1. Using the minimum distance as a metric for testing is ambiguous when fish have greater than two channels to move into. Fish GPS position during each day of sampling from radio telemetry results for two separate releases at Georgiana Slough are an example of this (Figure 4 and Figure 5). The same distance may result even though fish are at different locations in the network. In this case it may be more appropriate to compare the probability a particle is in a specific segment at a given time to observed fish locations at that time. Juvenile Salmon Movement in the Sacramento-San Joaquin Delta Challenges in Using Field Data to Validate Models Annjanette M. Dodd 2005 CALFED Science Fellow, Humboldt State University, Arcata, CA, amd2@humboldt.edu Model Testing Using Field Data Delta’s intricate channel network and influence of tides, combined with low observation frequency, make it difficult to determine exact routes of observed fish and challenging to test model results. Metrics for model testing that can be evaluated with field data include: minimum distance fish travels in a given time, direction a fish goes at a junction, probability a fish has passed a point at a given time, probability a fish is in a segment at a given time. Radio telemetry results for 18 fish released at Ryde were used to estimate the distance from the release location (Figure 3). Tidal influence on fish cannot be directly determined due to the extended periods with no monitoring. Estimates of minimum distance from release location to fish position at each time can be used to compare patterns between observed fish and simulated particles. Project Objective Create a model of juvenile movement by developing alternative models of fish behavior to merge with a particle transport model that incorporates Delta hydrodynamics. The strategy is not to produce a “realistic” model of fish movement, but to find the simplest model that is consistent with available data by hypothesizing alternative models of behavior and then testing them to determine which (by themselves or in combination) best reproduce patterns observed in the real system. Complex hydrodynamics resulting from river flow, tides, and diversions, make it difficult to distinguish movement due to behavior from that resulting from advection. The null hypothesis of juvenile salmon movement in the Delta is that the fish behave as passive particles and just “go with the flow”. This will initially be tested by comparing the results of radio telemetry studies with simulation results for passive particle movement. Introduction The Delta is an elaborate system of channels that drain the waters of the Sacramento River, San Joaquin River, and other smaller tributaries into San Francisco Bay located in northern California (Figure 1). A better understanding of how juvenile salmon move through the Delta is needed to improve water management actions. How juvenile salmon move through the Delta affects their probability of successfully smolting and migrating to the ocean. For example, managers need to understand how changes in flows, either natural or from water exported through diversion pumps, affect the probability of success of juvenile salmon in reaching the ocean. Next Steps Perform multiple simulations of passive particle movement within the Delta under the same conditions as the radio telemetry studies. Compare simulation results and radio telemetry study results to evaluate different metrics for model testing and test the null hypothesis of fish movement. Perform simulations of passive particle movement at the entire Delta scale to compare with coded wire tag release and catch data to further evaluate the different metrics and the null hypothesis of fish movement. Acknowledgements Research and transportation costs funded by the CALFED Science Fellows Program in Cooperation with the California Sea Grant College Program. I would like to thank Dr. Roland Lamberson, Humboldt State University; Steve Railsback, Lang Railsback & Associates; Mathew Nobriga, California Department of Water Resources; and Bret Harvey, USDA Forest Service for their continued support and assistance on this project. Figure 1: Sacramento-San Joaquin delta with radio telemetry study locations and approximate monitored fish paths during the studies. Figure 2: Radio telemetry results of 50 tagged juvenile salmon (red circles) released at two locations over the entire monitoring period in 2000. Area in red represented in Figure 2 Source: www.resourcescientists.com Downstream movement dominant Up and downstream movement with tides Ryde Rio Vista Georgiana Cache Slough Field Data - Radio Telemetry Studies Juvenile Chinook with attached miniature radio transmitters were released at various locations (Figures 1 and 2). Fish were tracked using boat mounted radio receivers between 9:00 and 17:30 hours over 3 to 4 days. About 18 different releases in all. Figure 3: Distance from release location for 18 fish released at Ryde on the Sacramento River in 2002 with stage and flow at Rio Vista. Negative distance means fish went upstream into Cache Slough (Figure 2). Delta Hydrodynamics and Particle Tracking Model Developed by the California Department of Water Resources Used to simulate passive particle movement through the Delta Spatially-detailed one-dimensional hydrodynamic model of the entire Delta Already calibrated to the Delta Particle Tracking is pseudo three-dimensional by applying transverse and vertical velocity variations as well as turbulent mixing on the one-dimensional channel velocities Figure 4: GPS locations during 4 day survey period of 13 fish released at Georgiana Slough in 2000. Black diamonds represent node locations in hydrodynamic model. Figure 5: GPS locations during 4 day survey period of 16 fish released at Georgiana Slough in 2002. Black diamonds represent node locations in hydrodynamic model. Georgiana Ryde Fish paths of travel for releases at Ryde and Georgiana Slough Source: CALFED Science Program San Joaquin River Mokelumne River Stems Sacramento River Georgiana Slough  N  N  N Central Delta South Delta