Edited by Nigel Haggan and Tony J. Pitcher.

DIRECTOR'S FOREWORD

Modelling marine ecosystems is not that difficult and it certainly is useful. This document, devoted to a model of the west coast of Scotland, illustrates this very well. Here, the usefulness of the model is based on two features:

1. It demonstrates that one can not only account for the feeding relationships of all groups within such a complex system, but also account for the distribution in space of these same group, i.e., their habitat preferences; 2. It helps identify data gaps and researchable topics which would lead to a better understanding of the waters of western Scotland.

Dr. Daniel Pauly
Director, UBC Fisheries Centre

ABSTRACT

Trophic flows of the west coast of Scotland (WCS) ecosystem (ICES zone VIa) for the period 1995 to 2000 were reconstructed using Ecopath. The Ecopath model is divided into 37 functional groups or trophic compartments ranging from phytoplankton and detritus to marine mammals and seabirds, and including harvested species of pelagic, demersal, and benthic environments. We present here details of the input data (biomass, production, consumption, catch, and diet composition) for each compartment used for modelling. This work represents the first ecosystem model for the west coast of Scotland and is the result of an important assemblage of data on the biological characteristics of species occurring in the zone VIa.

Fisheries data for 1995-2000 f0r all commercial species was collated from annual statistics provided by the appropriate fishing country. All fisheries combined are estimated to collect 1.36 tons of marine organisms per square kilometre per year, approximately 70% of what might be expected from ICES data for area VIa. Preliminary estimates of discards total 0.51 tokm-2 for a total extraction of 1.87 tokm-2, although the discard figure is thought to be low. The catches were divided into 8 gear types: demersal trawl, beam trawl, midwater trawl, dredge, purse seine, lines (handline and longline), creels and pots and miscellaneous gear types (Table 4).

Simple tests suggest that the Ecosim model for the west coast of Scotland is behaving in a reasonable manner under dynamic simulations. We assign the critical prey-predator vulnerability parameters using a short-cut method which has been validated by previous work, and use three procedures to test the model's dynamic performance. 1.) An equilibrium analysis determines for each commercial functional group the long-term catch rate and biomass level that would result under varying degrees of fishing mortality. 2.) Pulse fishing simulations reveal how quickly the ecosystem can recover from disturbance. 3.) All fishing pressure is removed from the dynamic simulation to show the recuperative potential of commercial groups and secondary trophic effects throughout the ecosystem. Much more can be done to improve the dynamic behaviour of Ecosim, but the model appears to be robust and free from instabilities.

Preliminary spatial simulations were set up by mapping the 31,085 km-2 model area into 450 cells of 69.1 km-2 each (nominally square, with 8.3 km sides). An approximate coastline was sketched in on the map, avoiding isolated bays at the edge of the model area, with 335 marine cells. Four habitats were allocated using depth zones 0-10 m (representing 19% of total model area); 10-100 m (46%); 100-200 m (28%); and 200-1000 m (2%). Primary production levels were allocated for each model square from the Sea Around Us database. The 37 functional groups and 8 fisheries in the model were allocated to their preferred habitats in suitable combinations of the 4 habitat categories. All of these allocations were performed in a preliminary fashion, and need to be validated with local data in future refinements to the model. Likewise, for each model group in its preferred and non-preferred habitats, using information from similar Ecospace models elsewhere, we adjusted default dispersal rates in kmoyear-1, the relative dispersal rate in bad habitats, the relative vulnerability to predation in bad habitats, and the relative feeding rate in bad habitats. Fishery management zones, termed "MPAs" in the software, were set up as examples; results may also be obtained in separate designated zones. The section concludes with a demonstration 50-year spatial simulation of the west coast of Scotland under default assumption of no changes to the existing fishing effort. Before attempting to analyse realistic spatial management scenarios, the present Ecospace model of the west coast of Scotland should be used in a diagnostic mode by running trial scenarios with large no-take zones, under progressive annual increases and decreases in fishing power, in order to refine the habitat-related dispersal parameters and the underlying Ecosim and Ecopath model structure.

The preliminary Ecopath, Ecosim and Ecospace models will benefit from more input from local experts. Workshops with scientific and lay experts on individual species fisheries and other areas would be highly recommended. The resulting models would then lend themselves to a number of research priorities identified in the course of the project, some of which are identified at the end of the report.

TABLE OF CONTENTS