francisco Estuary and the four study sites located in the central Sacramento–SanJoaquin Delta.

84 GRIMALDO ET AL.

manually maneuvering the boom. Ten-minutetows were conducted in each available habi-tat transect (marsh edge, shallow open-water,and river channel) per study site. Ichthy-oplankton were sampled about one neap tideper month during April–August 1998 andtwo neap tides per month during January–July1999. Zooplankton were sampled only oneneap tide per month during May–June 1998and March–June 1999. Water quality datawere measured either before or after each tow.Water temperature (oC) and specific conduc-tance (µS/cm) were measured just below thesurface (~0.5 m) using a YSI 85. Water clarity(cm) was estimated using a Secchi disk. Weestimated water depth (m) as the average ofdiscrete measurements by a boat-mountedfatho-meter. Tows were done irrespective oftidal stage or time of day. All samples wereindividually preserved in a 10% formalin so-lution and transferred to the laboratory foridentification to the lowest taxonomic leveland life stage. The densities of larval fish andzooplankton were calculated as the numberof organisms per volume (m3) of watersampled.

We did not do any shallow open-watertows at UM because the marsh edge abruptlytransitioned to the main channel. Further, wedid not sample any of the channels surround-ing MI because the site was almost entirelyenclosed by a levee. We assumed that any sub-stantial exchange of ichthyoplankton or zoop-

lankton between MI and outside channelswould occur mainly at the large northernbreach and would affect only a small portionof the site. Rather, we focused our examina-tion of ichthyoplankton and zooplankton vari-ability within MI habitats and its northern andsouthern regions because these two areas haddifferent hydrological and biological proper-ties (Lucas et al. 2002).

Data analysis

Intercorrelations among physical variableswere examined using Pearson product-mo-ment correlation tests to potentially reduce thenumber of highly correlated predicator vari-ables to a smaller number of uncorrelated vari-ables. This analytical step ensures a more con-cise comparison of spatial and temporaldynamics between taxa and environmentaldata (Turner et al. 1994). One-way analysis ofvariance (ANOVA) was used to determinewater column depth difference by habitattransect. Intersite and intermonth variation ofeach water quality variable (water tempera-ture, specific conductance, and water clarity)was examined using two-way ANOVA tests.Intrayear trends between each variable andconsecutive sample month were examinedusing Pearson product-moment correlationtests.

Multivariate analysis of covariance(MANCOVA) was used to test for significant

FIGURE 2. Diagram of side-towed ichthyoplankton net (505 mm mesh) and stylized cross-sectionaldiagram of subtidal sampling locations: (a) marsh edge, (b) shallow open-water, and (c) river channel.

tow rope

boom

a b c

SPATIAL AND TEMPORAL DISTRIBUTION OF NATIVE AND ALIEN ICHTHYOPLANKTON IN THREE HABITAT TYPES 85

differences between the most abundantichthyoplankton species (dependent vari-able) and three predicator variables: site,habitat type, and year. Water temperature(oC), specific conductance (mS/cm), and wa-ter clarity (cm) were included in the model ascovariates to test if ichthyoplankton abun-dance was influenced by water quality con-ditions at each habitat type. Interannualabundance differences were tested by includ-ing interaction terms in the MANCOVAmodel: (1) year × site, and (2) year × habitattype.

Monthly densities of native and alienfishes were plotted and visually examined forpatterns. Ichthyoplankton and zooplanktondensities were compared between north andsouth MI using parametric t-tests. Only fishspecies that contributed to 1% or more of thetotal relative abundance and occurred in 5%or more of the samples were used in statisti-cal analyses. Since we did not sample zoop-lankton at the same frequency as fish larvae,we only analyzed the five dominant taxa ofadult zooplankton to provide a general de-scription of how potential prey abundancesvaried by habitat, site, and season (three-wayANOVA). Fish and zooplankton densitieswere natural log-transformed (x + 1) prior toall statistical tests to reduce heteroscedasticityin the data.

To examine taxa-specific densities in re-lation to water quality variables, the mostabundant ichthyoplankton species were alsosubjected to canonical correspondence analy-sis (CCA) with the CANOCO software pro-gram (ter Braak and Smilauer 1998). Canoni-cal correspondence analysis is a directgradient analysis that is used to determine ataxon niche center (e.g., time and space)through calculation of weighted-average al-gorithms of abundance with environmentaldata (ter Braak and Verdonschot 1995). Nicheseparation among taxa is evaluated throughexamination of biplots, which depict ex-tracted synthetic gradients (ordination axes)of species abundances and environmentalvariables (ter Braak and Verdonschot 1995).The environmental variables included in thisanalysis were habitat type, water tempera-ture, specific conductance, Secchi depth, tidestage, and time of day.

