О Jorn - Impact of prey field variability on early cod larval survival a sensitivity study of a baltic cod - страница 1
Impact of prey field variability on early cod larval survival: a sensitivity study of a Baltic cod
OCEANOLOGIA, 50 (2), 2008.
Predator prey interactions Biophysical model Baltic Sea Spatial variations IBM
© 2008, by Institute of Oceanology PAS.
Jorn О. Schmidt
Leibniz Institute of Marine Sciences
at the University of Kiel (IFM-GEOMAR)
Diisternbrooker Weg 20, D-24105 Kiel, Germany;
* corresponding author
Received 18 February 2008, revised 3 June 2008, accepted 3 June 2008.
Existing coupled biophysical models for Baltic larval cod drift, growth and survival use idealised constructed mean prey fields of nauplius distributions. These simulations revealed the best feeding conditions for Baltic cod larvae longer than 6 mm. For shorter, first feeding larvae (between 4.5 and 6 mm) pronounced differences in growth and survival were observed, which depend on food availability and to a lesser degree on ambient temperature. We performed runs with an Individual-based Model (IBM) for Baltic cod larvae in order to demonstrate how natural variability in prey abundance influences the survival success of first feeding larvae. In the Baltic, this larval stage lives mainly between 20 and 40 m depth and feeds exclusively on the nauplii of different calanoid copepods (Acartia spp., Pseudocalanus acuspes, Temora longicornis and Centropages hamatus). Prey data
Hinrichsen et al. (2002) and others (Werner et al. 1996, Lough et al. 2005, Lough & Broughton 2007) were aware of the patchiness of larval prey, prey fields were averaged over large scales. In this paper we investigated the influence of prey field variability on cod larval survival in the Bornholm Basin (Baltic Sea) with field data obtained from a small spatial and temporal sampling in 2001 and from a basin-wide (mesoscale) monthly sampling in 2002. Further, we compared model outputs obtained from literature-based sizes of nauplii with a measured size distribution from samples obtained in the Baltic Sea in June 2001.
2. Material and methods
Copepod nauplii were sampled on cruises in June 2001 and in April, May and July 2002. In June 2001, six profiles were taken at one station covering day and night in a 48 h period (Figure 1). In 2002, sampling was performed at nine stations (Figure 1), the net being deployed only once at each station. In 2002, in parallel with the zooplankton sampling, hydrographic parameters (temperature and salinity) were recorded at each station using an ADM-CTD (Analoge und Digitale MeBsysteme GmbH). In 2001 this data was not recorded (see below).
Figure 1. The Bornholm Basin study area in the Baltic Sea (overview in the upper left corner); the black dot shows the station sampled in June 2001; the open circles show the stations sampled in April, May and July 2002
Table 1. Length [mm] and weight [/tin] of the different stages of the different species, and of sibling species where data for the original species was not available (Pseudocalanus minutus, Centropages typicus, Acartia clausi) taken from Ogilvie (1953); also the length-weight relationship given by Culver et al. (1985) for nauplius dry weight = a x lengthb (where a = 3.009 and b = 1.706)
Zooplankton samples were collected with a 0.25 m2 multi opening/closing net system (Multinet, Hydro-Bios Kiel). The Multinet was equipped with five nets of 50 /m mesh size. The gear was applied vertically with a down-and upward speed of 0.2 m s_1, resolving the water column into 10 m intervals from the bottom up to the surface, and filtering approximately 2.5 m3 of water. Clogging, with the corresponding change in filtering efficiency, was not observed. Nauplii were preserved in 4% borax buffered formalin seawater solution and sorted in the laboratory under a binocular microscope. All nauplii except Acartia spp. were identified to the species level. The abundance was calculated as the number per cubic metre for each species. The biomass was calculated as the dry weight, using the length for each nauplius stage taken from Ogilvie (1953) as the basis for the growth term in the model (Culver et al. 1985) (Table 1). In June 2001 the length of Pseudocalanus acuspes nauplii was additionally measured, irrespective of stage, in order to obtain real size distributions. The measurements were done using ImageJ software (2003) with a precision of ± 5 /m. For our model calculations, only the nauplius abundances in the 20-30 m and 3040 m depth layers were taken into consideration: these are the layers where first feeding cod larvae mainly live.
Model description and design of simulations
The individual-based model (IBM) used in this study was developed by Hinrichsen et al. (2002) using the generalised model described by Letcher et al. (1996) (Figure 2). This IBM tracks individual cod larvae through all larval phases including first feeding (4.5-6 mm). In the IBM, prey encounter rate, foraging, growth, survival and nutritional condition in terms of weight per length of larvae is simulated by specific sub-models in 6-h time steps.
