1. “IDEALIZED” WEST COAST SIMULATIONS Numerical domain Boundary conditions Forcings Wind stress: modeled as a Gaussian random process - Statistics (i.e., mean, variance & time scale) determined for a typical July using NDBC buoy data off Pt. Sur Alongshore pressure gradient: modeled as a body force - Determined using dynamic hight differences between Pt. Conception and Pt. Arena from July mean CalCoFI data SOURCE / DESTINATION RELATIONSHIP FOR LARVAL TRANSPORT Time series of larval settlement Alongshore travel distance of settlers Source/destination relationship for larval transport Short PLD larvae Long PLD larvae Stochastic Larval Settlement in Nearshore Marine Ecosystems S. Mitarai 1, D.A. Siegel 1 and K.B. Winters 2 1 Institute for Computational Earth System Science, University of California, Santa Barbara, Santa Barbara, CA 93106-3060 2 Integrative Oceanography Division, Scripps Institution of Oceanography, La Jolla, CA 92037 ABSTRACT Key to the predictive understanding of nearshore marine ecosystems is the transport of larvae by ocean circulation processes. Only a very few lucky larvae successfully settle upon suitable habitat and are able to recruit to adult life stages. Methodologies for predicting this source/settlement relationship for larval transport are still primitive and simple diffusive scaling analyses are used for many important applications. Here, we investigate source/settlement relationships of the larval transport using idealized Regional Ocean Model System simulations of time evolving coastal circulations and Lagrangian particles which are released and tracked as models of planktonic larvae. Simulation results are used to construct dispersal kernels which describe the source/destination relationships of larval transport. These dispersal kernels are strong functions of several time scales including the planktonic larval duration, the frequency and duration of larval release events and inherent coastal circulation time scales. For typical applications (such as fish stock assessment), larval dispersal is far from a simple diffusive process and consideration of the stochastic nature of larval dispersal is required. This work provides new insights into the persistence and spatial structure of nearshore fish stock abundances. SIMULATION FIELD AND VALIDATION Instantaneous field: sea level & surface velocity Mean temperature field: simulation vs. CalCoFI data PARTICLE TRACKING AND VALIDATION Sample particle trajectories Lagrangian statistics MODELING OF “LARVAL PARTICLES” Define “larval particles” Passive particles that move with local currents Each contains many (e.g., 10,000) larvae – considered as a bolus of larvae Larvae settle (stop advecting) when they arrive nearshore - Nearshore = 2 km from coast Two species considered Short PLD larvae: Need to settle within a window of 5 to 10 days Long PLD larvae: Need to settle within a window of 20 to 40 days Larval particle release One release each day for the entire season (90 days) Releases are made within 64 km from coast (every 4 km) Near surface SUMMARY & FUTURE PLANS Summary Idealized west coast simulations are presented, in which larval dispersion & settlement is investigated Simulations show a reasonable agreement with observation data Larval settlement is intermittent & heterogeneous when viewed on intergenerational time scales The modeling of larval settlement is significantly different from typical diffusion modeling approaches Future plans Develop other flow scenarios (e.g., a Southern California case, winter) Consider subgrid-scale dispersion and the dispersion of initially-adjacent larvae in determining the number of independent releases that are modeled Enable larval particles to have simplified “behaviours” and address its role in larval dispersion Construct a simple larval dispersion model for use in fish stock/harvest metapopulation model White circle: particle release point Red circles: particle location 30 days after particle release Numbers: particle released date Blue lines: particle trajectories Good qualitative agreement with CalCoFI seasonal mean ● Reasonable agreement with diffusion model ● Does not mean that source/destination relationships can be predicted using diffusion models PLD is time particles have spent in the plankton before settling Larval particle settlement to a 4 km subpopulation are shown Larval settlements are intermittent & ranges of PLD are seen Heterogeneous mapping in simulation while homogeneous in diffusion model Some sources produce many successful settlers while some produce none Travel distance (or pattern) differs depending on source locations Simulation data Surface drifter data (Swenson & Niiler,1996) Time scale Length scale Diffusivity zonal/meridional zonal/meridional zonal/meridional Shows vortex structures Rossby radius of deformation ~ 10 km 2.7 / 2.9 days 29 /31 km 4.0 / 4.3 x10 7 cm 2 /s 2.9 / 3.5 days 32 /38 km 4.3 / 4.5 x10 7 cm 2 /s Short PLD larvae Long PLD larvae Side view Top view 2-km horizontal resolution Coast GOALS FOR THIS STUDY Understand the role of larval transport in predicting nearshore fish stocks & its proper management Investigate source/destination relationships for Lagrangian particles which originate & settle in nearshore environments Use “idealized” realizations of coastal circulation tied to real data Statistically stationary & homogeneous in the alongshore direction Use these results to develop simple models of larval dispersal for use in fish stock / harvest models Periodic Free-slip Open BC: Inflow: nudging Outflow: radiation Wind stress Pressure Periodic Temperature nudging layer Simulation (mean over 180 days) CalCoFI data (Line #70, July) Results imply that one season (i.e., typical time interval for larval production) is not long enough to achieve homogeneous dispersion (as diffusion model does) http://www.icess.ucsb.edu/~satoshi/f3 Flow, Fish & Fishing – A Biocomplexity Project Simulation Diffusion model