Population Dynamics of Benthic Species and Shorebird Predation
321
and at the end of the season. The average individual mass was calculated as
the average biomass density (averaged over sites) divided by the average
numerical density. Production of those individuals that did not survive until
the first summer sampling could not be taken into account. For the summers
of 1971 and 1972 density and biomass density data were missing and production could not be calculated. For the summers 1973-1979 only data on biomass density were missing, and for each age-class a long-term average
individual mass was used instead for that period. Sampling variability of the
estimates was estimated by the bootstrap procedure (200 trials; at each trial
11 transects were re-sampled). Efron's first-percentile method was used to
obtain 90 % bootstrap confidence intervals (CI), i. e. the interval is given by
the values that exceed 5 and 95 %, respectively, of the generated bootstrap
distribution. The among-years variability (not to be confused with the
sampling variability) was expressed by the among-years standard deviation
(SD).
Monthly high-water roost count data for the knot and oystercatcher were
available from July 1975 to June 1998. The missing values for the knot, i.e. 41,
and for the oystercatcher, i. e. 39, (out of 276) were imputed by a log-linear
model with main effects year and month (Van der Meer et al. 1996). Food
demands were based on a field metabolic rate (power) of 4 W for the knot
(Wiersma and Piersma 1994). For the oystercatcher, we used a power of 8.8 W,
which is slightly higher than the estimate of 7.8 W at thermoneutrality for an
oystercatcher weighing 0.52kg that Zwarts et al. (1996) based on a literature
survey. Yet, winter temperatures in the Wadden Sea are slightly lower than the
critical temperature of 10°C below which the costs of thermoregulation increase (Kersten and Piersma 1987). We assumed that the available foraging
time per day equals 12 h, the energy density of the bivalves is 22 MJ/kg ashfree dry mass (Beukema 1997) and the gut absorption efficiency is 0.8. The
power values are then equivalent to a required food intake rate while foraging
of 0.45 and 1.0 mg/s for knot and oystercatcher, respectively.
For each winter and summer season in the period 1973-1998 estimates of
production were obtained. For M. balthica average production in summer
(1.36g/m2, SD 0.528, CI 1.13-1.67) and in winter (1.03g/m2, SD 0.693, CI
0.71-1.26) were almost equal. In their first winter the zero-year class individuals are too small (about 6 mm) to be profitable as a food source for the two
bird species. The older year classes, those that can be eaten, were responsible
for about two-thirds of the total winter production (0.69 g/m2, SD 0.516, CI
0.49-0.85). For C. edule average production in winter was 6.53 g/m2 (SD 4.57,
CI 3.97-9.36) and in summer 4.64g/m2 (SD 4.78, CI 2.98-6.95). The zero-year
class, which is edible for knots, contributed on average about one-fifth
(1.31 g/m2, SD 1.34, CI 0.70-2.07) to total winter production. Production
varied considerably from year-to-year in both species (Fig. 15.2). The same
holds for the production/biomass ratio, particularly for the cockle (Fig. 15.3).
321
and at the end of the season. The average individual mass was calculated as
the average biomass density (averaged over sites) divided by the average
numerical density. Production of those individuals that did not survive until
the first summer sampling could not be taken into account. For the summers
of 1971 and 1972 density and biomass density data were missing and production could not be calculated. For the summers 1973-1979 only data on biomass density were missing, and for each age-class a long-term average
individual mass was used instead for that period. Sampling variability of the
estimates was estimated by the bootstrap procedure (200 trials; at each trial
11 transects were re-sampled). Efron's first-percentile method was used to
obtain 90 % bootstrap confidence intervals (CI), i. e. the interval is given by
the values that exceed 5 and 95 %, respectively, of the generated bootstrap
distribution. The among-years variability (not to be confused with the
sampling variability) was expressed by the among-years standard deviation
(SD).
Monthly high-water roost count data for the knot and oystercatcher were
available from July 1975 to June 1998. The missing values for the knot, i.e. 41,
and for the oystercatcher, i. e. 39, (out of 276) were imputed by a log-linear
model with main effects year and month (Van der Meer et al. 1996). Food
demands were based on a field metabolic rate (power) of 4 W for the knot
(Wiersma and Piersma 1994). For the oystercatcher, we used a power of 8.8 W,
which is slightly higher than the estimate of 7.8 W at thermoneutrality for an
oystercatcher weighing 0.52kg that Zwarts et al. (1996) based on a literature
survey. Yet, winter temperatures in the Wadden Sea are slightly lower than the
critical temperature of 10°C below which the costs of thermoregulation increase (Kersten and Piersma 1987). We assumed that the available foraging
time per day equals 12 h, the energy density of the bivalves is 22 MJ/kg ashfree dry mass (Beukema 1997) and the gut absorption efficiency is 0.8. The
power values are then equivalent to a required food intake rate while foraging
of 0.45 and 1.0 mg/s for knot and oystercatcher, respectively.
For each winter and summer season in the period 1973-1998 estimates of
production were obtained. For M. balthica average production in summer
(1.36g/m2, SD 0.528, CI 1.13-1.67) and in winter (1.03g/m2, SD 0.693, CI
0.71-1.26) were almost equal. In their first winter the zero-year class individuals are too small (about 6 mm) to be profitable as a food source for the two
bird species. The older year classes, those that can be eaten, were responsible
for about two-thirds of the total winter production (0.69 g/m2, SD 0.516, CI
0.49-0.85). For C. edule average production in winter was 6.53 g/m2 (SD 4.57,
CI 3.97-9.36) and in summer 4.64g/m2 (SD 4.78, CI 2.98-6.95). The zero-year
class, which is edible for knots, contributed on average about one-fifth
(1.31 g/m2, SD 1.34, CI 0.70-2.07) to total winter production. Production
varied considerably from year-to-year in both species (Fig. 15.2). The same
holds for the production/biomass ratio, particularly for the cockle (Fig. 15.3).