Results

Environmental variables

Mean water column depths of marsh edge (1.60m ± 0.50 SD), shallow open-water (3.11 m ±1.20 SD), and river channel (10.17 m ± 3.11SD) habitat types were significantly different(P < 0.01). Water temperature and Secchi diskdepth varied significantly by month, but didnot vary among habitat types (Figure 3). In1999, specific conductance varied among habi-tat types and sample month. Both water tem-perature and specific conductance were sig-nificantly correlated with each consecutivesample month, r2 = 0.79 (P < 0.001) and r2 =0.45 (P < 0.001), respectively. Secchi disk depthwas not significantly correlated with consecu-tive sample month (r2 = 0.02; P > 0.05). Watertemperature and specific conductance werethe only two variables significantly correlatedwith each other (r2 = 0.21; P < 0.01). Since thesevariables were only weakly correlated andcould both independently affect larval fishdistribution, both were included in statisticalmodels as potential predicator variables.

Ichthyoplankton

We collected 25,863 individual fish larvae in240 tows (Table 1). Only six tows did not cap-ture larvae, suggesting that the gear method-ology provided high capture rates and ad-equate characterization of ichthyoplanktondistributions. Most larvae were identified tospecies, except for centrarchids, which wereidentified to genus and pooled for statisticalanalyses. A few fish (78) could only be identi-fied to the family Cyprinidae, while 9 otherlarvae could not be identified due to mutila-tion. Prickly sculpin Cottus asper, threadfinshad Dorosoma petenense, inland silversideMenidia beryllina , delta smelt Hypomesustranspacificus , bigscale logperch Percinamacrolepida, striped bass Morone saxatilis,golden shiner Notemigonus crysoleucas, andcentrarchid larvae accounted for 99% of thetotal number of fish larvae collected; these spe-cies and one family group were analyzed inMANCOVA and CCA analyses. All are alienspecies except for prickly sculpin and deltasmelt. Approximately 6,700 embryos repre-

86 GRIMALDO ET AL.

Marsh edge

Shallow

open-water

River channel

X

1998 1999

0

100

200

300

400

January February March April May June July0

100

200

300

400

0

10

20

30

0

10

20

30

0

20

40

60

80

100

120

April May June July August

0

20

40

60

80

100

120

January March May July

P > 0.17

P > 0.13

P > 0.72

P > 0.37

P < 0.01

P > 0.90

Wat

er t

emp

erat

ure

(oC

)S

pec

ific

condu

ctan

ce

(µS

/cm

) S

ecch

i d

epth

(c

m)

FIGURE 3. Mean values of physical variables by habitat type and month for each year. Significantdifferences among transects were tested using ANOVA models.

senting at least five species were collected dur-ing the study (Table 2); most of these embryoswere threadfin shad collected at MI.

The MANCOVA model indicated thatichthyoplankton abundance varied signifi-cantly by habitat type (P < 0.001), site (P <0.001), year (P < 0.05), and water temperature(P < 0.001). We did not detect a significant in-teraction effect between year and site (P > 0.34)or year and habitat type (P > 0.46), indicatingconsistent habitat-use by ichthyoplankton.Overall ichthyoplankton abundances variedbetween years, but this was expected giventhat we did not sample the same months eachyear. Although water temperature was indi-cated to have an effect on ichthyoplanktonabundance, we interpret this effect to indicateintra-annual differences with respect to sea-son, given that water temperature differed bymonth but not habitat type. Total ichthyo-plankton abundance was highest in marsh

edge habitat (Figure 4). Among sites, total abun-dance was highest at MI (Table 3). Within MI,the abundance of ichthyoplankton was sig-nificantly higher (P < 0.05) in the south (Fig-ure 5). Seasonal abundance trends in ichthy-oplankton were strongly patterned by recruit-ment peaks of native and alien fishes throughthe year (Figure 6). Generally, native fisheswere more abundant during winter andspring, whereas alien fishes were generallymore abundant during late summer. Goldenshiner and bigscale logperch were the excep-tion to this pattern, as they were most abun-dant during spring of both years.

The CCA indicated water temperature andhabitat type were important variables influ-encing fish abundance (Figure 7). Early sea-son recruits (prickly sculpin, delta smelt,golden shiner, and bigscale logperch) were lo-cated on the right side of canonical space, cor-responding with low water temperatures and

SPATIAL AND TEMPORAL DISTRIBUTION OF NATIVE AND ALIEN ICHTHYOPLANKTON IN THREE HABITAT TYPES 87

high specific conductance. In contrast, sum-mer recruits (inland silverside, threadfin shad,striped bass, and centrarchids) grouped in theopposite side of canonical space correspond-

ing with high water temperatures and lowspecific conductance. Delta smelt was centeredin the upper portion of the biplot correspond-ing with deep channel habitats. Inland silver-

TABLE 1. Frequency of occurrence (FO) and ranked summary of ichthyoplankton by total collected(N) during 240 tows between April 1998 and July 1999 in the Sacramento–San Joaquin Delta. Origin: A= Alien; N = native.

