Does copepod size determine food consumption of particulate feeding fish?

Mikael van Deurs1, Marja Koski1 and Anna Rindorf1

1DTU Aqua National Institute of Aquatic Resources, Technical University of Denmark (DTU), Jægersborg Alle 1,

Charlottenlund Castle, 2920 Charlottenlund, Denmark.

ABSTRACT: Climate-induced reduction in mean copepod size, mainly driven by a decrease in

abundance of the large Calanus finmarchicus around 1987, has been linked to low survival of fish

larvae in the North Sea. However, to what extent this sort of reduction in copepod size has any

influence on adult particulate feeding fish is unknown. In the present study we investigated the

hypothesis that availability of the large copepods determines food consumption and growth

conditions of Lesser sandeel (Ammodytes marinus) in the North Sea. Analysis of stomach content

suggested that food consumption is higher for fish feeding on large copepods, and additional

calculations revealed how handling time limitation may provide part of the explanation for this

relationship. Comparing stomach data and zooplankton samples indicated that Lesser sandeel

actively target large copepods when these are available. Finally, we observed that the length of

Lesser sandeel began to decrease in the late 1980s, simultaneously with the C. finmarchicus decline.

Key words: predator-prey interactions; optimal foraging; prey selection; North Sea regime-shift;

climate change; sandeel; sand lance; forage fish; Calanus finmarchicus; handling time limitation

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

INTRODUCTION

Copepods play a key role as prey for pelagic fish larvae and zooplanktivorous fish. Hence, as a

changing climate modifies copepod communities (Planque and Taylor 1998; Möllmann et al. 2003;

Beaugrand 2004; Helaouët and Beaugrand 2009; Kjellerup et al. 2012), cascading effects through

the foodweb are likely to affect fish populations (Beaugrand et al. 2003; Möllmann et al. 2003;

Frederiksen et al. 2006; van Deurs et al. 2009).

Long-term copepod abundances in the North Sea show a decreasing trend for relatively large

Calanus finmarchicus in spring and an increasing one for the equally large C. helgolandicus in fall

(Planque & Fromentin 1996; Planque et al. 1997). The spring decline in abundance of C.

finmarchicus has been linked to low survival of cod larvae (Gadus morhua) and sandeel larvae

(Ammodytes marinus) (Beaugrand et al. 2003; van Deurs et al. 2009). Besides cod and sandeel

larvae, also juvenile and adult zooplanktivorous fishes in the North Sea feed on the calanoid

copepods (De Silva and Balbontin 1974; Last 1987; Maes et al. 2005; Raab et al. 2012), and the

literature report northward shifts in the feeding grounds of North Sea herring in response to changes

in the timing of C. finmarchicus (Corten 2000).

While C. finmarchicus has declined in abundance, overall annual production of calanoid copepods

is seemingly unchanged (Beaugrand et al. 2002) and the spring reduction in C. finmarchicus may

therefore to some extent have been compensated for by an increase in C. helgolandicus and smaller

calanoids. Although little attention has been dedicated to the long-term changes in abundance and

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

biomass of smaller calanoids, such as genus Temora, Pseudocalanus, Centropages, Paracalanus,

Oithona, Microsetella and Oncaea, these are all abundant in the North Sea (Nielsen et al. 1993).

Similarly to C. finmarchicus, some of these species also have their reproduction peak in spring

(Kiørboe & Nielsen 1994; Broekhuizen and McKenzie 1995).

The weights of calanoids in the North Sea are around four to six orders of magnitude smaller than

the weight of zooplanktivorous fishes. Consequently, for a particulate feeding fish that snatches one

prey at the time, handling time (here defined as the average time it takes to prepare attack, capture

and swallow each prey) may represent a potential bottleneck on food consumption when feeding at

high prey concentrations (Jeschke et al. 2002). However, since Calanus are as much as ten to

twenty times larger than the small calanoids mentioned above, the strength of handling time

limitation may be dependent on the species and size composition of copepods within in the prey

field.

