Mixed-Resolution Models for Investigating Individual to Population Scale Spatial Dynamics

Progress Reports (PDF): FY 2006, FY 2005, FY 2004

Project Overview
This project forms part of a series of studies that are being conducted to facilitate the ambitious multi-national Oceanic Fisheries and Climate Change Project (OFCCP) of the International GLOBEC Program – a component of the International Geosphere-Biosphere Programme (IGBP), sponsored by the IOC and SCOR. The OFCCP will investigate the effect of climate change on the productivity and distribution of oceanic tuna stocks and fisheries in the Pacific with the goal of predicting short–to long term changes and impacts related to climate variability. The ultimate goal of the OFCCP project is to conduct simulation with ecosystem models that include the main tuna species, using an input data set predicted under a scenario of climate change induced by greenhouse warming as defined by the Intergovernmental Panel on Climate Change (IPCC). However, analyses of simulations based on retrospective series of oceanographic and fishing data sets are necessary to improve upon the predictive capacity of the ecosystem models, particularly the inter-annual (ENSO) and the decadal (Pacific Decadal Oscillations) time scales.

This project, Mixed Resolution Models for Investigating Individual to Population Scale Dynamics, addresses ways improve upon two classes of models: Individual Based Models (IBMs) and Advection Diffusion Reaction Models (ADRMs) that would help to model from ocean basin to individual scale as identified by the OFCCP objectives. Both these types of models have been successfully applied to predicting tuna behaviors; IBMs at the very fine scale and ADRMs at the population level.

The two classes of models can provide complimentary approaches to investigating the problems of scale integration when going from individual to the population level and from individual movements to advection-diffusion patterns. However, the approach needs a unifying framework combining large and small spatio-temporal scales i.e., the mixed resolutions in a same model domain. Mixed resolution models use a stretched grid system with greater resolution at one of multiple locations of the model domain.

The specific objectives of the project are:

• To develop a unifying framework based on the development of mathematical solutions for using mixed-resolution in ADRMs and combine IBM and ADRM approaches.
  
• To develop a range of IBMs for tunas (skipjack, yellowfin, bigeye, albacore) and other species of importance for tuna fisheries management. For instance, enhanced understanding of turtles' spatial dynamics might lead to effective bycatch mitigation.
  
• To produce predicted field of key variables (temperature, currents, primary and zooplankton production, forage biomass) at the scale of the Pacific basin, with several focus areas where the resolution will be enhanced. The predicted environment in these focus areas will serve for IBMs simulations.
  
• To select focus areas according to their interests and the existence of data allowing validation of the simulations and (consequently) parameterization of the models.
  
• To analyze scale effects and compare movement generated from IBMs and ADRMs between them and with observations.
  
• To investigate impacts of seamounts, anchored or drifting FADs on individual behavior from IBM simulations and observations, in the search for deducing consequences at large-scale population level.

Proposed activities

The mathematical and programming developments required for improving movement spatial population dynamics models would form two post-doctoral studies. The first study will focus on developing fundamental equations and their numerical solutions to be incorporated in the population dynamics models based on tuna tagging data, or on environmental constraints (Spatial Environmental Population Dynamics Model, SEPODYM) for using mixed resolution grids. Further work on the fundamental PDEs and their numerical solutions is necessary to combine movements at the scales of both populations and individuals. The possibility of using non-Fickian formulation of the diffusion equation will be explored. More accurate and faster solution methods will be sought in conditions where diffusive movements are greater than advective movements. Variable grid spacing finite differences methods will be developed that are consistent with the individual based models.

The second post-doc study would be devoted to the development of IBMs that will use the oceanic environment predicted by the ESSIC (Earth Science System Interdisciplinary Center) coupled physical-biogeochemical model (temperature, currents, plankton) and the SEPODYM model (forage). Behavior of tuna or other large pelagics predicted with IBMs would be compared to observed movements of individuals marked with electronic tags and to spatial patterns generated by ADRMs.

Three geographic regions of focus have been identified which will serve to compare several observed and predicted key variables:

1. The Transition Zone Chlorophyll Front (TZCF)
   The Transition Zone Chlorophyll Front stretches across 8000 km in the North Pacific with seasonal north-south shifts about 1000 km. Satellite telemetry data on movement of loggerhead turtles and detailed fisheries data for albacore tuna show that these apex predators travel along the TZCF as they migrate across the North Pacific. The position dynamics and the meandering in the TZCF, related to El Nino and La Nina events, have shown to alter the spatial distributions of loggerhead turtles and the catch rates of albacore tuna.  Proposed activities include simulations with the model SEPODYM that will be used to investigate the movement of albacore along the TZCF. IBM simulations will also be used to study the individual movement of tuna and turtles along the front.
   
   
2. Hawaii Islands
   A number of tuna tagging studies have been carried out (conventional tagging, archival tagging and acoustic telemetry) around the Hawaiian Islands. An array of "listening stations" is currently being deployed around the island of O'ahu to collect movement data of yellowfin and bigeye tuna equipped with coded transmitters. There are also plans to deploy listening stations around the other Hawaiian Islands and tagging other pelagic species (sharks, marlins, mahimahi, etc.) with transmitters. These meso-scale data will complement data collected in the past on very fine-scale movements (acoustic telemetry, archival tagging) and on larger-scale movements of tuna (conventional tagging). The study of individual behavior between the different sites around Hawaii Islands will be helpful to understand the mechanism of fine-scale aggregations around the floating objects and seamounts. The framework proposed here of combining IBMs and spatial population models will be used to assess the consequence of extrapolating the individual behaviors up to the population level.
   
   
3. Southern Subtropical Convergence Zone (STCZ)
   Studies on this geographic region will target on understanding the spatial dynamics of tuna forage using acoustic methods. During an upcoming research cruise, planned for mid 2003 in the STCZ area (R.V. Tangaroa, NIWA), measurements of micronekton target strength will be made in conjunction with target identification by opening-closing net and cameras. These data will permit discrimination of acoustic targets based on size and dB difference between frequencies so that krill and shrimps can by distinguished from mesopelagic fish. Another more recent discrimination technique will also be used to separate the backscatter of swimbladder from non-swimbladder organism. The combination of both methods allows the partitioning of acoustic backscatter into three broad categories: swimbladder fish, non-swimbladder fish/squid, and krill/shrimp. Net samples of microneckton will be collected for identification of species and size distributions in these three categories. The 2003 cruise data along with earlier data of 1997, 1999 and 2001 collected in cross-front transects will be analyzed to quantify the spatio-temporal variability of micronekton (tuna forage) across the frontal zone. The results of this analysis will be used to parameterize the forage component in the SEPODYM model (biomass transfer) and to validate the predicted spatial forage distribution. Other variables (temperature, salinity, currents, chlorophyll) collected during these cruises will also be useful for fine-scale validation of the fields predicted by the ESSIC model. Impacts of predicted forage (micronekton) spatial distribution on the spatial distribution of south Pacific albacore will be investigated using IBMs and SEPODYM.
   

Year 1 funding for this 2-year project to be awarded in early 2003.