Development, application and evaluation of a 1-D full life cycle anchovy and sardine model for the North Aegean Sea (Eastern Mediterranean)

Abstract

A 1-D full-life-cycle, Individual-based model (IBM), two-way coupled with a hydrodynamic/biogeochemical model, is demonstrated for anchovy and sardine in the N. Aegean Sea (Eastern Mediterranean). The model is stage-specific and includes a 'Wisconsin' type bioenergetics, a diel vertical migration and a population dynamics module, with the incorporation of known differences in biological attributes between the anchovy and sardine stocks. A new energy allocation/egg production algorithm was developed, allowing for breeding pattern to move along the capital-income breeding continuum. Fish growth was calibrated against available size-at-age data by tuning food consumption (the half saturation coefficients) using a genetic algorithm. After a ten-years spin up, the model reproduced well the magnitude of population biomasses and spawning periods of the two species in the N. Aegean Sea. Surprisingly, model simulations revealed that anchovy depends primarily on stored energy for egg production (mostly capital breeder) whereas sardine depends heavily on direct food intake (income breeder). This is related to the peculiar phenology of plankton production in the area, with mesozooplankton concentration exhibiting a sharp decrease from early summer to autumn and a subsequent increase from winter to early summer. Monthly changes in somatic condition of fish collected on board the commercial purse seine fleet followed closely the simulated mesozooplankton concentration. Finally, model simulations showed that, when both the anchovy and sardine stocks are overexploited, the mesozooplankton concentration increases, which may open up ecological space for competing species. The importance of protecting the recruit spawners was highlighted with model simulations testing the effect of changing the timing of the existing 2.5-months closed period. Optimum timing for fishery closure is different for anchovy and sardine because of their opposite spawning and recruitment periods.

Fig 1. Map of the North Aegean Sea showing the model domain.

The box indicates the location of Thracian Sea.

Fig 2. Representation of the anchovy and sardine model coupled with the lower trophic level (LTL) model.

Fig 3. Schematic illustration of the energy allocation and egg production algorithm.

SST: Sea surface temperature. L_m: length at maturity. E_buffer: energy in reproduction buffer. E_egg: batch energy.

Fig 4. Seasonal evolution of temperature and simulated mesozooplankton concentration.

Fig 5. Top panel: Simulated average mesozooplankton concentration (mgC m^-3) in the water column (0-100m) against calendar day.

Bottom panel: length-adjusted monthly mean weight (somatic condition) of fish samples collected onboard the Thracian Sea purse seine fleet in 2003–2008.

Fig 6. Mean length-at-age (±SD) of anchovy and sardine larvae, calibrated using the genetic algorithm and field data ([49], [65], [36]).

Fig 7. Evolution of mean weight and mean length of fish, calibrated using the genetic algorithm and mean weight- and length-at-age of adult fish (±SD) estimated from samples collected during the acoustic and egg production surveys in the North Aegean Sea, 2003–2008 [42].

Fig 8. Model-simulated anchovy and sardine biomass.

The mean biomasses of the two species in the N. Aegean Sea (based on acoustic data biomass estimations for the period 2003–2008) are also shown.

Fig 9. Model-simulated daily egg production (total number of eggs produced by the population) for recruit (age 1) and repeat spawners (age 2+).

The seasonal evolution of sea surface temperature (SST) is also shown.

Fig 10. Biomass of anchovy and sardine and mean mesozooplankton concentration for different combinations of exploitation rate (E.R.) of the two species.

E.R. = 0.4 is the reference point (maximum sustainable yield proxy) currently used in the management of small pelagic fish stocks in the Mediterranean Sea.

Fig 11

Mean anchovy and sardine biomass (upper panel) and annual population fecundity (middle panel) in relation to the timing of the 2.5 months fishing ban. Months 1, 2, 3,… etc correspond to closed period 15 Jan-Mar, 15 Feb-Apr, 15 Mar-May, …etc. The mean weight of individuals during the respective closed period is also plotted (lower panel).