AN ECOSYSTEM MODEL OF THE SACRAMENTO-SAN JOAQUIN

DELTA AND SUISUN BAY, CALIFORNIA USA

______

A Project

Presented

to the Faculty of

California State University, Chico

______

In Partial Fulfillment

of the Requirements for the Degree

Master of Science

in

Environmental Science:

Professional Science Master Option

______

by

Marissa Bauer

Summer 2010 AN ECOSYSTEM MODEL OF THE SACRAMENTO-SAN JOAQUIN

DELTA AND SUISUN BAY, CALIFORNIA, USA

A Project

by

Marissa Bauer

Summer 2010

APPROVED BY THE DEAN OF GRADUATE STUDIES AND VICE PROVOST FOR RESEARCH:

Katie Milo, Ed.D.

APPROVED BY THE GRADUATE ADVISORY COMMITTEE:

John Nishio, Ph.D., Chair

Larry Brown, Ph.D.

Michael P. Marchetti, Ph.D. ACKNOWLEDGEMENTS

First and foremost I thank Dr. Larry Brown for his unfailing patience, support and guidance through this project. Many thanks are due to Dr. Howard Townsend, my greatest Ecopath mentor whose sense of humor made the world of ecosystem modeling a tolerable beast. I also thank my committee members, Dr. John Nishio and Dr. Michael

Marchetti for their advice and comments on drafts.

A huge thank-you to those who graciously contributed their valuable time, data expertise and knowledge, I could not have put this model together without your help:

Fred Feyrer (USBR), Wim Kimmerer (SFSU), Bill Bennett (UCD), Jan Thompson

(USGS), and Erik Loboschefsky (UCD). Thanks also to the California Department of

Fish and Game and the Interagency Ecological Program for the collection and maintenance of long term data sets.

I am beyond fortunate to have the unflinching support and never-ending encouragement of my parents, Jim and Cheryl Bauer, without whom I would not be where I am today. Thank you so much.

To Frank, who made me stay inside and type those last few pages. I love you.

______

This work was supported by funding from the Interagency Ecological

Program to Larry Brown. The work was conducted as part of a working group convened at the National Center for Ecological Analysis and Synthesis in cooperation with the

Interagency Ecological Program. iii TABLE OF CONTENTS

PAGE

Acknowledgements...... iii

List of Tables ...... v

List of Figures...... vi

List of Symbols and Nomenclature...... vii

Abstract...... ix

CHAPTER

I. Introduction ...... 1

A Need for Ecosystem Modeling...... 4

II. Methods...... 7

Introduction to the Ecopath with Ecosim Methodology...... 7 Ecopath ...... 9 Ecopath Input Parameters for the Delta Model...... 14 Parameter Estimate Methodology...... 18 Diet Composition Matrix ...... 23 Balancing the Ecopath Model...... 28

III. Results ...... 29

IV. Discussion ...... 37

Future Work...... 41

References...... 43

iv LIST OF TABLES

TABLE PAGE

1. Ecopath Parameters for the Preliminary 1982 Delta Model ...... 15

2. Ecopath Parameters for the Preliminary 1982 Delta Model ...... 16

3. Diet Matrix for the Preliminary 1982 Delta Ecopath Model ...... 24

4. Diet Matrix for the 1982 Delta Ecopath Model Continued ...... 25

5. Diet Matrix for the 1982 Delta Ecopath Model Continued ...... 26

6. Diet Composition Sources for each Functional Group...... 27

7. Ecological Characteristics for each Functional Group ...... 35

8. Summary Statistics of the Delta Model ...... 36

9. Trophic Transfer Efficiencies for each Trophic Group (%) ...... 36

v LIST OF FIGURES

FIGURE PAGE

1. Sacramento-San Joaquin Delta and Suisun Bay, California USA ...... 9

2. Visual representation of Ecopath with Ecosim model: Data needs, modeling abilities, research and application are linked...... 10

3. A simplified food-web model of the Delta in 1982...... 34

vi LIST OF SYMBOLS AND NOMENCLATURE

B: EwE Biomass parameter (t · km2)

BA: Biomass Accumulation (t · km2)

CDWR: California Department of Water Resources

cm: centimeter (metric unit of measurement)

CVP: Central Valley Project

DC: Diet Composition (%)

DOC: Dissolved organic carbon

DFG: California Department of Fish and Game

Ei: net migration rate (emigration – immigration)

EE: Ecotrophic efficiency

EwE: Ecopath with Ecosim

F: instantaneous rate of fishing mortality year-1

FMT: Fall Midwater Trawl Survey (performed by DFG)

IEP: Interagency Ecological Program

K: k parameter of the von Bertalanffy growth equation; K = (dL/dt)/( L∞-L) km: kilometer

L∞: theoretical Length at Infinite age

Lmat/L∞: Length at Maturity divided by theoretical Length at Infinite age

M: instantaneous rate of natural Mortality (year-1)

m: meter (metric unit of measurement) vii P: Production of biomass by a functional group (t/km2 year-1)

POC: Particulate organic carbon

P/B: Production per unit Biomass parameter (year-1) (equivalent to Z)

P/Q: Production/Consumption (year-1)

POD: Pelagic Organism Decline

Q: Consumption of biomass by a functional group (t·km2 year-1)

Q/B: Consumption per unit Biomass parameter (year-1)

SWP: State Water Project t · km-2 metric tons per square kilometer, units used for biomass inputs

USFWS: US Fish and Wildlife Services

USGS: US Geological Survey

VBGF: von Bertalanffy growth function

Wmat/W∞: Weight at Maturity divided by theoretical Weight at Infinite age

W∞: Weight at Infinite age

YI Yield, catch in weight, or fishery harvest (note that Yi = FiBi where F is

the fishing mortality rate) (year-1)

Z: instantaneous rate of total mortality (year-1)

viii ABSTRACT

AN ECOSYSTEM MODEL OF THE SACRAMENTO-SAN JOAQUIN

DELTA AND SUISUN BAY, CALIFORNIA USA

by

Marissa Bauer

Master of Science in Environmental Science:

Professional Science Master Option

California State University, Chico

Summer 2010

The San Francisco Estuary is a highly exploited system that has been drastically changed from its natural state through urban development, agriculture, and water management. Ecopath with Ecosim (EwE 6) modeling software was used to explore direct and indirect drivers of ecosystem dynamics of the Sacramento-San Joaquin

Delta and Suisun Bay, with emphasis on the Pelagic Organism Decline species (delta smelt (Hypomesus transpacificus), threadfin shad (Dorosoma petenense), longfin smelt

(Spirinchus thaleicthys), and striped bass (Morone saxatilis)). An Ecopath model was built to simulate the Delta and Suisun Bay ecosystem for the year 1982 based on best available data. The Ecopath modeling approach incorporates trophic interactions and community food web structure to help understand the dynamic interactions between trophic groups. The model uses parameters such as biomass, production, and mortality

ix estimates of 40 functional groups, including birds, fish, benthic invertebrates, zooplankton, phytoplankton, and detritus. Model parameters were calculated from published and unpublished data taken from the literature or estimated by local experts when appropriate.

The balanced Delta Ecopath model gave some surprising results about the food-web within the Delta. Of the total system biomass (270.5 t km-2) 37 % is from trophic level 3, which consists primarily of the fish species. Trophic level 1 including both detritus and phytoplankton composes 55 % of total system biomass. The final 8 % of total system biomass is composed of second and fourth trophic levels. Trophic transfer efficiencies for the Delta average 8.3%, with the highest trophic transfer efficiency at trophic level 1 (17.6%) and the lowest at trophic level 4 (3.7%).These results are consistent with the general ecological principle of a decreasing trend in transfer efficiency with increasing trophic level. The Delta Ecopath model also brings attention to the fact that the microbial and detrital loops may play an important role in the Delta food web and that further investigation into these pathways may prove useful. It was also apparent that the current Delta model was attempting to describe both the pelagic and littoral food webs of the Delta system and that while these two food webs share some similar species composition, further development of the model is needed to better understand and address this issue.

The development of an ecosystem-wide Ecopath model of Suisun Bay and the

Sacramento-San Joaquin Delta is a major step in exploratory ecosystem modeling of this highly altered system. The basic input parameters are a solid foundation for the

x development of the more in-depth Ecosim model, the time-dynamic simulation model driven by the mass-balanced Ecopath model. The EwE modeling approach can be used to address multiple areas of management concern. For example, it can be used to evaluate ecosystem effects of fishing mortality, explore management policy options, predict movement and accumulation of contaminants, and model effects of environmental changes.

xi CHAPTER I

INTRODUCTION

The San Francisco Estuary, California is the largest estuary on the west coast of the United States and supports several endangered species, as well as many commercially and recreationally important fish species that spend all or part of their life cycle in the estuary (Bennett 1995; Matern et al., 2002; Brown 2007; Rosenfield and

Baxter 2007). The organisms in the ecosystem are subject to a wide range of stressors as the San Francisco Estuary (herein referred to as the Estuary) is a highly exploited system that has been drastically changed from its natural state through urban development, agriculture, and water management (Nichols 1986; Bergamaschi et al., 2001, Kimmerer

2004).

The Estuary, including San Francisco Bay, San Pablo Bay, Suisun Bay, and the Sacramento-San Joaquin Delta has a long history of ecosystem decline beginning with the influx of settlers to California during the 1800’s (Brown and Moyle 2005). The

Estuary was a natural focal point for shipping, commerce and settlement, followed by agriculture and urbanization. As development proceeded through the decades, Estuary modification soon came in the form of state and federal water projects designed to meet water demands in the Central Valley and Southern California. The Estuary has also been assailed by multiple invasive species (Cohen and Carlton 1998), and has lost approximately 95% of its historical wetlands (Mount 1995, Sommer et al., 2007). These

1 2 changes and other associated environmental alterations to important habitat for San

Francisco Estuary fishes have been associated with declines in native fish species (Moyle

2002, Brown and Moyle 2005).

The Interagency Ecological Program (IEP) is a collaboration of several state and federal agencies that has been instrumental in collecting ecological monitoring data within the Estuary since the 1970’s. An invaluable asset to ecological study, long-term

IEP datasets have been used to understand trends in the Estuary and have played a key role in recognizing the recent ecological shifts. A frequently used IEP dataset, the fall midwater trawl survey (FMT), conducted by the California Department of Fish and Game

(DFG) has been monitoring pelagic fish abundance within the delta since 1967. Prior to

2000, the four most abundant resident pelagic fishes surveyed by the trawl included two native species, delta smelt (Hypomesus transpacificus) a federal listed endangered species, and longfin smelt (Spirinchus thaleicthys) a state listed endangered species, and two introduced species, threadfin shad (Dorosoma petenense) and striped bass (Morone saxatilis)).

Environmental conditions within estuaries are constantly changing and annual and seasonal variability in a species population is often a product of this dynamic habitat.