The basis of this IBM is the standard bioenergetic supply-demand function (Beyer & Laurence 1980, Carlotti & Hirche 1997), where growth is represented as the difference between the amount of food ingested by a larva and the metabolic costs of its daily activities. To fill gaps in the data for Baltic cod, length-weight relationships from Otterlei et al. (1999) were applied to compute the expected weight under superabundant prey concentrations. Deviations of the simulated weight from the bioenergetic model and the maximum weight per length were used to run the starvation model. Larvae were defined as dying from starvation if the final weight per a specific length fell below specific threshold values, i.e. 75% of the maximum weight per length of larvae (Letcher et al. 1996).
Larval prey conditions were taken from in situ zooplankton measurements as described above. Each sample was assumed to be a 'potential prey field' that individual larva experienced during their first feeding phase.
Prey density, size *
Swimming speed, Reactive area, Turbulence
Capture success, Handling
Assimilation efficiency, Metabolism
Figure 2. A scheme of the individual-based model (IBM) used in this study; the investigated parameter is marked with an asterisk. Modified after Letcher et al.
For June 2001 six profiles times two depth bins resulted in 12 'potential prey fields', and for 2002 nine profiles each month times two depth strata resulted in 18 'potential prey fields'.
To calculate the survival of the larvae experiencing the prey fields taken in June 2001, the IBM was coupled to the circulation model described in Hinrichsen et al. (2002). This was done primarily to obtain an estimate of the small-scale temperature variability the larvae experienced in this investigation period (Table 2). In the simulation, a cohort of larvae was
Table 2. Minimum and maximum temperature during the investigation periods at 20-40 m depths. The temperature in 2002 was measured in situ; the temperature in June 2001 was derived from the circulation model
5.8 - 7.9
2.7 - 3.9
3.2 - 6.9
4.5 - 17.2
released and tracked for a period of 15 days. This time span approximately covers the first feeding period of larvae at June temperatures (Hinrichsen et al. 2002). 550 larval drifters were released at depths where first feeding larvae mainly live (26, 28, 30, 32 and 34 m) on a regularly spaced grid with a horizontal resolution of 500 m. To enable comparison of the model results with different prey field compositions, at the start of the model run (6 June) all larvae were defined as being of equal length and weight (4.5 mm). The larvae experienced specific temperature conditions along the drift trajectories within the coupled model. Only a one-dimensional version of the IBM was performed during all the investigation periods in 2002; larval drift was not taken into account. The temperature variability obtained from basin-wide measurements from each sampling period within the depth range of 20-40 m (Table 2) was split into 0.1°C steps for the model. One model run was performed for each 'prey field' and each temperature category. We assumed that with respect to the temporal development of nauplii, their abundances and sizes remained constant for the whole duration of the simulation periods, which were between 12 and 14 days, depending exclusively on the ambient temperature conditions. The sizes of the different copepod stages for each species were obtained from the literature (Ogilvie 1953). Where data for a species was not available we used data for a sibling species (Table 1).
Another set of model runs was performed using the abundances of Pseudocalanus acuspes in the 2001 samples with the real size distribution obtained from measurements and with the average size for P. acuspes taken from the literature (Ogilvie 1953). Two threshold values for nauplius abundance were calculated: (i) to ensure larval survival, and (ii) to gain maximum weight-at-age for the size range of 4.5 mm to 6 mm. To obtain some idea of the behaviour of the model, the prey density for a given prey weight needed for a larva with a given length at 5°Ctosurviveone model step was calculated, assuming that the larva enters the model with maximum weight per length.
3. Results Prey fields
In June 2001, Pseudocalanus acuspes nauplii were the most abundant of the four species within the 20-40 m depth stratum. Differences in abundance between samples were more than tenfold, from 369 to 5358 individuals per cubic metre (Figure 3). Peak abundance of the other species was only 120 individuals for Temora longicornis, 200 for Centropages hamatus and 80 for Acartia spp.
5000.0 " 500.0
50.0 5.0 0.5
50.0 5.0 0.5
50.0 5.0 0.5
5000.0 500.0 50.0 5.0 0.5
Figure 3. Abundances of the four copepod species estimated from multinet samples on a logarithmic scale; ■ - median, box - 25 and 75 percentiles, _L -minimum, T - maximum; note the different scaling