Taxa Origin FO (%) N

Threadfin shad A 43 11,822Prickly sculpin N 73 9,516Centrarchidaea A 32 2,260Inland silversides A 18 1,590Delta smelt N 24 216Bigscale logperch A 20 151Striped bass A 12 88Cyprinidae Unknown 9 78Golden shiner A 11 50American shad Alosa sapidissima A 5 34White catfish Ameiurus catus A 2 12Sacramento blackfish Orthodon microlepidotus N 3 10Unidentified 3 9Splittail Pogonichthys macrolepidotus N 5 8Shimofuri goby Tridentiger bifasciatus A 3 6Wakasagi Hypomesus nipponensis A 5 4Common carp Cyprinus carpio A <1 3Longfin smelt Spirinchus thaleichthys N <2 1Staghorn sculpin Leptocottus armatus N <1 1Goldfish Carassius auratus A <1 1Sacramento sucker Catostomus occidentalis N <1 1Total 25,863a includes Lepomis spp., Pomoxis spp., and Micropterus spp.

TABLE 2. Summary of embryos collected between 1998 and 1999 in the Sacramento–San JoaquinDelta. Substrate codes indicate if embryos were unattached (U) or attached to either emergent vegeta-tion (EV) or submerged aquatic vegetation (SAV). Site codes are VC = Venice Cut; UM = upperMandeville Tip; LM = lower Mandeville Tip; and MI = Mildred Island.

MonthsCommon name collected Site Habitat Substrate N

Common carp June VC Shallow open U 11water

Threadfin shad late April–June UM, MI Marsh edge EV, SAV 6,785Inland silverside May–June MI, VC Marsh edge EV 30Striped bass May LM River channel U 1Unidentified April VC Marsh edge SAV 23

cyprinidaeUnidentified May–July UM, MI, LM, VC Marsh edge, EV, U 115

shallow openwater

Total 6,966

88 GRIMALDO ET AL.

FIGURE 4. Mean ichthyoplankton abundance (±SE) by habitat type. Data were averaged from all sitessampled during 1998 and 1999. Asterisk indicates significant difference from other groups by one-wayANOVA (P < 0.05).

side, centrarchids, prickly sculpin, threadfinshad, and bigscale logperch were centeredmore towards the lower region of the biplotcorresponding to nearshore transects (edgeand shallow open-water habitats).

Zooplankton

The most common zooplankton groups col-lected were calanoid copepods and cladocer-ans. Of the calanoids collected, adult Eury-temora affinis was significantly (P < 0.001) moreabundant in marsh edge habitat and duringlate spring (i.e., April and May). Abundanceof adult Pseudodiaptomus forbesi did not varysignificantly among habitat types (P > 0.40)but was significantly more abundant duringJune (P < 0.001; Figure 8). Another calanoid,Sinocalanus doerrii, did not vary by habitat type(P > 0.23) or month (P > 0.21). Of the cladocer-ans, Daphnia spp. were more abundant in chan-nel habitat and during late spring. Bosminaspp. abundance did not vary among habitats,but they were higher during late spring aswell. Overall, zooplankton abundance did notdiffer among habitat types (P > 0.30) but washighest at MI compared to the other study sites(P < 0.05). Abundance in MI did not differ be-

tween the northern and southern regions (Fig-ure 5).

Discussion

Although estuarine environments are com-plex, many researchers have documented es-tuarine organisms aggregating within plumefronts (Govoni et al. 1989), during certain tidalstages (Melville-Smith et al. 1981; Drake andArias 1991; Joyeux 1999), and verticallythroughout the water column (Laprise andDodson 1989b; Kimmerer et al. 1998; Bennettet al. 2002). In this study, we found thatichthyoplankton abundance and compositiondiffered among subtidal horizontal transects.Specifically, we found that marsh edge habi-tats supported higher abundances of mostfishes, suggesting that they provided favor-able rearing conditions. Because the habitatuse results were consistent between years, webelieve that the results are robust and reflectstrategic life history traits. For example, marshedge habitats potentially are favorable rear-ing habitats because they typically supporthigh prey abundances (Cardinale et al. 1998)and provide cover from predators (Crowderand Cooper 1982; Rozas and Odum 1988).

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Marsh edge Shallow open-water River channel

Mea

n i

chth

yopla

nkto

n d

ensi

ty (

no./

m3) *

SPATIAL AND TEMPORAL DISTRIBUTION OF NATIVE AND ALIEN ICHTHYOPLANKTON IN THREE HABITAT TYPES 89

TA

BL

E 3

. Mea

n d

ensi

ties

of

the

eigh

t m

ost

abu

ndan

t ic

hthy

opla

nkto

n ta

xa b

y si

te a

nd h

abit

at t

rans

ect

betw

een

Ap

ril

1998

and

Ju

ly 1

999.

Mea

nd

ensi

ties

(no.

/m

3 * 1

00).