For example, delta smelt, longfin smelt, threadfin shad and striped bass have all shown fluctuations in abundance over time attributed to a variety of factors, including amount of freshwater flow into the Estuary (Kimmerer 2002, Nobriga et al., 2008). Beginning in the early 2000’s, abundance indices of the four pelagic fishes showed concurrent and alarming declines in multiple surveys (Sommer et al., 2007). Between survey years,

2002-2005 delta smelt and age-0 striped bass experienced record abundance lows, while 3 threadfin shad and longfin smelt experienced near record lows (Sommer et al., 2007). The most alarming aspect of this decline was that it occurred during a series of relatively wet years, when these fishes would be expected to experience good survival and increasing populations (Sommer et al., 2007). The multi-species decline, termed the Pelagic

Organism Decline (POD), has been verified by statistical analyses (Thomson et al., 2010) and is under investigation by IEP and others, including multiple state, federal, and university researchers.

In order to address and define potential causes of the POD, the IEP-POD management team has developed a conceptual model including the following 4 major areas of focus (Baxter et al., 2008): 1) Prior fish abundance, meaning stock-recruitment effects between adult population and juvenile recruitment; 2) Habitat, including water quality, disease, and contaminants; 3) Top-down effects, including mortality due to predation and entrainment from water project pumps and 4) Bottom up effects, the effects of food web interactions, including reduced food availability and nutritional quality, on populations.

The potential causes of the POD are multifaceted and interconnected. An ecosystem model could be used to explore the trophic interactions of the POD species with other species in the ecosystem, providing a viable research tool for addressing top down and bottom up effects of this complex problem. Such a model, with input parameters grounded in real data, could be used to explore many different hypotheses that have been proposed to help explain the decline. The development of an ecosystem modelling approach for the Delta is also consistent with IEP’s need to explore different research directions regarding the POD. 4

A Need for Ecosystem Modeling

Ecosystem modeling is a research method that can be used to project ecosystem responses to disturbances and to identify processes and relationships within the ecosystem that are not easily measured, have been overlooked, or are unknown

(Fulton et al., 2003; Field et al., 2006; Chen et al., 2008). Ecosystem models that are driven by long-term data sets, such as those generated by the modeling software tool,

Ecopath with Ecosim, have enormous potential to be used as an adaptive management tool by policy makers, scientists, and managers (Ma et al., 2010). A model that is built from the best available data can also give insight into gaps or weaknesses in ecological knowledge of a system that would benefit from additional research.

Using model simulations to characterize the structure and function of an ecosystem can identify critical processes and relationships that are not initially apparent to those studying or managing the system. This, in turn, can lead to the development of new monitoring protocols or changes in management or policy, such as revisions to an ecosystem restoration plan. Ecosystem models can also be used to quantify the direct and indirect effects of anthropogenic changes to a system by allowing the researcher to compare a baseline model built from actual survey data to future projections generated by the model. Ecological forecasting by models is a useful management and policy tool

(Clark et al., 2001).

Ecopath with Ecosim (EwE), is a suite of static and spatial modeling applications based on an approach initiated by Polovina (1984) and further developed by researchers at the University of British Columbia’s Fishery Centre (Pauly et al., 2000).

The EwE suite of models includes the Ecopath, Ecosim, Ecospace, and Ecotrace 5 modules. Ecopath is used to organized ecosystem data including trophic interactions of the species in the system, population size, and other biological data. This data is then used to create a mass-balance “snapshot” of the ecosystem. Ecopath is the foundation for all other EwE modules, making it a key component of the EwE software package. Ecosim builds dynamic predictions by combining the Ecopath data with foraging arena theory, providing dynamic simulations of ecosystem changes. Ecospace is a spatial dynamic module used to incorporate special habitat preferences and predation risks. Ecotrace can be used to explore ecosystem effects of contaminants. To date, over 100 ecosystem models using the EwE software have been published. The EwE modeling software was initially built for estimating biomass and food consumption rates of multiple functional groups, (e.g. species) in a steady-state aquatic system (Christensen and Pauly, 1992).

EwE is grounded in theoretical ecology applications of analyzing energy flows within a system (Odum, 1969; Ulanowicz, 1986). The ability to examine more than one species at a time in a model was one of the key features that set Ecopath apart from its single- species model counterparts.

EwE software has been used in a wide range of ecosystem analyses including, food web research, fisheries catch analysis, and ecosystem response to anthropogenic change. EwE was used to identify keystone functional groups within Prince William

Sound and the Gulf of Thailand (Libralato et al., 2006), to look at the interactions of mercury with marine mammals and climate change (Booth and Zeller 2005), and to explore the roll of introduced species in a lake ecosystem (Villanueva et al., 2008). A detailed EwE ecosystem model for the Chesapeake Bay containing 45 functional groups at all trophic levels is being developed to meet a management need for a quantified 6 estimate of trophic pathways within the Bay (Christensen et al., 2008). EwE will help analyze how one species affects another within the food web and how various species impact both target and non-target species. The Sacramento-San Joaquin Delta and Suisun

Bay Ecopath model is being developed with similar goals in mind. The objective of this project is to develop an Ecopath model for Suisun Bay and the Sacramento-San Joaquin

Delta for 1982. This model will provide a basis for application of the entire suite of EwE software. CHAPTER II

METHODS

The Ecopath model describes the Sacramento-San Joaquin Delta and Suisun

Bay (herein referred to as the Delta) portion of the Estuary (Figure 1). The modeled area

of the Delta was defined by the extent of DFG’s FMT survey, which is the main source of

fish information used in the model, and covers approximately 244 km2.

The Delta receives approximately 50 percent of California’s freshwater

streamflow and collects runoff from 40 percent of California’s total land area (Ingebritsen

et al., 2000). The state operated State Water Project (SWP) and federally operated Central

Valley Project (CVP) export up to 7.5 million acre-feet per year from two major pumping

facilities in the south Delta (CDWR 1993). More than twenty million people,

approximately half of California’s population, get at least part of their drinking water

from the Delta. California’s extensive agriculture industry also depends heavily on water

exports from the Delta.

Introduction to the Ecopath with Ecosim Methodology

Ecopath’s mass-balanced modelling approach incorporates trophic interactions and community food web structure to help understand the dynamic interactions between trophic groups (Figure 2). Once all basic input parameters have been determined by the user, the Ecopath model balances the input and output of each

7 8

Figure 1. Sacramento-San Joaquin Delta and Suisun Bay, California USA.

functional group with two a linear equations of production and consumption, using varying rates of respiration for adjustment. The mass-balanced linear equations of

Ecopath are then re-expressed as coupled differential equations, so that they can be used

9 by the Ecosim model to simulate changes in the different functional groups over time.

Multiple model runs are generated by the Ecosim model and compared with actual time- series data, with the best fit being chosen to represent the system (Christensen and Pauly,

1992, Christensen and Walters, 2004). Once the initial values are generated, changes can be made to various components of the Ecopath or Ecosim model to examine the consequences of such changes in the ecosystem. For example, one could increase the production of amphipods (Data, Box Biol) and project forward in time (Model, Box Time

Dynamic) to determine how other functional groups in the ecosystem will be affected

(Research) and how that result might impact other areas of interest e.g. policy, management decisions (Application) (Figure 2).

Ecopath

The Ecopath model uses mass-balance principles to link functional biomass

groups within a dynamic system to create a static snapshot of the resources and energy

flows within the ecosystem (Christensen and Pauly, 1992; Pauly et al., 2000; Christensen

and Walters, 2004). A functional group can represent a group of tropically similar

species, a single species, or a group of species split into age categories (“multi-stanza

groups”). Ecopath is used to specify initial conditions for biomass (B) and production (P) of each functional group. By defining the biomass and production of each functional group at a single point in time, Ecopath provides a static, time-invariant description of the ecosystem.

The parameters of an Ecopath model are based on satisfying two principle equations. The first equation describes how the production term for each group can be

Figure 2. Visual representation of Ecopath with Ecosim model: Data needs, modeling abilities, research and application are linked. 10

11

divided for a user defined time period. For most EwE models, including the present one,

production is accounted for on an annual basis.

Production = catch + predation + net migration + biomass accumulation + other mortality

In mathematical terms:

n i )/( iiii   j )/( j DCBQBEYEEBPB ji (1) j1

where for each functional group (i),

-2  Bi is total biomass (t · km ) during the period of question

-1 -2 -1  (P/B)i is the production to biomass ratio (year ), where P (t · km year ) is the production of functional group i and B (t · km-2) is the biomass of functional group i

 EEi is the ecotrophic efficiency, defined as the fraction of the production that is

consumed within or harvested from the system

 YI is the yield, catch in weight, or fishery harvest (note that Yi = FiBi where F is

the fishing mortality rate) (year-1).

 Ei is the net migration rate (emigration – immigration)

-1  (Q/B)j is the food consumption per unit biomass for consumer j (year ), where

Qj (t · km-2 year-1) is the biomass of prey (i) consumed by the consumer or predator j and

-2 Bj (t · km ) is the biomass of the consumers or predators of (i)

 DCji is the average fraction of i in the diet of j (note that DCji = 0 when j does

not eat i)

It is important to note that production refers to the generation of tissue by a

functional group over the time period specified by the model. In most cases, this mostly

12

equates to changes in the number of individuals rather than the growth of individual

organisms. Mass-balance constraints assume that production over biomass (P/B) is equal

to the instantaneous rate of total mortality (Z) used by fisheries biologists (Allen, 1971).

Total (fish) mortality is a commonly used term in fisheries science and refers to the

removal of fish from a stock, or population. Total mortality is comprised of two types of

mortality, natural mortality (M) and fishing mortality (F). Natural mortality refers to

disease, cannibalism, predation, or any other source of naturally occurring mortality.

Fishing mortality is caused by removal of fish from the population by any type of fishing

activity. Thus, (M) and (F) are instantaneous rates that sum to Z (e.g. M+F=Z=P/B). It

follows, that under the stipulations assumed for the construction of the mass balanced

Ecopath model, estimates of Z can be used interchangeably with the P/B input. If there is

no fishing mortality on a functional group, then Z will only include sources of natural

mortality. For example, zooplankton have a high P/B ratio because they live a short time

and are rapidly eaten or removed from the system, thus the high ratio helps illustrate their

high mortality rate over multiple generations during a one year period. In contrast, a bird

lives a much longer life and has relatively fewer predators, thus having a lower P/B ratio.

The model is taking a one year “snapshot” of the ecosystem, thus the (t km-2) units in the

P/B ratio cancel out (t km-2 year-1/t km-2), leaving units of year-1.

The consumption over biomass (Q/B) ratio is the consumption or intake of

food by a functional group over a specified time period (e.g., one year). Ecopath

expresses absolute consumption as an energy flow expressed in (t · km-2 year-1), while the subsequent Q/B is on a per year basis (Christensen et al., 2005). Thus, the Q/B input

13

parameter entered into Ecopath for each functional group in this model is expressed as

Q/B year-1.

For each functional group, Ecopath requires input of DC, Y, and three of the

following four parameters, B, P/B, QB, and EE. If only 3 of the 4 parameters are known,

mass-balance principles are used to estimate the fourth parameter (Christensen et al.,

2005). If all four parameters are known, then Ecopath can be used to estimate biomass

accumulation rate, or net migration rate. Equation (1) enables the estimates of biomass,

production, and consumption to be used in constructing a diagram of energy flow

(Mackay, 1981, Ulanowicz, 1986).