Lesser sandeel is a particulate feeding zooplanktivorous fish throughout its life. They are abundant

in the Dogger area (central North Sea) on the southerly distribution limit of the large C.

finmarchicus, which suggests a particular susceptibility to fluctuations in the spring invasion of this

copepod species. The highly resident nature of sandeel renders them unable to move between areas

(Jensen et al. 2011) and hence they are likely to be more sensitive to changes in copepod abundance

than, for example, Atlantic herring that is known to redistribute according to food density

(Dragesund et al. 1997; Misund et al. 1998; Corten 2001). Furthermore, adult Lesser sandeel feed

mainly during spring (Macer 1966; Winslade 1974; MacLeod et al. 2007). From the point of view

of the sandeel, increasing C. helgolandicus concentrations in late summer are thus not likely to

compensate for a spring-time reductions in C. finmarchicus.

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

In the present study we examined the hypothesis that climate-driven reductions in copepod size

affect the growth of adult particulate feeding fish. If this hypothesis is valid, we would expect that

(i) fish on a diet of small copepods have diminished food consumption relative to fish on a diet of

large copepods, (ii) fish actively select for large copepods when these are available, and (iii) size at

age for particulate feeding fish in the North Sea have decreased simultaneously with the spring time

reduction in C. finmarchicus in the late 1980s. We investigated these assumptions by (1) using

stomach content of Lesser sandeel from the North Sea to analyze the relationship between prey size

and daily food consumption, (2) carrying out theoretical calculations to address the role of handling

time limitation, (3) comparing mesozooplankton data to the stomach content to shed light on prey

size preference and (4) investigating long-term trends in length at age for Lesser sandeel before and

after the late 1980s.

MATERIALS AND METHODS

Stomach analysis

Samples were collected from commercial fishing grounds in the Dogger area in 2006, 2007, 2009

and 2010 by fishermen. A total of 30 samples (containing between 16 and 86 sandeels each) were

used (Fig. 1). Each sample was taken by fishermen on commercial sandeel fishing vessels, frozen

on board the fishing vessel, and labeled with catch date, time of capture, and position. This enabled

sufficient spread in time and space and thereby a wide range of stomach contents. Fish from all 30

samples were measured (weight and length) and the stomachs were dissected, dapped on paper

towel, and weighed to the nearest mg. After weighing the stomachs, some were preserved in

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

alcohol. An overview of samples and number of alcohol preserved stomachs is presented in Table 1.

In order to get the wet weight of the stomach content, the weight of the empty stomach must be

subtracted from the total weight of the stomach. Therefore, 30 stomachs covering fish lengths (total

length) from 8 to 18 cm were emptied and the relationship between the wet weight of empty

stomachs (WES) [g] and fish length (FL) [cm] was established. A power law model provided the best

fit to data: WES = 9 × 10-5 × FL2.30 (R2 0.87).

The alcohol preserved stomach contents were spread evenly on Petri-dish with added water. A sub-

area of 4 cm2 in the middle of the petri-dish was photographed using Leica MZ12 stereo-microscope

with a 6.5 × magnification and a Leica DFC290 camera Image Pro Plus software was used to

determine the size of copepods by digitally measuring the length of all intact copepod prosomes

[mm ± 0.1]. Countable biomass constituted on average 45% of the stomach content, and of the

countable biomass 20% was available for counting in the 4 cm2 counting frame. A total of 4368

copepods were measured. Mean copepod size was calculated for each of the 30 samples giving each

stomach the same weight, independent of the number of measured prosomes. Due to the state of

degradation (e.g. lost appendixes and antennas) copepods could not be quantified per species.

In order to validate whether a 4 cm2 subarea provides a representative impression of the entire

stomach content, we spread out the content from 4 stomachs onto four Petri-dishes and divided each

dish into five zones of even size. Thereafter we measured all measurable prosomes within each of

the five zones and calculated the coefficients of variance. The average coefficient of variance was

0.056 (min = 0.032, max = 0.074).