Ind

ivid

ual

icht

hyop

lank

ton

den

siti

es w

ere

sign

ific

antl

y d

iffe

rent

by

wet

land

sit

e an

d h

abit

at ty

pe

(MA

NC

OV

A; P

< 0

.05)

. The

aver

age

den

sity

for

all

eigh

t fis

hes

by w

etla

nd s

ite

wer

e ex

amin

ed u

sing

one

-way

AN

OV

A.

P-v

alue

Up

per

Low

erM

and

evill

e V

enic

e M

and

evill

e M

ildre

dTi

p C

ut

Tip

Isla

nd

Mar

shR

iver

Mar

shSh

allo

wR

iver

Mar

shSh

allo

wR

iver

Mar

shSh

allo

wed

gech

anne

led

geop

en-

chan

nel

edge

open

-ch

anne

led

geop

en-

wat

erw

ater

wat

er

Big

scal

e lo

gper

ch 0

.53

0.03

2.7

30.

020.

04 0

.81

0.05

0.00

0.7

50.

08C

entr

arch

idae

0.4

40.

09 4

.78

0.59

0.00

0.9

30.

260.

16 1

7.12

1.33

Del

ta s

mel

t 0

.40

0.20

0.5

00.

911.

08 0

.32

0.34

2.37

0.1

80.

09G

old

en s

hine

r 0

.22

0.14

0.9

10.

030.

24 0

.76

0.23

0.15

0.0

70.

00In

land

silv

ersi

de

4.1

50.

08 1

5.14

0.02

0.10

1.6

30.

000.

04 9

.50

0.05

Pric

kly

scul

pin

6.3

23.

82 3

0.68

15.4

117

.71

21.

9011

.93

10.7

6 2

8.77

10.5

5St

rip

ed b

ass

0.0

20.

00 0

.11

0.49

0.21

0.2

30.

380.

50 0

.08

0.17

Thr

ead

fin

shad

12.

829.

24 7

.56

8.62

2.00

13.

334.

031.

64 4

8.41

18.4

7A

vera

ge d

ensi

ty 1

9.62

39.

40 2

6.05

68.

98 <

0.00

1p

er s

ite

90 GRIMALDO ET AL.

Others have documented the importance ofedge habitats for larval fishes in other ecosys-tems (Paller 1987; Dewey and Jennings 1992).

We found a strong parallel betweenichthyoplankton distributions documented inthis study with distribution of juvenile andadult fishes observed in a companion inshore–offshore investigation (authors’ unpublished

data). Specifically, we found that the marshedge supported a distinct group of residentlarvae belonging to many of the same speciesthat were observed in these habitats as juve-niles and adults. The most prominent mem-bers of this marsh edge assemblage werecentrarchids, which are often key members ofedge habitats in other aquatic ecosystems

0

4000

8000

12000

16000

20000

0

0.4

0.8

1.2

1.6

Mildred Island

North South

Mea

n z

oopla

nkto

n

den

sity

(no./

m3)

Mea

n i

chth

yopla

nkto

n

den

sity

(no./

m3)

P > 0.16

P < 0.05

FIGURE 5. Comparisons between mean densities (±SE) of zooplankton (top) and ichthyoplankton(bottom) in southern and northern Mildred Island combined from 1998 and 1999. Significance (P < 0.05)was determined using Student’s t-tests.

SPATIAL AND TEMPORAL DISTRIBUTION OF NATIVE AND ALIEN ICHTHYOPLANKTON IN THREE HABITAT TYPES 91

(Crowder and Cooper 1982; Paller 1987; Rozasand Odum 1987; Killgore et al. 1989; Deweyand Jennings 1992; Cardinale et al. 1998;Johnson and Jennings 1998). Larvae of resi-dent fishes, such as bigscale logperch, goldenshiner, and inland silverside, were also veryabundant in marsh edge habitat, but the juve-nile stages of these fishes were most oftenfound in nearshore unvegetated habitats. Incontrast, juvenile centrarchids were mostlyfound within dense SAV beds (authors’ un-published data). We could not examine thequestion as to whether the interior of denseSAV exhibited higher or lower densities of thisedge-associated fauna. Paller (1987) foundthat centrarchids were more abundant in SAVcompared to the edge ecotone.

The marsh edge may also serve as impor-tant transitional habitat for many pelagic anddemersal fishes. For example, we found thathigh larval threadfin shad densities and manyof their embryos attached to vegetation in theedge transects. As juveniles and adults, thread-fin shad are most abundant in offshore areas

(authors’ unpublished data), indicating theymove from nearshore to offshore habitats dur-ing the larval–juvenile transition phase. Thisobserved ontogenetic shift indicates a strongdependence on marsh edge habitats, whichprovide both spawning substrate and rearinghabitat for threadfin shad. Benthic fishes, suchas native prickly sculpin larvae, were alsomore abundant in marsh edge compared toopen shallow-water and river channel. Thisfinding could indicate a sampling bias, par-ticularly in the river channel areas where asmaller proportion of the water column wassampled (i.e., only the surface) compared toopen shallow-water and marsh edge habitats.However, because prickly sculpin larvae areplanktonic (Wang 1986) and juveniles arefound in SAV (authors’ unpublished data), ourresults suggest that marsh edge habitats areimportant presettlement aggregation zones forthis species.