The second equation used in Ecopath is based on the principle of conservation of matter within a functional group, and is as follows:

Consumption = production + respiration + unassimilated food (2)

The equation balances the energy flows into and out of each functional group and develops the mass-balance idea that consumption by a group equals the production by the group, plus “waste.” The energy flowing into a group equals that flow out of the group, and it does so for all functional groups. This mass-balance restriction is necessary

to constrain the initial parameterization of the model, so that future Ecosim simulations

are reasonable and grounded in data, as much as possible. The constraint also provides the basis for estimations of the unknown parameter in Equation (1), if only 3 of the 4 are

available, as described above.

14

Ecopath Input Parameters for the Delta Model

Basic input parameters for the 40 user-defined functional groups of the 1982

Delta Ecopath model are presented in Tables 1 and 2. The 40 functional groups that

contain species or groups of species found in the Delta were chosen because of their

known or potential importance to or on the POD fish species. For example, the four zooplankton species were singled out as functional groups because of their know importance as POD fish food. Mysids were included because of their importance in the longfin smelt’s diet. Functional group inclusion in the model was based on suggestions made by Delta experts who have extensive, long-term knowledge of the system similar to what was done in Mac Nally et al., (2010). While other fish, invertebrates, and functional groups could be included in the model, it was important to limit the inputs to the model, at least initially, so that some basic inferences could be drawn. Ecopath allows for great flexibility in that the user has the ability to make changes to the model if desired,

meaning that if another functional group were deemed necessary/unnecessary, it could be

added or subtracted with minimal effort. Model parameters in Tables 1 and 2 were

calculated from publicly available data, unpublished data, taken from the literature, or, if

necessary, estimated.

The modeled area of the Delta was defined by the extent of DFGs FMT

survey, which is the main source of fish information used in the model. The FMT

samples the pelagic habitat of the Delta and was originally started to monitor the potential

effects of the CVP and SWP on the survival of young striped bass. The USFWS beach

seine survey samples the littoral habitat of the Delta and was initiated to monitor the

15

TABLE 1. ECOPATH PARAMETERS FOR THE PRELIMINARY 1982 DELTA MODEL. Biomass Parameter Functional Group (t • km-2 ) Source EE Piscivorous birds 0.001 - 0.000 Mollusc eating birds 0.001 - 0.000 Waterfowl 0.001 - 0.000 Striped Bass 0 0.015 b 0.931 Striped Bass 1-2 0.187 b 0.484 Striped Bass 3+ 0.6 Petersen MR estimate 0.009 Largemouth Bass 0 0.097 USFWS beach seine 0.564 Largemouth Bass 1+ 0.2 b 0.657 Sturgeon 0.001 DFG data 0.000 Chinook 0 0.028 a 0.950 Catfish 0.09 Schaffter (1997) 0.635 Other cenntrarchids 2.3 USFWS beach seine 0.349 Delta Smelt 0.33 FMT 0.001 Longfin Smelt 0 0.089 FMT 0.013 Longfin Smelt 1+ 1.5 FMT 0.000 Tule Perch 1.08 USFWS beach seine 0.973 Starry Flounder 0.009 DFG bay study 0.953 Pikeminnow 1 0.227 DFG bay study 0.180 American Shad 0.668 FMT 0.196 Splittail 0 1.59 FMT 0.114 Threadfin Shad 1.2 FMT 0.596 Gobies/Sculpins 3.2 DFG bay study 0.332 Silversides 2.3 USFWS beach seine 0.377 Jellyfish 0.001 Forced 0.000 Crangon f. 0.702 Benthic survey 0.051 other Shrimp 4.441 a 0.950 Corbicula clams 2 Benthic survey 0.449 Corbula clams 0.001 Forced 0.009 Mysids 2 DFG zooplankton survey 0.845 Amphipods 4.558 a 0.950 Other Epi/infauna 5.725 a 0.950 Cladocerans 3.762 a 0.980 Calanoids 5.759 a 0.980 Cycloploid 3.719 a 0.980 Harpticoids 3.258 a 0.980 Limnoithona 0.001 Forced 0.232 SAV 1 Forced 0.136 Phytoplankton 3.577 a 0.990 Other Micro-algae 7.458 a 0.950 Detritus (DOC-POC) 700 a 0.714 a) Parameter calculated by model b) Multi-stanza group, model calculates parameter based on single input

Note: Values in bold were calculated by the model. EE = Ecotrophic Efficiency. The biomass column is followed by a source column, defining where the value came from. Parameters taken from the literature are cited, while parameters that were calculated give the primary data source, with a brief description of the methodology provided in the text.

16

TABLE 2. ECOPATH PARAMETERS FOR THE PRELIMINARY 1982 DELTA MODEL. P/B Parameter Q/B Parameter Functional Group (year-1) Source (year-1) Source Piscivorous birds 0.17 Preikshot (2007) 120 Preikshot (2007) Mollusc eating birds 0.20 Preikshot (2007) 160 Preikshot (2007) Waterfowl 0.20 Preikshot (2007); Anderson (1975) 200 Preikshot (2007) Froese and Pauly Striped Bass 0 - Stevens et al., (1985) 13.25 (2000) Froese and Pauly Striped Bass 1-2 - Stevens et al., (1985) 4.38 (2000) Froese and Pauly Striped Bass 3+ - Stevens et al., (1985) 2 (2000) Froese and Pauly Largemouth Bass 0 - Randall and Minns (2000) 14.57 (2000) Froese and Pauly Largemouth Bass 1+ - - 4.25 (2000) Boreman (1997);Froese and Sturgeon 0.3 DFG data; Beamesderfer (2007) 1.8 Pauly (2000) Cech and Myrick Chinook 0 6.5 Randall and Minns (2000) 30 (1999) Froese and Pauly Catfish 0.28 Schaffter and Kohlhorst (1997) 2.2 (2000) Froese and Pauly Other centrarchids 0.33 Randall and Minns (2000) 4.7 (2000) Froese and Pauly Delta Smelt 2.3 a 7.3 (2000) Froese and Pauly Longfin Smelt 0 - - 46.55 (2000) Froese and Pauly Longfin Smelt 1+ - - 15 (2000) Froese and Pauly Tule Perch 0.51 Froese and Pauly (2000) 8.9 (2000) Froese and Pauly Starry Flounder 0.5 Field et al., (2006) 4.1 (2000) Froese and Pauly Pikeminnow 1 2 Ruzicka et al., (2007) 3.7 (2000) Froese and Pauly American Shad 0.6 Crecco et al., (1983) 4.4 (2000) Froese and Pauly Splittail 0 2.5 Moyle et al., (2004) 25.1 (2000) Froese and Pauly Threadfin Shad 1.2 McLean et al., (1985) 19.5 (2000) Froese and Pauly Gobies/Sculpins 1.5 Randall and Minns (2000). 15.1 (2000) Froese and Pauly Silversides 1.7 Froese and Pauly (2000) 22 (2000) Jellyfish 1.5 Field et al., (2006) 15 a Field et al., (2006); Ruzicka et al., Crangon f. 2 (2007) 12 Field et al., (2006)

17

TABLE 2. (CONTINUED) Functional Group P/B Parameter Q/B Parameter (year-1) Source (year-1) Source Field et al., (2006); Ruzicka et al., other Shrimp 2 (2007) 12 Field et al., (2006) Corbicula clams 2 Christensen et al., (2004) 10 a Corbula clams 2 Jorgensen et al., (2000) 10 a Ruzicka et al., Mysids 7 Ruzicka et al., (2007) 28 (2007) Ruzicka et al., Amphipods 7 Ruzicka et al., (2007) 28 (2007) Other Epi/infauna 1 Jorgensen et al., (2000) 5 a Cladocerans 20 b 100 b Calanoids 20 b 100 b Cycloploid 20 b 100 b Harpticoids 20 b 100 b Limnoithona 20 b 100 b SAV 100 b 0 - Phytoplankton 160 b 0 - Other Micro-algae 80 b 0 - Detritus (DOC-POC) - - - -

Note: Values in bold or denoted as (a) were calculated by the model. P/B = production rate, Q/B = consumption rate. B = average literature value. Each parameter is followed by a source column, defining where the value came from. Parameters taken from the literature are cited, while parameters that were calculated give the primary data source, with a brief description of the methodology provided in the text.

abundance and distribution of salmon fry that use the Delta as rearing grounds. DFG’s

San Francisco Bay Study (Bay Study) was established to investigate the effects of freshwater outflow on the abundance and distribution of fish and mobile crustaceans in the San Francisco Estuary. Sampling includes two different trawls; 1) the otter trawl, which samples demersal fishes, shrimp, and crabs, and 2) the midwater trawl, which samples pelagic fishes (Orsi, 1999). The Bay Study captures many of the same fish as the

FMT so only data from the Bay Study’s otter trawl was used to fill gaps for demersal fish. The benthic and zooplankton surveys sample benthic and pelagic habitats respectively.

18

Parameter Estimate Methodology

Multi-Stanza Groups

Multi-stanza functional groups were created for striped bass (age 0, 1-2, and

3+), largemouth bass (age 0 and 1+), and longfin smelt (age 0 and 1+). Establishing multi-stanza functional groups is a way to examine the roles and processes of multiple life stages within a single functional group. The multi-stanza life stage divisions mirror behavioral changes exhibited by the species in the Delta and were based on discussions with local experts. For multi-stanza groups the basic input parameters were slightly different from standard Ecopath functional groups. The trophic ontogeny of multi-stanza species was modeled explicitly with each stage containing individuals with similar mortality rates and diet compositions. Biomass and Q/B values for one leading stanza

(often one for which assessment data is available) were entered and the biomass and Q/B are calculated for the other stanzas by Ecopath, which assumes that body growth follows a von Bertalanffy growth function (VBGF) curve, and that the species population initially has stable mortality and relative recruitment resulting in a stable age-size distribution

(Christensen et al., 2005). The VBGF predicts the length of a fish as a function of its age,

(-kt) L = L(t): L(t) = L∞ - (L∞ - L0) . Thus, in order to allow Ecopath to calculate unknown

biomass and Q/B values, the user enters values for the von Bertalanffy (1938) curvature

parameter (K), a recruitment power value (between 0 and 2), a biomass accumulation rate

(BA), a value for weight at maturity divided by asymptotic weight (Wmat/W inf), and a

start age for each stanza of that species.

In most cases K was estimated in FishBase where K equals the rate per year at which the asymptotic length is approached = (dL/dt)/( L∞-L)) (Froese and Pauly 2000).

19

The recruitment power value sets the degree of density dependence in juvenile survival for juveniles outside the modeled area (Christensen et al., 2005). In all cases, the recruitment power value was set to 1 to simulate juvenile presence outside the Delta.

Biomass accumulation is a flow term (t · km-2 year-1) where the default value is zero

which indicates no biomass accumulation. A negative BA value implies a reduction in

biomass during the time period modeled (e.g. one year). The BA was set to 0 for each of

the three multi-stanza groups on the Delta Ecopath model.