The effect of copepod size on food consumption

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

Stomach analysis revealed that the variation in the size of ingested copepods was smaller within

samples than between samples. A bimodal distribution separated samples into two groups: those

with mean prosome length significantly above 1.3 mm and those with mean prosome length

significantly below 1.1 mm (see results). Based on this finding two feeding groups were defined:

sandeels from samples with mean copepod size <1.3 mm (group-1) and sandeels from samples with

mean copepod size >1.3 mm (group-2). We chose 1.3 mm over 1.1 mm as the threshold, since

among the common copepods in the North Sea only Calanus exceed 1.3 mm (Pitois et al. 2009)

plus it represented the local minimum when a density function was fitted to the bimodal

distribution.The effect of copepod size on consumption rate was subsequently investigated by

testing whether the mean consumption rate of sandeels belonging to group-1 differed significantly

from that of group-2. How this was tested is described below.

Assuming that gastric evacuation is an exponential process, we approximated the weight specific

consumption rate (C [g (g fish) -1 d-1]) for each fish using the following equation (e.g. Eggers 1977):

(1)

where S is the weight of the stomach content [g], W is the wet weight of the fish [g] and α(T) is the

gastric evacuation coefficient as a function of temperature (derived from van Deurs at al. 2010):

(2)

Modeled sea bottom temperature for each sample was provided by the Danish Meteorological

Institute based on predictions from the DMI-BSHcmod model (http://baltic.mersea.eu.org/). The

DMI-BSHcmod model is a 3 dimensional finite difference hydrodynamic model set up in spherical

coordinates. Calculations are done on an Arakawa C grid (Kleine 1994).

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

In order to estimate the effect of copepod size it was necessary also to account for other potential

predictors (i.e. fish size, year, season, latitude, longitude) and a random effect of sample. This was

achieved by estimating the parameters in a linear mixed effects model on the following form:

(3)

where and , , , , , and are parameters to be estimated, Lon is longitude, Lat

is latitude and indices CS, FS, y, s and sample denote copepod size group, fish size group, year,

season and sample, respectively. Copepod size was implemented as a class variable with two

discrete levels describing whether a fish belonged to group-1 or group-2. Fish size also had two

levels; sandeels <12 cm and sandeels >12 cm. The interaction effect between copepod size and fish

length was tested in order to determine whether the effect of copepod size depends on fish size.

Year and season had four levels (2006, 2007, 2009 and 2010) and two levels (April/May and June)

respectively. The p-values associated with parameter estimates for copepod size, fish size and the

interaction between copepod size and fish size were derived by comparing the full model with a

reduced model (likelihood ratio test).

Handling time limitation

To investigate if food consumption is potentially handling time limited, we calculated the average

time period available to handle (preparing attack, capture and swallow) each prey given a

hypothetical situation where search time is zero, all attacks are successful, and no time is spent on

other activities such as escaping predators. We refer to this time period as the maximum allowed

handling time per prey (MAH [s prey-1]). The assumption is that if MAH is sufficiently small,

handling time limitation is likely to affect food consumption. Mean MAH was calculated for each

fish as:

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

(4)

where C is derived from the equation (3), W is the wet weight of the fish [g], Pm is the copepod

individual wet weight [g], and τ is the daily foraging period [s]. Sandeels are visual feeders and

spend the night buried in the sand. The day length in May in the North Sea is roughly 15 hours.

However, as indicated by van Deurs et al. (2011) it takes at least one hour for the school to form in

the morning and another hour for the school to disintegrate after completing foraging. The τ-

parameter was therefore set to 46800 s (= 13 h). Calculations were made for each of the following

three scenarios of daily ration and copepod size: (A) Daily ration as observed in group-1 and

prosome length = 0.85 mm (equal to 0.000035 g), (B) Daily ration as observed in group-2 and

prosome length = 0.85 mm, and (C) Daily ration as observed in group-2 and prosome length = 1.75

mm (equal to 0.00044 g). Pm was derived from the length-weight relationships for Calanus and

Acartia; representing large and small copepods respectively (Uye 1982). For the dry weight to wet

weight conversion we applied a ratio of 5 adopted from Yamaguchi and Ikeda (2000).