The abundance of native migratory larvalfishes (delta smelt and splittail ) varied be-tween years. This was expected given that

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Apri

l

May

June

July

August

January

Febru

ary

Marc

h

Apri

l

May

June

July

Native

Alien

Mea

n i

chth

yo

pla

nkto

n d

ensi

ty (

no

./m

3)

1998 1999

FIGURE 6. Mean densities (±SE) of native and alien fishes by month during the study period.

92 GRIMALDO ET AL.

mechanisms influencing recruitment of thesefishes have been found to vary inter-annu-ally and with upstream conditions (Sommeret al. 1997). Splittail larvae, for example, wasprimarily observed during 1998. In contrast,delta smelt larvae were present in 52% oftows (N = 92) during March–May 1999 com-pared to only 20% of tows (N = 30) duringthe same time period in 1998. Delta smelt isprimarily an annual species that spawns intidal freshwater habitats, migrating from thelower estuary to upstream habitats duringspring (Moyle et al. 1992). During 1999, Cali-fornia Department of Fish and Game moni-toring surveys revealed that adult delta smeltspawning was concentrated in the same geo-graphic region as our study sites (Dege andBrown 2004, this volume). We did not collectany delta smelt embryos to verify spawning

microhabitat; however, we did find that deltasmelt larvae were most abundant in riverchannel habitat, the only fish other thanstriped bass to exhibit such a distributionpattern. Residing in channels and using tidalcurrents may be an important strategy totransport delta smelt downstream to optimalrearing habitats.

Natural and restored wetlands

Given that ichthyoplankton densities at thenatural site were low and mostly dominatedby alien fishes, our results suggest that thenatural site offers little suitable habitat for na-tive fishes. It was our expectation that thenatural site would provide habitats with natu-ral conditions; however, it became evidentduring the study that natural conditions

FIGURE 7. Canonical correspondence analysis (CCA) plot of important environmental correlationvectors and species scores in the first two CCA dimensions for the eight most abundant fish speciescollected.

Specific conductance

Secchi depth

Habitat transect

Water temperature

prickly sculpin

inland silverside

threadfin shad

centrarchid spp.

delta smelt

striped bass

bigscale logperch

golden shiner

High temperature Low temperature

Shallow

DeepAlien

Native

SPATIAL AND TEMPORAL DISTRIBUTION OF NATIVE AND ALIEN ICHTHYOPLANKTON IN THREE HABITAT TYPES 93

might have been altered by the extensive colo-nization of E. densa in the subtidal regions ofthis study site. Evidence from the juvenile fishstudy (authors’ unpublished data) indicatesthat UM supports a large number of alienfishes. We suggest that centrarchids limit thenumber of native fish in the area, largelythrough predatory influences.

Among restored sites, ichthyoplanktonabundances were similarly low at VC andLM and highest at MI. We suspect that twokey mechanisms are responsible for thesefindings. First, MI supported enhanced lowertrophic conditions, such as higher zooplank-ton abundances (this study) and high zoop-lankton growth rates (Mueller-Solger et al.2002). These conditions likely promoted lar-val recruitment success. Second, larvae wereprobably more aggregated in MI comparedto LM and VC because MI has only one large

breach where ichthyoplankton can be ex-changed tidally with outside channels (Lucaset al. 2002). Our results provide some sup-port of this hypothesis as we found higherichthyoplankton abundance in the southernregion of MI, which experiences minimaltidal exchange influence (Lucas et al. 2002).In contrast, VC and LM were more open tomain channels (i.e., had fewer intact levees)and supported similar ichthyoplankton as-semblages. Because the subtidal regions ofVC and LM were much shallower than thoseof MI, these areas were colonized more ex-tensively by E. densa. As a result, we believethat we may have underestimated the num-bers and perhaps composition of larvae atthese sites, particularly of some centrarchids,since they are often found at higher densitieswithin SAV compared to the edge (Paller1987).