DFG Fall Midwater Trawl Survey

DFG’s fall midwater trawl survey has been conducted since 1967, between

September and December at stations throughout the Delta. The net used in the survey has

a mouth opening of 3.7 m2. The mesh size decreases from 10.2 cm mesh in the forward

panel to 1.3 cm mesh at the cod end.

Biomass estimates using abundance data from FMT surveys were calculated

using only data collected from stations located within the Delta. Fish abundance data

were then transformed into biomass (t · km-2) using local length-to-weight ratios from

Kimmerer et al., (2005) and Nobriga et al., (2006). Length-to-weight ratios were obtained

from FishBase (Froese and Pauly 2000), when local information was not available. The final biomass estimate was extrapolated to the whole Delta area using a scaling factor that transformed the amount of biomass collected in the area sampled by the FMT net to an estimated biomass for the entire Delta.

USFWS Beach Seine Survey

Biomass estimates using abundance data from USFWS beach seine surveys

were calculated using data collected from stations located within the Delta. Fish

20

abundance data was then transformed into biomass (t · km-2) using local length-to-weight ratios from Kimmerer et al., (2005) and Nobriga et al., (2006). Length-to-weight ratios were obtained from FishBase (Froese and Pauly 2000), when local information was not available. The final biomass estimate was extrapolated to the whole Delta area using a scaling factor that transformed the amount of biomass in the area sampled by the Beach

Seine Surveys to an estimated biomass for the entire Delta.

DFG Bay Study

Biomass estimates using abundance data from DFG Bay Study surveys were

calculated using only data collected from stations located within the Delta. Fish

abundance data were then transformed into biomass (t · km-2) using local length-to-

weight ratios from Kimmerer et al., (2005) and Nobriga et al., (2006). Length-to-weight

ratios were obtained from FishBase (Froese and Pauly 2000), when local information was

not available. The final biomass estimate was extrapolated to the whole Delta area using

a scaling factor that transformed the amount of biomass in the area sampled by the Bay

Study to an estimated biomass for the entire Delta.

Peterson Mark-Recapture Estimate

Biomass estimates for age 3+ striped bass were calculated using data based on

DFG’s Petersen mark-recapture estimates of population size and striped bass length data.

Abundance data were transformed into biomass data using the striped bass length-to-

weight ratio in Kimmerer et al., (2005). The resulting biomass calculation was then

scaled back to the Delta model’s area because the original abundance data included the

greater San Francisco estuary area.

21

Sturgeon Data

Sturgeon biomass was estimated from DFG tagging studies, fishing report

cards, abundance estimates, and survival data. White sturgeon contributed the most to

overall biomass. White sturgeon biomass was estimated from annual estimated

abundance indices using white sturgeon mark-recapture data. Since there was no

abundance estimate for 1982, the number was extrapolated from pre and post 1982 data

points using regression. Average length and weight data was used to estimate average

biomass within the Delta.

Benthic Survey

Benthic data from the Department of Water Resources’ Environmental

Monitoring Program (under the IEP umbrella), benthic monitoring program (Benthic

Survey) were used to obtain historical trends for Corbicula fluminea (Corbicula), and

California bay shrimp (Crangon franciscorum). The survey has been conducted from

1975 to the present. The study area currently contains ten sites that are sampled monthly using a hydraulic winch and Ponar dredge. The Ponar dredge samples a bottom area of approximately 0.052 m2 to a depth that varies with the type of sediment and the ability of

the dredge to go through it (Fields and Messer, 1999).

DFG Zooplankton Survey

The DFG Zooplankton and Neomysis survey has been conducted since 1971,

with more than 80 stations located throughout the Delta and greater San Francisco

Estuary. The survey monitors the density of Neomysis mercedes and other mysid shrimp

(summed as mysids) and zooplankton in the Delta. Surveys are conducted twice per

22

month from March to November. Biomass estimates using data from this survey were

calculated by W. Kimmereer, SFSU (Personal communication).

Q/B estimate via FishBase

For most of the fish groups, consumption (Q/B) values were estimated by the

empirical equation available in FishBase (Froese and Pauly 2000) based on work by

Palomares and Pauly (1998) using the following equation:

log Q/B = 7.964–0.204 logW∞–1.965T’+0.083A+0.532h+0.398d (3)

Where estimates are provided for weight at infinite age (W∞), average environmental

temperature (T), the ratio of the square of the height of the caudal fin to its surface area

(A) which is a measure of the swimming and metabolic activity of a fish (a higher ratio

implies more swimming and greater activity), and food type (detritivore, herbivore,

omnivore, carnivore). Where (h) and (d) are variables indicating herbivores (h=1, d=0),

detritivores (h=0, d=1) and carnivores (h=0, d=0).

Forced Biomass

Corbula amurensis (Corbula) and Limnoithona tetraspina (Limnoithona),

invaded the Delta after 1982. Native species of submerged aquatic vegetation (SAV) were present in low abundance before the 1980s but an invasive species of SAV, Egeria

densa, became abundant between the 1980s and 2000s. Jellyfish were present in the

system in 1982, but in extremely low abundances. Thus, in 1982 there was no known

biomass for these functional groups, so a place holder of 0.0001 t · km-2 was designated

so that the population can be “introduced” or forced into the system at a later date, when

using Ecosim to simulate the introductions or increases in abundance and resulting

changes in the ecosystem.

23

Diet Composition Matrix

Diet composition data for each functional group is presented in the following diet matrixes for the 1982 Ecopath model (Tables 3-5). Diet composition sources are presented in Table 6. Local diet data were used whenever possible and is denoted by an asterisk. Diet composition information is presented in proportion (0-1) of prey item, to represent diet composition linkages.

Diet information was taken from literature sources as close to 1982 as possible to simulate the diet linkages among functional groups during that time. When this was not possible, diet information was taken from later studies and if feasible, modified to reflect known diet composition in the early 1980s. Diet information for longfin smelt,

American shad, splittail, threadfin shad, and gobies/sculpins from Feyrer et al. (2003), is actually from data collected between the years 1979-1983, hence representing 1982 diet composition fairly well. The functional group SAV refers to the invertebrates and other lower trophic level food items that live within the SAV. Presence of SAV in a functional group’s diet is reflecting this concept and does not necessarily mean the predator is

preying on, or receiving nutritional value from the SAV itself. In the diet composition

matrix (Tables 3-5) the term ‘import’ refers to the consumption of prey that is not part of

the modeled system. For the Delta model, the bird functional groups get the majority of

their food from outside the system, hence their high import numbers. Diet information

from more than one source were averaged to reflect the breadth of information when

from the same time period.

24

TABLE 3: DIET MATRIX FOR THE PRELIMINARY 1982 DELTA ECOPATH MODEL—PART 1. Piscivorous Mollusc Striped Striped Striped Largemouth Largemouth Other Prey - Predator birds eating birds Waterfowl Bass 0 Bass 1-2 Bass 3 Bass 0 Bass 1+ Sturgeon Chinook 0 Catfish cenntrarchids Piscivorous birds Mollusc eating birds Waterfowl Striped Bass 0 0.006 0.007 0.009 0.001 Striped Bass 1-2 0.005 0.000 0.003 0.060 Striped Bass 3 0.005 Largemouth Bass 0 0.009 0.002 0.042 0.021 Largemouth Bass 1+ 0.005 0.022 Sturgeon Chinook 0 0.009 0.002 0.012 0.015 Catfish Other centrarchids 0.009 0.002 0.030 0.042 0.016 Delta Smelt 0.006 Longfin Smelt 0 0.009 Longfin Smelt 1+ Tule Perch 0.002 0.007 0.036 0.044 Starry Flounder 0.002 0.002 Pikeminnow 1 0.003 0.007 American Shad 0.041 0.002 0.030 Splittail 0 0.009 0.019 0.012 0.015 Threadfin Shad 0.009 0.415 0.258 0.012 0.015 Gobies/Sculpins 0.009 0.192 0.020 0.165 0.116 Silversides 0.009 0.124 0.065 0.060 0.096 0.102 Jellyfish Crangon f. 0.060 other Shrimp 0.046 0.054 0.255 0.030 0.090 0.110 Corbicula clams 0.250 0.400 Corbula clams 0.000 0.000 Mysids 0.184 0.134 0.024 0.100 0.100 0.029 Amphipods 0.639 0.114 0.000 0.410 0.036 0.050 0.044 Other Epi/infauna 0.250 0.150 0.026 0.025 0.182 0.060 0.054 0.500 0.550 0.500 0.066 Cladocerans 0.010 0.060 0.072 0.100 0.088 Calanoids 0.033 0.100 Cycloploid 0.037 0.100 Harpticoids 0.007 0.100 Limnoithona 0.000 0.000 SAV 0.350 0.018 0.320 0.241 0.293 Phytoplankton Other Micro-algae Detritus (DOC-POC) 0.400 Import 0.900 0.500 0.500 Note. Diet composition units are in proportion (0-1) of total diet.

25

TABLE 4: DIET MATRIX FOR THE 1982 DELTA ECOPATH MODEL—PART 2. Delta Longfin Longfin Tule Starry American Threadfin Prey - Predator Smelt Smelt 0 Smelt 1+ Perch Flounder Pikeminnow 1 Shad Splittail 0 Shad Gobies/Sculpins Silversides Jellyfish Piscivorous birds Mollusc eating birds Waterfowl Striped Bass 0 0.021 0.001 Striped Bass 1-2 0.025 0.003 Striped Bass 3 0.003 Largemouth Bass 0 0.082 Largemouth Bass 1+ 0.003 Sturgeon Chinook 0 0.082 Catfish Other centrarchids Delta Smelt Longfin Smelt 0 0.082 Longfin Smelt 1+ Tule Perch 0.025 Starry Flounder Pikeminnow 1 American Shad Splittail 0 0.082 Threadfin Shad Gobies/Sculpins 0.025 Silversides 0.025 Jellyfish Crangon f. other Shrimp 0.011 0.169 0.082 Corbicula clams 0.027 0.960 0.056 Corbula clams 0.000 0.000 0.000 Mysids 0.015 0.269 0.233 0.309 0.085 0.021 0.082 Amphipods 0.045 0.028 0.030 0.929 0.528 0.083 0.152 0.056 0.156 0.082 Other Epi/infauna 0.030 0.298 0.083 0.085 0.082 Cladocerans 0.099 0.003 0.083 0.126 0.250 0.189 0.082 Calanoids 0.626 0.618 0.648 0.003 0.220 0.001 0.615 0.250 0.385 0.082 Cycloploid 0.137 0.042 0.045 0.003 0.220 0.001 0.104 0.250 0.170 0.082 Harpticoids 0.051 0.042 0.045 0.003 0.001 0.085 0.022 0.125 0.100 0.082 Limnoithona 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 SAV 0.008 0.125 Phytoplankton 0.008 Other Micro-algae 0.008 Detritus (DOC-POC) 0.010 0.008 0.423 Note. Diet composition units are in proportion (0-1) of total diet.