Zooplankton samples

The inclusion of zooplankton samples was a post hoc decision taken based on the outcome of the

stomach analysis, showing that fish caught on a given position and date contained either small

copepods or large copepods, but never a mixture. Assuming that active selection for small copepods

is not an option; the simplest pair of mutually exclusive explanations is: (a) locations where

Calanus is common has relatively low concentrations of smaller calanoids (inverse correlation

between Calanus and smaller calanoids), or (b) sandeels avoid small calanoids when Calanus is

available (positive selection for larger copepods). The purpose of including zooplankton data was to

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

explicitly test explanation a. Since fish samples were collected onboard commercial fishing vessels,

parallel zooplankton samples could not be obtained. Instead we used a large mesozooplankton data-

set collected during the LIFECO program from late March to May in 2001 (Arendt et al. 2005) to

analyze the general covariation between small copepods (Acartia, Temora, Centropages,

Pseudocalanus, Oithona, Microcalanus, Oncaea, and Microsetella) and large copepods (Calanus).

Although no sampling stations were directly overlapping the stomach samples, all 62 samples used

overlapped with the continuum of sandeel habitat extending north-east from Dogger along the

southern boundary of the C. finmarchicus distribution (Fig. 1). Hence, they were found suitable for

testing the very general aspect of explanation a. Samples were typically taken from 50-70 m depth

to the surface (if the station was shallower, ca. 10 m above the bottom), using a submersible pump

connected to a 30 µm plankton net. The pump was towed vertically from 50-70 m to the surface

with an average rate of 15 cm s-1 and with a sampling volume of 1.3 m3 min-1. The samples were

preserved in 4% borax buffered formalin and counted from 1/4 to 1/64 fraction of the total sample

to achieve a total of >200 individuals for the dominant species (Jónasdóttir & Koski 2011). The

copepods were identified to the genus or species level, and the life-stages were grouped to nauplii,

copepodites (I-V), females and males, as well as placed into size categories at 50 µm intervals. The

average prosome length of each species was estimated for 5 stations per cruise combining the

copepodites and adults (not nauplii) by calculating the weighted mean length of each species. The

average length was converted to wet weight using length-weight conversions from the literature

(Cohen and Lough 1981; Breteler et al. 1982; Uye 1982) and a wet weight/dry weight ratio of 5 to

convert dry weight to wet weight (Yamaguchi and Ikeda 2000). In most cases length-weight

conversions at the genus level could be found in the literature. Exceptions were Microcalanus and

Oncaea, where the length-weight conversion for Oithona was applied. Lastly, we multiplied

weights and numbers and thereby derived the wet biomass [g m-2] of small and large copepods for

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

each sampling station. The spearman correlation was applied to test if the concentration of Calanus

was related to the concentration of small copepod species or vice versa

Trends in length at age of sandeels over time

Lesser sandeel time series of mean total length at age 1 and age 2 were analyzed to investigate if

periods of changing growth rates could be detected. In particular, we investigated whether the

growth rate of Lesser sandeel in the Dogger area slowed down around 1987, the year of the reported

change in the North Sea copepod community (Beaugrand et al. 2004). The time-series represent

length at age in June (1976-2011) from commercial samples.

A linear mixed effect model was applied to test if the period after 1987 differed from the period

prior to 1987 in terms of length at age:

where and and are constants within indices to be estimated in the model. With this

formulation, slope and intercept of a linear trend line are estimated for each combination of age and

period. The linear mixed effect model allowed us to account for correlation between length at age 1

and age 2 in a given year through the addition of a random effect . Finally, the effect of Period

was tested by comparing the presented model with a version without Period. Length at age trends

were held up against development in total stock biomass and fishing mortality going back to 1983

(ICES 2012), to allow us to discuss potential alternative explanations, such as density dependence

and fishing induced evolution of growth rates (Olsen et al. 2004).