600

0

200

800

1000

400

600

0

200

800

1000

400

600

0

200

800

1000

400

600

0

200

800

1000

400

1500

0

500

2000

2500

1000

1500

0

500

2000

2500

1000

1500

0

500

2000

2500

1000

1500

0

500

2000

2500

1000

Mea

n d

ensi

ty (

no./

m3)

Mea

n d

ensi

ty (

no./

m3)

E.affinis*

P. forbesi

S. doerrii

E.affinis*

P. forbesi

S. doerrii

Bosmina

Daphnia**

xBosmina

Daphnia**

x Bosmina**

Daphnia*

Bosmina**

Daphnia*

x

E. affinis**

P. forbesi**

S. doerrii

E. affinis**

P. forbesi**

S. doerrii

Marsh

edge Shallow

open-water

River

channelMarch April May June

(i) (ii)

FIGURE 8. Mean densities (±SE) of adult calanoid copepods and cladocerans by habitat type (i) andmonth (ii). Data were averaged for 1998 and 1999. Significant differences of individual taxa by monthand habitat are denoted in the figure legend: * = P < 0.05, ** = P < 0.001.

94 GRIMALDO ET AL.

Seasonal abundance patterns

The recruitment patterns exhibited by ichthy-oplankton during our study period are con-sistent with those observed in other investi-gations conducted in the San FranciscoEstuary (Meng and Matern 2001; Feyrer 2004,this volume) and nearby adjoining regions(Rockriver 1998; Marchetti and Moyle 2000;Sommer et al. 2004, this volume). In particu-lar, our results substantiate the previous ob-servations that larvae of native fishes are earlyseason recruits compared to larvae of alienfishes, which were more common during sum-mer. The recruitment pattern differences arelikely initiated by environmental conditionsmeant to cue adult spawning migrations ormaximize larval success after hatching. Forexample, splittail year-class strength has beenfound to be positively linked with winter andspring inundation of upstream floodplainhabitats (Sommer et al. 1997). For other nativefishes, increased river flows are believed toprovide optimal rearing conditions (Meng etal. 1994; Jassby et al. 1995; Rockriver 1998;Marchetti and Moyle 2000; Meng and Matern2001).

Our results suggest that the bioenergeticsand agents of larval mortality (e.g., competi-tion and predation) vary between native andalien fishes, since trends in water quality pa-rameters and zooplankton assemblages varyseasonally. For example, we documented thatE. affinis, the main diet item of delta smelt(Nobriga 2002), was most abundant duringspring coincident with the presence of deltasmelt larvae. This suggests that survival of deltasmelt larvae may depend upon E. affinis densi-ties and growth may depend on water tempera-ture in spring. In contrast, recruitment of manyalien fishes hinges on densities of zooplank-ton common in the summer, mainly P. forbesi,and summer conditions, such as elevated wa-ter temperatures or low flow conditions.

Restoration opportunities

Whether native fishes can benefit throughhabitat restoration in this region of the deltaremains in doubt given that alien fishes werefound among all habitat types and sites. Weargue that the best restoration strategy may be

to create or restore wetlands that only floodduring winter and spring, the period whennative fishes spawn and recruit in the estu-ary. This strategy may not eliminate use byalien fishes altogether, but it may limit theiruse on a perennial basis, thereby limitingspawning or recruit success. Clearly, the needto understand the underlying mechanismsgoverning both native and alien recruitmentis warranted given the strong impetus to re-store wetlands in the estuary (CALFED 2000).

Acknowledgments

We gratefully acknowledge funding from theCALFED Bay-Delta Program and the Inter-agency Ecological Program. We thank C.Simenstad, J. Toft, D. Reed, and J. Cordell forhelpful study design suggestions. We thankB. Cavallo, M. Dege, F. Feyrer, B. Harrell, R.Kurth, Z. Matica, C. Messer, M. Nobriga, T.Veld-huizen, and J. Yokimizo for field assis-tance. We also thank L. Lynch, S. Skelton, andJ. Wang for identifying fish and zooplanktonsamples and G. O’Leary and J. Sears for sum-marizing the raw data. We are indebted to E.Santos for assistance with developing gearmethodology and dedicating countless hoursin the field as a boat operator. We also thankD. Holmgren, L. Rivard, and T. Sommer forreviewing the manuscript and providingmany insightful comments.

References

Atwater, B. F., S. G. Conard, J. N. Dowden, C. W. Hedel,R. L. MacDonald, and W. Savage. 1979. History,landforms, and vegetation of the estuary’s tidalmarshes. Pages 347–385 in T. J. Conomos, editor.San Francisco Bay: the urbanized estuary. Ameri-can Association for the Advancement of Science,San Francisco.

Bennett, W. A., and P. B. Moyle. 1996. Where have all thefishes gone? Interactive factors producing fish de-clines in the Sacramento-San Joaquin Estuary.Pages 519–542 in J. T. Hollibaugh, editor. San Fran-cisco Bay: the ecosystem. American Associationfor the Advancement of Science, San Francisco.

Bennett, W. A., W. J. Kimmerer, and J. R. Burau. 2002.Plasticity in vertical migration by native and alienestuarine fishes in a dynamic low-salinity zone.Limnology and Oceanography 47:1496–1507.