26

TABLE 5: DIET MATRIX FOR THE 1982 DELTA ECOPATH MODEL PART 3. Crangon other Corbicula Corbula Other Prey - Predator f. Shrimp clams clams Mysids Amphipods Epi/infauna Cladocerans Calanoids Cycloploid Harpticoids Limnoithona Piscivorous birds Mollusc eating birds Waterfowl Striped Bass 0 Striped Bass 1-2 Striped Bass 3 Largemouth Bass 0 Largemouth Bass 1+ Sturgeon Chinook 0 Catfish Other centrarchids Delta Smelt Longfin Smelt 0 Longfin Smelt 1+ Tule Perch Starry Flounder Pikeminnow 1 American Shad Splittail 0 Threadfin Shad Gobies/Sculpins Silversides Jellyfish Crangon f. other Shrimp Corbicula clams Corbula clams Mysids Amphipods 0.111 Other Epi/infauna Cladocerans 0.105 0.105 0.105 0.250 0.111 Calanoids 0.105 0.105 0.105 0.250 0.111 Cycloploid 0.105 0.105 0.105 0.250 0.111 Harpticoids 0.105 0.105 0.105 0.250 0.111 Limnoithona 0.000 0.000 0.000 0.000 0.000 SAV 0.111 Phytoplankton 0.105 0.105 0.330 0.330 0.105 0.111 0.330 0.330 0.330 0.330 0.330 Other Micro-algae 0.105 0.105 0.340 0.340 0.105 0.111 0.330 0.330 0.330 0.330 0.330 Detritus (DOC-POC) 0.368 0.368 0.330 0.330 0.368 0.111 0.340 0.340 0.340 0.340 0.340 Note. Diet composition units are in proportion (0-1) of total diet.

27

TABLE 6: DIET COMPOSITION SOURCES FOR EACH FUNCTIONAL GROUP. Diet Composition Functional Group Source Piscivorous birds Kushlan (1976) Mollusc eating birds Poulton et al., (2002)* Waterfowl Euliss (1989) Striped Bass 0 Steve Slater, unpublished data* Striped Bass 1-2 Nobriga and Feyrer (2008)* Striped Bass 3+ Nobriga and Feyrer (2008)* Largemouth Bass 0 Nobriga and Feyrer (2007)* Largemouth Bass 1+ Nobriga and Feyrer (2007)* Sturgeon McKechnie and Fenner (1971)*; Kogut (2008)* Chinook 0 MacFarlane and Norton (2002)* Catfish Turner (1966)* Other centrarchids Turner (1966)*; Froese and Pauly (2000) Delta Smelt Lott (1998)*; Moyle et al., (1992)*; Nobriga (2002)* Longfin Smelt 0 Feyrer et al., (2003)* Longfin Smelt 1+ Feyrer et al., (2003)* Tule Perch Turner (1966)*; Moyle (2002)*; Feyrer et al., (2003)* Starry Flounder McCall (1992); Moyle (2002)*; Feyrer et al., (2003)* Pikeminnow 1 Nobriga et al., (2006)* American Shad Feyrer et al., (2003)* Splittail 0 Feyrer et al., (2003)*; Moyle et al., (2004)* Threadfin Shad Feyrer et al., (2003)* Gobies/Sculpins Feyrer et al., (2003)* Silversides Moyle (2002)* Jellyfish Purcell (2003) Crangon f. Wahle (1985)* other Shrimp Wahle (1985)* Corbicula clams Lauritsen (1986)* Corbula clams Carlton et al., (1990)* Mysids Mauchline (1980); Orsi and Knutson (19979)* Amphipods/epi/infauna Oakden (1984) Rollwagen Bollens and Penry (2003)*; Hooff and Bollens Zooplankton species (2004)*; Bouley and Kimmerer (2006)*

Note. Local diet information was used whenever possible and is denoted by an asterisk (*).

28

Balancing the Ecopath Model

Once the basic input parameters are entered into the Ecopath model, the model is balanced using a series of steps recommended by the Ecopath software engineers including reducing cannibalism within functional groups, examining diet discrepancies among unbalanced groups, and making biomass adjustments when ecologically reasonable (Christensen et al., 2005).

One of the primary concerns of balancing an Ecopath model is the EE parameter, which should be between 0 and 1. An EE value near or equal to 1 means that the functional group is being heavily preyed on, or that fishing pressure/entrainment is high and few to no individuals are left to die of old age. An EE value of zero means that no other functional group preys upon that particular group, nor is it exploited in any way

(e.g. over fishing, commercial harvesting, entrainment).

CHAPTER III

RESULTS

The initial input data for the four zooplankton species was too low to allow the model to balance. Even with diet composition variations and other recommend

balancing steps from Ecopath experts, the software could not balance the model with local data inputs. The next step was to let the model estimate zooplankton biomass based on an EE of 0.95. The model balanced after estimating a zooplankton increase of between

one and three orders of magnitude in comparison to local data. This means that Delta zooplankton biomass is underestimated, or the linkages in the model are incorrect. The

San Francisco Estuary (including the Delta) is recognized to have much lower primary production than other similar systems (Day et al., 1989, Jassby, 2008) and this may be playing a factor in this discrepancy.

Values of P/B ranged from 160 year-1 for phytoplankton to 0.18 year-1 for piscivorous birds (Table 2). The low value for piscivorous birds makes ecological sense because they are top predators in the food web and have fewer sources of mortality compared to common prey species, such as shrimp and zooplankton, which are consistent food items for TL 3. Sturgeon also have a relatively low P/B value (Table 2), most likely because this functional group is not split up into multi-stanzas and mortality is averaged over the group’s life history. Adult sturgeons are long-lived, have few predators, with relatively low fishing mortality (Table 2). The high values for zooplankton and algae

29 30 result from their rapid turnover with multiple generations during the annual time step and high exploitation of these groups by predators.

Q/B values range from 200 year-1 for waterfowl to 1.8 year-1 for sturgeon. The birds in the system are once again at the high end of the range. The high Q/B values for the birds, makes ecological sense because they are homoeothermic and have higher basal metabolic rates than the remaining species in the model, which are all ectothermic.

Waterfowl also have to store energy for major migrations. Hence, consumption of many times their body weight in food in a year seems reasonable. Zooplankton also have a high

Q/B value, (100 year-1). The high value implies that zooplankton must consume large quantities of food to support the rapid turnover and high predation rates on the population. The Q/B value for age 0 longfin smelt (46.55 year-1) seems high and needs to be reviewed. Since it is part of a multi-stanza group, the value was actually calculated by the model based on the input for age 1+ longfin smelt (Table 2). It is likely that characteristics specific to longfin smelt resulted in the equation giving a high value compared to other species.

Diet composition data for the Delta functional groups show that planktivore fish species such as delta smelt, longfin smelt and American shad are highly dependent on calanoid copepods (Table 3-5). Mysids and amphipods are important for some juvenile fish including striped bass, longfin smelt and American shad. The piscivores, including birds, largemouth bass, and other centrarchids are consuming a wide variety of species ranging from trophic level 1-3. The occurrence of SAV is highest in waterfowl, largemouth bass (all age groups) and other centrarchids. This implies that these species are getting a significant portion of their diet from the invertebrates and organisms that

31

live within the SAV. The presence of invasive species that were not in the Delta in 1982

including Corbula and Limnoithona are still present in the diet of various functional

groups because they need to maintain a placeholder until they are brought into the model

with Ecosim at a later date.

Diet information for the Delta Ecopath model was taken primarily from the

published literature (Table 6). In some cases, diet information was taken from recent

studies and may not accurately reflect diet linkages in the 1982 Delta. Some functional

groups had very little diet information, or the information was not from local sources

(e.g. bird groups). Also, diet information for functional groups that were grouped together, such as centrarchids, was averaged over all diet items, potentially decreasing its accuracy. Other diet linkages are not included in the model because of insufficient data.

For example, it has been hypothesized that ciliates provide an important link between algae, DOC/POC, and zooplankton, however, there is insufficient information to include it in the model. Regardless of these issues, much of the Delta Ecopath diet information is based on sound, local data and adds great strength to the model. The Delta model can be updated as new or improved information becomes available.

Ecopath calculates a variety of ecological characteristics, including the

Production: (consumption ratio (P/Q), flow to detritus, net efficiency and the omnivory index. The P/Q ratio is sometimes referred to as the gross efficiency rate of a functional group. P/Q values normally range from 0.05 to 0.3, meaning that the consumption of most groups is about 3-20 times higher than their production. Exceptions to this rule included top predators in the model, which can have lower P/Q values, and small fast-

growing fish, larvae or other organisms, whose P/Q values can be much higher

32

(Christensen et al., 2008). In the Delta model mysids, clams, shrimp, gobies/sculpin, and age 1+ largemouth bass have the highest P/Q ratios. The birds in the system and also the carnivores ‘other centrarchids’ have the lowest P/Q ratios.

Flow to detritus consists of the non-assimilated food and portions of the functional group that die of old age, diseases, or other sources of mortality. The flow to detritus should be positive for all groups. Delta Ecopath results show that all flow to detritus is positive, with the greatest contribution coming from trophic levels 1 and 2

(Table 7). Net efficiency is calculated as production divided by the assimilated portion of the food, based on the equation Net efficiency = Production / (production + respiration).

The net efficiency is a dimensionless fraction, always positive and almost always less than 1. In this system, net efficiency is highest for the clams (0.714), followed by gobies/sculpins (0.568) and largemouth bass (age 0 at 0.429 and age 1+ at 0.529). Lowest net efficiency values come from the bird groups (0.001-0.002).

The omnivory index is a statistical measure of the number of different trophic levels a particular consumer exploits. For example, if the omnivory index of a functional group is zero, than that consumer is feeding on a single trophic level. A large value indicates that the consumer feeds on many trophic levels. The omnivory index is dimensionless (Christensen et al., 2008). Largemouth bass, other centrarchids, catfish and birds feed on multiple trophic levels in the Delta model, thus having a high (0.639-.0908) omnivory index. A low omnivory index value (e.g. 0) is for groups feeding at one tropic level such as clams, amphipods, and zooplankton.

Trophic Level (TL) is a dimensionless number that is calculated by the model based on the position of the functional group in the food web relative to primary

33

producers or detritus. The TL is calculated as 1 + the weighted average of the trophic

levels of all prey items. The highest TL in this model was found for the piscivorous birds

(TL 4.2), followed by striped bass (TL 4), with the majority of the other fish species falling into TL 3. Phytoplankton, SAV, micro-algae and detritus make up TL 1, with

zooplankton and other primary consumers filling out TL 2.

The trophic flows and relative biomass amounts in each trophic group can also

be displayed graphically (Figure 3). The zooplankton groups are largely supported by

phytoplankton production, which appears relatively small in comparison to other TL 1

production. Of the total system biomass (270.5 t km-2) 37% is from TL 3, which consists

primarily of the fish species (Table 8). TL 1, including both detritus and phytoplankton,

composes 55% of total system biomass. The final 8% of total system biomass is composed of the second and fourth trophic levels. Ecopath calculates transfer efficiencies as the fraction of total flows at each trophic level that are either exported or transferred to another trophic level through the process of consumption. Transfer efficiencies for TL 1-

4 range from 17.6-3.7% (Table 9).