CALFED. 2000. Programmatic Record of Decision. Au-gust 28, 2000. CALFED Bay-Delta Program. Sac-

SPATIAL AND TEMPORAL DISTRIBUTION OF NATIVE AND ALIEN ICHTHYOPLANKTON IN THREE HABITAT TYPES 95

ramento, California. Available online at http://wwwcalfed.water.ca.gov/current/ROD.html.

Cardinale, B. J., V. J. Brandy, and T. M. Burton. 1998.Changes in the abundance and diversity of coastalwetland fauna from the open water/macrophyteedge towards shore. Wetlands Ecology and Man-agement 6:59–68.

Cohen, A. N., and J. T. Carlton. 1998. Accelerating inva-sion rate in a highly invaded estuary. Science279:555–558.

Cole, R. A., and J. R. MacMillan. 1984. Sampling larvalfish in the littoral zone of western Lake Erie. Jour-nal of Great Lakes Research 10:15–27.

Conomos, T. J., R. E. Smith, and J. W. Gartner. 1985.Environmental setting of San Francisco Bay.Hydrobiologia 129:1–12.

Crowder, L. B., and W. E. Cooper. 1982. Habitat struc-tural complexity and the interaction between blue-gills and their prey. Ecology 63:1802–1813.

Cushing, D. H. 1983. Are fish larvae too dilute to affectthe density of their food organisms? Journal ofPlankton Research 5:847–854.

Dege, M. D., and L. R. Brown. 2004. Effect of outflow onspring and summertime distribution and abun-dance of larval and juvenile fishes in the upper SanFrancisco Estuary. Pages 49–65 in F. Feyrer, L. R.Brown, R. L. Brown, and J. J. Orsi, editors. Early lifehistory of fishes in the San Francisco Estuary andwatershed. American Fisheries Society, Symposium39, Bethesda, Maryland.

Dewey, M. R., and C. A. Jennings. 1992. Habitat use bylarval fishes in a backwater lake of the upper Mis-sissippi River. Journal of Freshwater Ecology 7:363–371.

Dodson, J. J., J. C. Dauvin, R. G. Ingram, and B.d’Anglejan. 1989. Abundance of larval rainbowsmelt (Osmerus mordax) in relation to the maxi-mum turbidity zone and associated macroplanktonfauna of the middle St. Lawrence Estuary. Estuar-ies 12:77–81.

Drake, P., and A. M. Arias. 1991. Ichthyoplankton of ashallow coastal inlet in south-west Spain: Factorscontributing to colonization and retention. Estua-rine, Coastal and Shelf Science 32:347–364.

Feyrer, F. 2004. Ecological segregation of native and alienlarval fish assemblages in the southern Sacramento–San Joaquin Delta. Pages 67–79 in F. Feyrer, L. R.Brown, R. L. Brown, and J. J. Orsi, editors. Early lifehistory of fishes in the San Francisco Estuary andwatershed. American Fisheries Society, Symposium39, Bethesda, Maryland.

Gadomski, D. M., and C. A. Barfoot. 1998. Diel anddistributional abundance patterns of fish embryosand larvae in the lower Columbia and Deschutesrivers. Environmental Biology of Fishes 51:353–368.

Govoni, J. J., D. E. Hoss, and D. R. Colby. 1989. Thespatial distribution of larval fishes about the Mis-sissippi River plume. Limnology and Oceanogra-phy 34:178–187.

Houde, E. D., and E. S. Rutherford. 1993. Recent trends

in estuarine fisheries-predications of fish produc-tion and yield. Estuaries 16:161–176.

Houde, E. D., and J. D. Alpern-Lovdal. 1985. Patternsof variability in ichthyoplankton occurrence andabundance in Biscayne Bay, Florida. Estuarine,Coastal and Shelf Science 20:79–103.

Houde, E. D. 1987. Fish early life history dynamics andrecruitment variability. Pages 17–29 in R. D. Hoyt.10th annual larval fish conference. American Fish-eries Society, Symposium 2, Bethesda, Maryland.

Jassby, A. D., W. J. Kimmerer, S. G. Monismith, C. Ar-mor, J. E. Cloern, J. E. Powell, T. M. Schubel, and T.J. Vendlinski. 1995. Isohaline position as a habitatindicator for estuarine populations. EcologicalApplications 5:272–289.

Johnson, B. L., and C. A. Jennings. 1998. Habitat asso-ciations of small fishes around islands in the upperMississippi River. North American Journal of Fish-eries Management 18:327–336.

Joyeux, J. C. 1999. The abundance of fish larvae in estu-aries: within-tide variability at inlet and immigra-tion. Estuaries 22:899–904.

Kimmerer, W. J., J. R. Burau, and W. A. Bennett. 1998.Tidally-orientated vertical migration and popula-tion maintenance of zooplankton in a temperateestuary. Limnology and Oceanography 43:1697–1709.

Killgore, K. J., R. P. Morgan, I. I., and N. B. Rybicki.1989. Distribution and abundance of fishes associ-ated with submersed aquatic plants in the PotomacRiver. North American Journal of Fisheries Man-agement 9:101–111.

Laprise, R., and J. J. Dodson. 1989a. Ontogenetic changesin the longitudinal distribution of two taxa of lar-val fish in a turbid well-mixed estuary. Journal ofFish Biology 35:39–47.

Laprise, R., and J. J. Dodson. 1989b. Ontogeny and im-portance of tidal vertical migrations in the reten-tion of larval smelt Osmerus mordax in a well-mixedestuary. Marine Ecology Progress Series 55:101–111.

Lucas, L. V., J. E. Cloern, J. K. Thompson, and N. E.Monsen. 2002. Functional variability of shallowtidal habitats in the Sacramento-San Joaquin Delta:restoration implications. Ecological Applications12:1528–1547.

Marchetti, M. P., and P. B. Moyle. 2000. Spatial andtemporal ecology of native and introduced fish lar-vae in lower Putah Creek, California. Environmen-tal Biology of Fishes 58:75–87.

Melville-Smith, R., D. Baird, and T. Wooldridge. 1981.The utilization of tidal currents by the larvae of anestuarine fish. South Africa Journal of Zoology16:10–13.

Meng, L., P. B. Moyle, and B. Herbold. 1994. Changes inabundance and distribution of native and intro-duced fishes of Suisun Marsh. Transactions of theAmerican Fisheries Society 123:498–507.

Meng, L., and S. A. Matern. 2001. Native and intro-duced larval fishes of Suisun Marsh California: the

96 GRIMALDO ET AL.

effects of freshwater flow. Transactions of theAmerican Fisheries Society 130:750–765.

Moyle, P. B., B. Herbold, D. E. Stevens, and L. W. Miller.1992. Life history and status of delta smelt in theSacramento–San Joaquin Estuary, California.Transactions of the American Fisheries Society121:67–77.

Mueller-Solger, A. B., A. D. Jassby, and D. C. Mueller-Navarra. 2002. Nutritional quality of food resourcesfor zooplankton (Daphnia) in a tidal freshwatersystem (Sacramento–San Joaquin River Delta).Limnology and Oceanography 47:1468–1476.

Nobriga, M. L. 2002. Larval delta smelt diet composi-tion and feeding incidence: environmental andontogentic influences. California Fish and Game88:149–164.

Paller, M. H. 1987. Distribution of larval fish betweenmacrophyte beds and open channels in a south-eastern USA floodplain swamp. Journal of Fresh-water Ecology 4:191–200.

Rakocinski, C. F., J. Lyczkowski-Shultz, and S. L.Richardson. 1996. Ichthyoplankton assemblagestructure in Mississippi Sound as revealed by ca-nonical correspondence analysis. Estuarine,Coastal, and Shelf Science 43:237–257.

Rockriver, A. K. C. 1998. Spatial and temporal distribu-tion of larval fish in littoral habitats of the CentralValley, California. Master’s thesis. California StateUniversity, Sacramento.

Rozas, L. P., and W. E. Odum. 1987. Fish andmacrocrustacean use of submerged plant beds intidal freshwater marsh creeks. Marine EcologyProgress Series 38:101–108.

Rozas, L. P., and W. E. Odum. 1988. Occupation ofsubmerged aquatic vegetation by fishes: testing theroles of food and refuge. Oecologia 77:101–106.

Sommer, T., R. Baxter, and B. Herbold. 1997. The resil-ience of splittail in the Sacramento–San JoaquinEstuary. Transactions of the American Fisheries So-ciety 126:961–976.

Sommer, T. R., W. C. Harrell, R. Kurth, F. Feyrer, S. C.Zeug, and G. O’Leary. 2004. Ecological patterns ofearly life stages of fishes in a large river-floodplainof the San Francisco Estuary. Pages 111–123 in F.Feyrer, L. R. Brown, R. L. Brown, and J. J. Orsi,editors. Early life history of fishes in the San Fran-cisco Estuary and watershed. American FisheriesSociety, Symposium 39, Bethesda, Maryland.

ter Braak, C. J. F., and P. Smilauer. 1998. CANOCOreference manual and user’s guide to CANOCOfor Windows: software for canonical communityordination (version 4). Microcomputer Power.Ithaca, New York.

ter Braak, C. J. F., and P. F. M. Verdonschot. 1995. Ca-nonical correspondence analysis and related mul-tivariate methods in aquatic ecology. Aquatic Sci-ences 57:255–289.

Turner, T. F., J. C. Trexler, G. L. Miller, and K. E. Toyer.1994. Temporal and spatial dynamics of larval andjuvenile fish abundance in a temperate floodplainriver. Copeia 1:174–185.

Wang, J. C. S. 1986. Fishes of the Sacramento-San JoaquinEstuary and adjacent waters, California: a guide totheir early life histories. Interagency Ecological Pro-gram. California Department of Water Resources,Technical Report 9, Sacramento, California.