34

Figure 3. A simplified food-web model of the Delta in 1982. Estimated trophic level is on the y-axis, the size of the boxes is scaled to the standing biomass and the width of the lines represents biomass flow of prey to predators. The pelagic system groups are largely supported by phytoplankton production, while the benthic system groups are supported by detritus.

35

TABLE 7. ECOLOGICAL CHARACTERISTICS FOR EACH FUNCTIONAL GROUP. Flow to Detritus t Net Omnivory Trophic Functional Group P/Q km-2 year-1 Efficiency Index Level Piscivorous birds 0.001 0.024 0.002 0.838 4.2 Mollusc eating birds 0.001 0.024 0.002 0.736 3.3 Waterfowl 0.001 0.04 0.001 0.573 2.5 Striped Bass 0 0.189 0.151 0.236 0.161 3.7 Striped Bass 1-2 0.194 0.987 0.242 0.073 4 Striped Bass 3 0.325 2.072 0.406 0.109 3.7 Largemouth Bass 0 0.343 0.47 0.429 0.87 3.3 Largemouth Bass 1+ 0.424 0.43 0.529 0.908 3.5 Sturgeon 0.167 0.007 0.208 0.099 3.4 Chinook 0 0.09 0.227 0.113 0.35 2.5 Catfish 0.127 0.049 0.159 0.639 3 Other centrarchids 0.07 1.05 0.088 0.873 3.3 Delta Smelt 0.315 0.112 0.394 0.045 3.1 Longfin Smelt 0 0.043 0.121 0.054 0.057 3.1 Longfin Smelt 1+ 0.063 0.79 0.079 0.057 3.1 Tule Perch 0.109 1.72 0.137 0.042 3.9 Starry Flounder 0.109 0.114 0.136 0.023 3 Pikeminnow 1 0.027 1.812 0.034 0.287 3.9 American Shad 0.25 8.165 0.313 0.046 3.3 Splittail 0 0.243 4.516 0.304 0.056 2.9 Threadfin Shad 0.136 16.933 0.17 0.056 3.1 Gobies/Sculpins 0.455 5.463 0.568 0.109 2.9 Silversides 0.085 4.894 0.106 0.131 3.2 Jellyfish 0.1 0.045 0.125 0.245 3.5 Crangon f. 0.4 147 .189 0.5 0.244 2.4 other Shrimp 0.4 27.554 0.5 0.244 2.4 Corbicula clams 0.571 1.519 0.714 0 2 Corbula clams 0.571 0.003 0.714 0 2 Mysids 0.4 117.9 0.5 0.244 2.4 Amphipods 0.14 38.546 0.175 0 3 Other Epi/infauna 0.1 7.498 0.125 0.444 2.7 Cladocerans 0.2 134.34 0.25 0 2 Calanoids 0.2 178.593 0.25 0 2 Cycloploid 0.2 132.519 0.25 0 2 Harpticoids 0.2 123.546 0.25 0 2 Limnoithona 0.2 0 0.25 0 2 SAV - 141.804 - 0 1 Phytoplankton - 0.957 - 0 1 Other Micro-algae - 0.957 - 0 1 Detritus (DOC-POC) - 0 - 0.377 1 Note. P/Q = Production/Consumption. Trophic level is calculated by the model.

36

TABLE 8. SUMMARY STATISTICS OF THE DELTA MODEL. Parameter Value Units Sum of all consumption 3621.52 t · km-2 year-1 Sum of all exports 280.78 t · km-2 year-1 Sum of all respiratory flows 2083.71 t · km-2 year-1 Sum of all flows into detritus 1403.14 t · km-2 year-1 Total system throughput 7389.16 t · km-2 year-1 Sum of all production 2877.73 t · km-2 year-1 Calculated total net primary production 2064.23 t · km-2 year-1 Total primary production/total biomass 7.62 Total biomass/total throughput 0.04 Total biomass (excluding detritus) 270.80 t · km-2

TABLE 9. TROPHIC TRANSFER EFFICIENCIES FOR EACH TROPHIC GROUP (%).

Trophic Level Transfer Efficiencies 2 17.6

3 7.8

4 3.7 Transfer efficiencies From primary producers: 8.1%; From detritus: 7.8%; Total: 8.0% Note. Calculated as a proportion of energy transferred from one trophic level to the next.

CHAPTER IV

DISCUSSION

The balanced Delta Ecopath model gave some interesting results about the food-web within the Delta during 1982. It called attention to the fact that the model is missing a potentially critical link that involves describing and quantifying the microbial loop and its contribution to the Delta food web through microzooplankton and zooplankton. The microbial loop refers to the bacterial driven process that transforms particulate organic carbon (POC) and dissolved organic carbon (DOC) into bioavailable carbon. Research aimed at describing and understanding the microbial loop in the Delta is minimal, but the available data indicate the importance of the microbial loop in the upper

San Francisco Estuary and Delta systems as well as the importance of the detrital food supplies (Hollibaugh and Wong, 1999; Sobczak et al., 2002). Without better representation of microbial and detrital food web inputs, the Delta model may not represent Delta food web accurately, because of the missing critical lower trophic level food availability and linkages.

Trophic transfer efficiencies generally vary around 10%, so that one-tenth of the energy that enters a trophic level is transferred to the next trophic level (Lindeman

1942). Transfer efficiencies are usually greater at the beginning of the food web compared with higher trophic levels, because of varying characteristics of organisms at different levels in the food web (Christensen and Pauly, 1993). Transfer efficiencies for

37 38 the Delta average 8.3%, with the highest trophic transfer efficiency at TL 1 (17.6%) and the lowest at TL 4 (3.7%).These results are consistent with the general trend of decreasing transfer efficiency with increasing TL. Average transfer efficiency for the

Delta model falls within the range of 8-15% described by Christensen and Pauly (1993) and Wolff (1994). Using the trophic efficiencies calculated for the Delta Ecopath model

(Table 7), there is clearly a discrepancy in trophic efficiency when accounting for biomass production at TL 3. This discrepancy may be largely attributed to the lack of information on immigration and emigration in the current Ecopath model. It is also important to note that the model is not accounting for primary production within the

SAV. The production used by the organisms is likely the growth of epiphytic diatoms on the surface of the SAV, not the SAV itself. So, the current model is missing a large source of production.

The uncertainty in the zooplankton biomass estimates brings attention to the relative lack of biomass in TL 2 (approximately 6% of total system biomass), compared with 37% from TL 3. The traditional view on food webs is that biomass decreases up the food web’s trophic levels. Primary producers (TL1) contain the greatest amount of biomass, followed by the primary consumers (TL2), secondary consumers (TL3), and finally the tertiary consumers, or top predators (TL4). A potential contributing factor to the low amount of biomass in TL 2 is that zooplankton populations are food limited in the

Delta. Kimmerer et al., (2005) and Sobczak et al., (2005) show that zooplankton primarily feed on phytoplankton and not detritus within the Delta and that this factor is further compounded by the low phytoplankton productivity within the Delta. Therefore, if zooplankton have a limited amount of food, then they too will be limited in their

39

productivity, generating a relatively small TL 2 biomass. A major contributor to the high

amount of biomass in TL 3, and the unusually high trophic transfer efficiencies (Table 9)

is that the model does not account for the fact that some of the species groups are

growing (i.e. attaining biomass) largely outside of the domain of the model. For example,

the FMT catches longfin smelt as they are moving back into the Delta to spawn, hence

their presence in the model (contributing biomass to TL 3) may be overestimated because

their biomass is based on FMT data. In actuality, a very high percentage of longfin smelt

rear outside of the Delta in the San Francisco Bay and near-shore ocean so most of the

biomass is being created outside of the model. This is also the case for striped bass and

sturgeon because individuals move in and out of the domain of the model. Similarly,

Chinook salmon mainly move through the Delta rather than rear there.

The high TL 3/low TL 2 biomass result may also be due to the fact that the

Delta model is actually describing parts of two different food webs, a pelagic food web, and a littoral food web. The littoral food web (modeled by the presence of SAV) includes the array of food items that are contained within the SAV and not explicitly defined in the model. The SAV food web is likely almost completely supporting several of the large fish functional groups including largemouth bass, some catfish, and other centrarchids.

This fish biomass is not supported by phytoplankton or detritus so it will distort the model outputs. A future version of the Delta model will attempt to incorporate the SAV food web, including the invertebrates that live and feed within the SAV and the fish that feed on the invertebrates and use the SAV as refuge. The current Delta Ecopath model does not account for the high amount of immigration and emigration of fish species in the

40

Delta, nor does it account for the SAV associated food web, both potential causes of the high TL 3 biomass. Clearly, this should be a major focus of future research.

The major objective of the present project was to use Ecopath to systematically describe the Delta ecosystem and to explore its various properties using mass balance principles. The Delta ecosystem is an open system that experiences seasonal variation in environmental conditions (e.g. fresh water inflow, salinity, temperature) and the distribution and abundance of many of the organisms. The Ecopath model assumes a steady-state, as this is the major condition needed to solve the system of equations that drives the model. Thus, these results can only give a preliminary snap-shot of the Delta based on the available data for the Delta in 1982.

Results from the present model are preliminary and illustrate the need for better inputs and development of the model. There is uncertainty in the parameter estimates for many species and trophic pathways in the Delta ecosystem. While modeling can be valuable, it is important to remember that model outputs are only projections, and although based on the best available data, they do not predict the future. Model projections enable researchers and policy makers to study scenarios without having to implement them. Models can also capture elements of a system that can lead to discovery of unknown processes and identifying needs or gaps in current data.

The Delta Ecopath input parameters are based on the best available data at the time of development. Many of the parameters were calculated from long term IEP data sets, which are the only available data from the system. Input parameters were reviewed by local experts whenever possible. Unfortunately, not all input parameters were grounded in real data, some were estimated and others taken from the literature.

41

Improving the data is a continual process that should continue for as long as the model is being used for the Delta. A positive aspect of such an in-depth data mining exercise is that it identifies ecosystem parameters that are poorly known, highlighting for managers the types of data needed to improve understanding and model performance. This in turn could instigate new or improved monitoring and research programs.

The current model is built on an annual time step, meaning that all data was averaged over the year before being used in the model. A potential improvement to the model could be made by incorporating seasonality through a seasonal time step. This could help improve the model’s reproduction of such things as seasonal phytoplankton fluxes, migration in and out of the system, and seasonal changes I species biomass.

Another alternative would be to build four different models, one to represent each season with input parameters based on season specific data. Comparison of production, biomass, consumption, and trophic flows in each season could lead to new understanding of variability in habitat suitability and trophic conditions in the Delta throughout the year.

Future Work

The development of an ecosystem-wide Ecopath model of the Suisun Bay and

Sacramento-San Joaquin Delta is a major step in exploratory ecosystem modeling of this highly altered system. The basic input parameters provide a preliminary foundation for the development of the more in-depth Ecosim model, the time-dynamic simulation model based on the mass-balanced Ecopath model. Using the Delta Ecosim model, different external factors can be applied as forcing functions to drive the trophic interactions such as invasive species, salinity, and primary production. Assessment of the relative

42 improvement in model fit provided by considering additional factors can be use as a method to determine what combinations of factors best explain the POD while still accounting for the relative stability of other organisms.

After successful development of the Delta Ecopath and Ecosim models, development of a spatial component of the EwE model, Ecospace, will allow further exploration of the freshwater, low salinity, pelagic, and littoral habitats within the Delta.

A spatial modeling component of the Delta EwE model will allow for finer resolution hypothesis testing and exploration of the habitat and food web changes in the Delta. A spatial model will allow habitat shifts to emerge from the physical parameters of the model and allow the distribution and abundance of fish to shift as a result. Compared to the coarse forcing function used in Ecosim, spatial modeling will help explore the role of invasive species and their effects on the food web in both the littoral and pelagic habitats of the Delta. By more explicitly capturing habitat dynamics, more realism can be incorporated into the model, making an even more useful tool for researchers, managers and policy makers.

REFERENCES CITED

1

REFERENCES CITED

Allen, K.R., 1971, Relation between production and biomass: Journal of the Fisheries Research Board of Canada, v. 28, p. 1573–1584.

Anderson, D.R., 1975, Population ecology of the mallard, V: Temporal and geographic estimates of survival, recovery, and harvest rates: U.S. Fish and Wildlife. Service Resource Publication, v. 125, p. 110.

Baxter, R., Breuer, R., Brown, L., Chotkowski, M., Feyrer, F., Gingras, M., Herbold, B., Mueller-Solger, A., Nobriga, M., Sommer, T., Souza, K. 2008. Pelagic organism decline progress report: 2007 synthesis of results. Interagency Ecological Program for the San Francisco Estuary, Technical Report 227, Sacramento CA: Department of Water Resources, p. 67.

Beamesderfer, R.C.P., 2007, Use of life history information in a population model for Sacramento green sturgeon: Environmental Biology of Fishes, v. 79. doi: 10.1007/S10641-006-9145-x

Bennett, W.A., 1995, Larval striped bass condition in a drought-stricken estuary: Evaluating pelagic food-web limitation, v. 5, p. 1995-1692. doi:10.2307/1941977

Bergamaschi, B.A., Kuivila, K.M., and Fram, M.S., 2001. Pesticides associated with suspended sediments entering San Francisco Bay following the first major storm of water year 1996: Estuaries, v. 24, p. 368–380.

Bertalanffy, L. von, 1938, A quantitative theory of organic growth (Inquiries on growth laws. II): Human Biology, v. 10, p. 181-213.

Booth, S and Zeller, D., 2005, Mercury, food webs, and marine mammals: implications of diet and climate change for human health: Environmental Health Perspectives, v. 133, p. 521–526. doi: 10.1289/ehp.7603.

Boreman, J., 1997, Sensitivity of North American sturgeons and paddlefish to fishing mortality: Environmental Biology of Fishes, v. 48, p. 399-405. doi: 10.1023/A: 1007345806559

Bouley, P. and Kimmerer, W., 2006, Ecology of a highly abundant, introduced cyclopoid copepod in a temperate estuary: Marine Ecology Progress Series, v. 324, p. 2006- 2228. doi:10.3354/meps324219.

44 45

Brown, L.R. and Moyle, P.B., 2005, Native fish communities of the Sacramento-San Joaquin watershed, California: a history of decline, in Rinne, F., Hughes, R., and Calamusso, R., eds., Fish Communities of Large Rivers of the United States: American Fisheries Society, Bethesda, Maryland, p. 75-98.

Brown, L.R., 2007, Littoral fish assemblages of the alien-dominated Sacramento-San Joaquin Delta, California, 1980-1983 and 2001-2003: Estuaries and Coasts, v. 30, p.186-200. doi: 10.1007/BF02782979.

California Department of Water Resources, 1993, Sacramento-San Joaquin Delta atlas: California Department of Water Resources, 1 p. 21.

Carlton, J.T., Thompson, J.K., Schemel, L.E., and Nichols, F.H. 1990. Remarkable invasion of San Francisco Bay (California, USA) by the Asian clam Potamocorbula amurensis. I. Introduction and dispersal: Marine Ecology Progress Series, v. 66, p. 81-94.

Cech, J.J.J., and Myrick, C.A., 1999, Steelhead and Chinook Salmon bioenergetics: Temperature, ration, and genetic effects: UC Berkeley, University of California Water Resources Center, p. 72.

Chen, Z., Qiu, Y., Jia, X., and Xu, S., 2008, Using an ecosystem modeling approach to explore possible ecosystem impacts of fishing in the Beibu Gulf, Northern South China Sea: Ecosystems, v. 11, p. 1318-1333. doi: 10.1007/s10021-008-9200-x

Christensen, V., and Pauly, D., 1992, Ecopath II - a software for balancing steady-state ecosystem models and calculating network characteristics: Ecological Modeling, v. 61, p. 169-185. doi:10.1016/j.ecolmodel.2003.09.003.

Christensen, V. and Pauly, D., editors, 1993, Trophic Models of Aquatic Ecosystems: Manila, International Center for Living Aquatic Recourses Management, p. 390.

Christensen, V., and Walters, C.J., 2004, Ecopath with Ecosim: methods, capabilities and limitations: Ecological Modeling, v. 172, p. 109-139. doi:10.1016/j.ecolmodel. 2003.09.003.

Christensen, V., Beattie A., Buchanan, C., Martell, S.J.D., Latour, R.J.,Preikshot, D., Sigrist, M., Uphoff, J.H., Walters, C.J., Wood, R.J., and Townsend, H., Fisheries ecosystem model of the Chesapeake Bay: Methodology, parameterization, and model exploration. Draft Technical Report for submission to NOAA Technical Report Series. Current revision: 9 June 2008.

Christensen, V., Walters, C.J., and Pauly, D., 2005, Ecopath with Ecosim: a User’s Guide. Fisheries Centre, University of British Columbia, Vancouver. November 2005 edition, p 154.

46

Clark, J.S., Carpenter, S.R., Barber, M., Collins, S., Dobson, A., Foley, J.A., Lodge, D.M., Pascual, M., Pielke, R. Jr., Pizer, W., Pringle, C., Reid, W.V., Rose, K.A., Sala, O., Schlesinger, W.H., Wall, D.H., and Wear, D., 2001, Ecological forecasts: an emerging imperative: Science, v. 293, p. 657-660. doi: 10.1126/science.293.5530.657

Cohen, A. N., and J. T. Carlton., 1998, Accelerating invasion rate in a highly invaded estuary: Science, v. 279, p. 555-558. 10.1126/science.279.5350.555.

Crecco, V., Savoy, T., and Gunn, L., 1983, Daily mortality rates of larval and juvenile American shad (Alosa sapidissima) in the Connecticut River with changes in year-class strength: Canadian Journal of Fisheries and Aquatic Sciences, v. 40, p. 1719-1728. doi: 10.1002/rra.1111.

Euliss, N.H., Jr., R.L. Jarvis, and D.S. Gilmer, 1991, Feeding ecology of waterfowl wintering on evaporation ponds in California: The Condor, v. 93, p. 582-590.

Feyrer, F., Herbold, B., Matern, S.A., and Moyle, P.B. 2003. Dietary shifts in a stressed fish assemblage: Consequences of a bivalve invasion in the San Francisco Estuary. Environmental Biology of Fishes, v. 67:277-288. 10.1023/A: 1025839132274

Field, J.C., Francis, R.C., and Aydin, K., 2006, Top-down modeling and bottom-up dynamics: linking a fisheries-based ecosystem model with climate hypotheses in the Northern California current: Progress in Oceanography, v. 68, p. 238–270. doi:10.1016/j.pocean.2006.02.010.

Fields, W., and Messer, C., 1999, Life on the bottom: Trends in species composition of the IEP-DWR benthic monitoring program: IEP Newsletter, v. 12, p. 38-41.

Froese, R. and Pauly, D. Editors. (2000) FishBase 2000. Concepts design and data sources: ICLARM, Los Banos, Laguna, Philippines, p. 344.

Fulton, E.A., Smith, A.D., and Johnson, C.R., 2003, Effect of complexity on marine ecosystem models: Marine Ecology Progress Series, v.256, p. 1-16. doi: 10.3354/meps06987

Grimaldo, L.F., Stewart, A.R., and Kimmerer, K., 2009, Dietary segregation of pelagic and littoral fish assemblages in a highly modified tidal freshwater estuary: Marine and Coastal Fisheries: Dynamics, Management, and Ecosystem Science. v. 2, p. 200-217. DOI: 10.1577/C08-013.1.

Hollibaugh, J.T. and Wong, P.S., 1999, Microbial processes in the San Francisco Bay estuarine turbidity maximum: Estuaries, v. 22 p. 848-862. 10.2307/1353066.

47

Hooff, R. C. and Bollens, S. M., 2004, Functional response and potential predatory impact of Tortanus dextrilobatus, a recently introduced carnivorous copepod in the San Francisco estuary: Limnology and Oceanography, v. 277, p. 167–179.

Jorgensen, S.E., Patten, B.C., and Straskraba, M., 2000, Ecosystems emerging. 4. Growth: Ecological Modeling, v. 126. p. 249–284. doi:10.1016/S0304- 3800(00)00268-4.

Kogut, N.J., 2008, Overbite Clam, Corbula Amurensis, defecated alive by white sturgeon, Acipenser Transmontanus: California Fish and Game, v. 94, p. 143-149.

Kushlan, J.A., 1976, Feeding behavior of North American herons: Auk, v. 93, p. 86-94.

Kimmerer, W. J., 2002, Effects of freshwater flow on abundance of estuarine organisms: physical effects or trophic linkages? Marine Ecology Progress Series, v. 243, p. 39-55. doi:10.3354/meps243039.

Kimmerer, W.J., 2004, Open water processes of the San Francisco estuary: From physical forcing to biological responses: San Francisco Estuary and Watershed Science, v. 2, p. 142.

Kimmerer, W., Avent, S.R., and Bollens, S.M., 2005, Variability in length-weight relationships used to estimate biomass of estuarine fish from survey data: Transactions of the American Fisheries Society, v. 134, p. 481-495. doi: 10.1577/T04-042.1.

Kimmerer, W.J., Ferm, N., Nicolini, M.H. and Penalva, C., 2005, Chronic food limitation of egg production in populations of copepods of the genus Acartia in the San Francisco Estuary: Estuaries, v. 28, p. 541-550. doi: 10.1007/BF02696065

Lauritsen, D.D., 1986, Filter-feeding in Corbicula fluminea and its effect on seston removal: Journal of the North American Benthological Society, v. 5, p. 165-172.

Libralato, S., Christensen, V., and Pauly, D., 2006, A method for identifying keystone species in food web models: Ecological Modeling, v. 195, p. 153-171. doi: 10.1016/j.ecolmodel.2005.11.029.

Lott, J., 1998, Feeding habits of juvenile and adult delta smelt from the Sacramento-San Joaquin River estuary. Interagency Ecological Studies Program for the Sacramento-San Joaquin Estuary Newsletter, Winter 1998.

Lindeman, R.L., 1942, The trophic-dynamic aspect of ecology: Ecology, v. 23, p. 399- 418. doi:10.2307/1930126

48

Ma, H., Townsend, H., Zhang, X., Sigrist, M., and Christensen, V., 2010, Using a fisheries ecosystem model with a water quality model to explore trophic and habitat impacts on a fisheries stock: A case study of the blue crab population in the Chesapeake Bay: Ecological Modeling, v. 221, p. 997-1004. doi: 10.1016/j.ecolmodel.2009.01.026.

MacFarlane, R.B., and Norton, E.C., 2002, Physiological ecology of juvenile Chinooksalmon (Oncorhynchus tshawytscha) at the southern end of their distribution, the San Francisco estuary and the Gulf of the Farallones: Fisheries Bulletin, v. 100, p. 244-257.

Matern, S.A., Moyle, P.B., and Pierce, L.C., 2002, Native and alien fishes in a California estuarine marsh: twenty-one years of changing assemblages: Transactions of the American Fisheries Society, v. 131, p. 797-816. doi: 10.1577/1548- 8659(2002)131<0797:NAAFIA>2.0.CO;2.

Mauchline, J., 1980, The biology of mysids and euphausids: .Advances in Marine Biology, v. 18, p. 681.

Mackay, A., 1981, The generalized inverse: Practical Computing, p. 108-110.

McCall, J. N., 1992, Source of harpactacoid copepods in the diet of juvenile starry flounder: Marine Ecology Progress Series, v. 86, p. 41-50.

McKechnie, R.J., and Fenner, R.B., 1971, Food habits of white sturgeon, Acipenscr transmontanus in San Pablo and Suisun bays, California: California Fish and Game, v. 57, p. 209-212.

McLean, R.B., Griffith, J.S., and McGee, M.V., 1985, Threadfin shad, Dorosoma petenense Günther, mortality: causes and ecological implications in a south- eastern United States reservoir: Journal of Fish Biology, v. 27, p. 1-12. 10.1111/j.1095-8649.1985.tb04004.x.

Mount, J.F., 1995, California rivers and streams: Berkeley, University of California Press, p. 376 .

Moyle, P.B., 2002, Inland fishes of California: Berkeley, University of California Press, p. 502

Moyle, P.B., Baxter, R.D., Sommer, T., Foin, T.C., and Matern, S.A., 2004, Biology and population dynamics of Sacramento splittail (Pogonichthys macrolepidotus) in the San Francisco estuary: A review: San Francisco Estuary and Watershed Science, v. 2, p. 47.

49

Moyle, P.B., Herbold, B., Stevens, D.E., and Miller, L.W., 1992, Life History and Status of Delta Smelt in the Sacramento-San Joaquin Estuary, California, Transactions of the American Fisheries Society, v. 121, p. 67-77.

Nichols, F.H., Cloern, J.E., Luoma, S.N., and Peterson, D.H., 1986, The modification of an estuary: Science, v. 231, p. 567-573. doi: 10.1126/science.231.4738.567.

Nobriga, M.L., 2002, Larval delta smelt diet composition and feeding incidence: Environmental and ontogenetic influences: California Fish and Game, v. 88, p. 149-164.

Nobriga, M.L., Feyrer, F., and Baxter, R.D., 2006, Aspects of the Sacramento pikeminnow biology in nearshore habitats of the Sacramento- San Joaquin Delta, California: Western North American Naturalist, v. 66, p. 106-114. doi: 10.3398/1527-0904(2006)66[106:AOSPBI]2.0.CO;2.

Nobriga, M. and F. Feyrer., 2007, Shallow-water piscivore-prey dynamics in the Sacramento–San Joaquin Delta: San Francisco Estuary and Watershed Science, v. 5, p. 13.

Nobriga, M.L., and Feyrer, F., 2008, Diet composition in San Francisco estuary striped bass: Does trophic adaptability have its limits?: Environmental Biology of Fishes, v. 83:495-503. doi:10.1007/s10641-008-9376-0.

Nobriga, M.L., Sommer, T.R., Feyrer, F., and Fleming, K., 2008, Long-term trends in summertime habitat suitability for delta smelt (Hypomesus transpacificus): San Francisco Estuary and Watershed Science, v. 6, p. 13 p.

Nobriga, M., Feyrer, F., Baxter, R., and Chotkowski, M., 2005, Fish community ecology in an altered river delta: spatial patterns in species composition, life history strategies, and biomass: Estuaries, v. 28, p. 776–785.

Oakden, J.M., 1984, Feeding and substrate preference in five species of Phoxocephalid amphipods from central California: Journal of Crustacean Biology, v. 4, p. 233- 247.

Odum, E.P., 1969, The strategy of ecosystem development: Science, v. 164, p. 262–270.

Orsi, J.J., editor, 1999, Report on the 1980-1995 fish, shrimp, and crab sampling in the San Francisco Estuary, California. Technical Report 63. Sacramento, California: The Interagency Ecological Program for the Sacramento-San Joaquin Estuary. p. 501

50

Orsi, J. J. and Knutson, A. C. 1979. The role of mysid shrimp in the Sacramento-San Joaquin estuary and factors affecting their abundance and distribution, p. 401– 408.In T. J. Conomos (ed.), San Francisco Bay: The urbanized estuary. Pacific Division, American Association for the Advancement of Science, San Francisco, California.

Palomares, M.L.D. and Pauly, D., 1998, Predicting the food consumption of fish populations as functions of mortality, food type, morphometrics, temperature and salinity: Marine and Freshwater Research, v. 49, p. 447-453. doi:10.1071/MF98015.

Pauly, D., Christensen, V., and Walters, C., 2000, Ecopath, Ecosim, and Ecospace as tools for evaluating ecosystem impact of fisheries: ICES Journal of Marine Science, v. 57, p. 697-706. doi:10.1006/jmsc.2000.0726.

Polovina, J.J., 1984, Model of a coral reef ecosystem I. The ecopath model and its application to French Frigate Shoals, Coral Reefs, v. 3, p. 1-11. doi: 10.1007/BF00306135.

Poulton, V.K., Lovvorn, J.R., and Takekawa, J.Y., 2002, Clam density and scaup feeding behavior in San Pablo Bay, California: Condor, v. 1, p. 518-527. doi: 10.1650/0010-5422(2002)104[0518:CDASFB]2.0.CO;2.

Preikshot, D., 2007, The influence of geographic scale, climate and trophic dynamics upon North Pacific oceanic ecosystem models [Ph.D. thesis]: University of British Columbia, Vancouver, p. 227.

Purcell, J.E., 2003, Predation on zooplankton by large jellyfish, Aurelia labiata, Cyanea capillata and Aequoreaaequorea, in Prince William Sound, Alaska: Marine Ecology Progress Series, v. 246, p. 137-152.

Randall, R.G., and Minns, C.K., 2000, Use of fish production per unit biomass ratios for measuring the productive capacity of fish habitats: Canadian Journal of Fisheries and Aquatic Sciences, v. 57, p. 1657-1667. doi:10.1139/cjfas-57-8-1657.

Rollwagen Bollens, G.C. and Penry, D.L., 2003, Feeding dynamics of Acartia spp. copepods in a large, temperate estuary (San Francisco Bay, California): Marine Ecology Progress Series, v. 257, p. 139–158.

Rosenfield, J.A. and Baxter, R.D., 2007, Population dynamics and distribution patterns of longfin smelt in the San Francisco Estuary: Transactions of the American Fisheries Society, v. 136, p.1577–1592. doi: 10.1577/T06-148.1.

Ruzicka J.J., Brodeur R.D., and Wainwright, T.C., 2007, Seasonal food web models for the Oregon inner-shelf ecosystem: investigating the role of large jellyfish: California Cooperative Oceanic Fisheries Investigations, v. 48, p. 106–128.

51

Schaffter R.G., and Kohlhorst, D.W., 1997, Mortality rates of white catfish in California's Sacramento-San Joaquin Delta: California Fish and Game, v. 83, p. 45-56.

Schaffter, R.G., 1997, Growth of white catfish in California's Sacramento-San Joaquin Delta: California Fish and Game, v. 83, p. 57-69.

Sobczak, W.V., Cloern, J.E., Jassby, A.D., and Muller-Solger, A.B., 2002, Bioavailability of organic matter in a highly disturbed estuary: The role of detrital and algal resources: PNAS, v. 12, p. 8101-8105. doi:10.1073/pnas122614399.

Sobczak, W.V., Cloern, J.E., Jassby, A.D., Cole, B.E., Schraga, T.S., and Arnsberg, A., 2005, Detritus fuels ecosystem metabolism but not metazoan food webs in San Francisco estuary’s freshwater delta: Estuaries and Coasts, v. 28, p. 124-137. doi: 10.1007/BF02732759.

Sommer, T., Armor, C., Baxter, R., Breuer, R., Brown, L., Chotkowski, M., Culberson, S., Feyrer, F., Gingras, M., Herbold, B., Kimmerer, W., Mueller-Solger, A., Nobriga, M., and Souza, K., 2007, The Collapse of Pelagic Fishes in the Upper San Francisco Estuary: El Colapso de los Peces Pelagicos en La Cabecera Del Estuario San Francisco: Fisheries, v. 32, p. 270-277. doi: 10.1577/1548- 8446(2007)32[270:TCOPFI]2.0.CO;2.

Stevens, D.E., Kohlhorst, D.W., and Miller, L.W., 1985, The decline of striped bass in the Sacramento-San Joaquin estuary, California: Transactions of American Fisheries Society, v. 114, p. 12-30. doi: 10.1577/1548- 8659(1985)114<12:TDOSBI>2.0.CO;2.

Thomson, J., Kimmerer, W., Brown, L., Newman, K., Mac Nally, R., Bennett, W., Feyrer, F., and Fleishman, E., 2010, Bayesian change-point analysis of abundance trends for pelagic fishes in the upper San Francisco Estuary: Ecological Applications, v. 20, p. 1417-1430. doi:10.1890/09-1724.1.

Turner, J.L., 1966, Ecological studies of the Sacramento-San Joaquin Delta Part II Fishes of the Delta: Distribution and food habits of centrarchid fishes in the Sacramento- San Joaquin Delta, Fisheries Bulletin, v. 136, p. 144-153.

Ulanowicz, R.E., 1986, Growth and development: Ecosystem phenomenology: New York, Springer-Verlag, p. 203

Villanueva, M.C.S., Isumbisho, M., Kaningini, B., Moreau, J., and Micha, J., 2008, Modeling trophic interactions in Lake Kivu: What roles do exotics play?: Ecological Modeling, v. p. 212, 422-438. doi:10.1016/j.ecolmodel.2007.10.047.

Wahle, R.A., 1985, The feeding ecology of Crangon franciscorum and Crangon nigricauda in San Francisco Bay, California. Journal of Crustacean Biology, v. 5, p. 311-326.

52

Wolff, M., 1994, A trophic model for Tongoy Bay: A system exposed to suspended scallop culture (Northern Chile): Journal of Experimental Marine Biology and Ecology, v, 182, p. 149-168. doi:10.1016/0022-0981(94)90048-5.