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

RESULTS

Stomach analysis: Variation in the size of ingested copepods was much smaller within samples

than between samples. When samples was aligned and sorted according to mean prosome length a

bimodal distribution was revealed, that separated samples into two groups, those with mean

prosome length significantly above 1.3 mm and those with mean prosome length significantly

below 1.1 mm (Fig. 2, the use of “significantly” was based on the standard errors associated with

each sample). Only one sample (id 30) fell in between 1.1 and 1.3 mm. Each sample was

represented by only a subset of stomachs (Table 1). However, mean prosome length between

stomachs from the same sample varied little: For 88% of all stomachs analyzed, it applied that the

mean prosome length of copepods in a stomach was on the same side of 1.3 mm as the combined

mean prosome length of all stomachs from the sample to which this stomach belonged.

Furthermore, the coefficient of variance of prosome lengths within a given stomach was on average

0.22. Together these findings indicate that sandeels from the same sample had been foraging on the

same narrow size range of copepods, either small (<1.3 mm) or large (>1.3 mm) copepods. It

should, however, be noted that a statistically significant relationship existed between fish size and

prey size at the level of the sample (p < 0.01, mixed effect model with random sample effects). The

estimated coefficient of this relationship was 0.027, corresponding to an increase in prey size of

0.027 mm for each 1 cm increase of the fish.

The effect of copepod size on food consumption: Sandeels feeding on large

copepods (group-2) displayed on average 130% (for small sandeels) and 100% (for large sandeels)

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

higher stomach content weight compared to sandeels feeding on smaller copepods (group-1). This

pattern was unchanged after converting stomach content to weight specific food consumption (Fig.

3). Weight specific food consumption of group-1 sandeel was significantly lower than that of

group-2 sandeel (p = 0.001; Gaussian criteria of the model was satisfied and residuals were without

trend). This pattern was consistent among sampling years with an average consumption rate of

~0.010 (g g-1 fish) for group 1 and ~0.022 (g g-1 fish) for group-2 (Table 2). Further, weight-specific

consumption of large sandeels was significantly lower than the weight-specific consumption of

small sandeels (p = 0.04), but the interaction-effect between copepod size and fish length was

insignificant. In addition, the food consumption was influenced both by longitude and latitude (p =

0.01 and p = 0.04, respectively), whereas the season and year had no effect (p > 0.05).

Handling time limitation: Mean maximum allowed handling time (MAH) was approx. 10 seconds

for large sandeels in scenario A and for both small and large sandeels in scenario B. For small

sandeels in scenario A MAH was approx. 30 seconds, whereas it was well above 100 seconds for

both small and large sandeels in scenario C (Fig. 4).

Zooplankton data: Zooplankton samples from 2001 revealed that Calanus concentration within

sandeel feeding grounds was positively correlated to the concentration of smaller copepods (Fig. 5),

indicating that sandeels experience a prey field containing both small and large copepods. On all

sampling stations smaller copepods exceeded Calanus in terms of numbers, whereas the biomasses

were in the same order of magnitude.

Length at age: Length at age time-series for age 1 and age 2 Lesser sandeel in the Dogger area

indicated a turning-point around 1987, with increasing length at age in the period before 1987 and a

gradual decrease in the period after 1987 (Fig. 6). The effect of Period was significant (p = 0.02)

and data satisfied model assumptions as indicated by inspection of residuals.

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

Total stock biomass displayed large fluctuations between 1983 and 2002, but with no clear temporal

trend or apparent turning point around 1987. In 2003 stock biomass dropped to the lowest level

experienced in the entire time series followed by an increasing trend. The long-term pattern in

fishing mortality was also seemingly disconnected from the overall patterns in length at age.

Fishing mortality exceeded 1 in the period 1999 to 2004, for the remaining of the time period (going

back to 1983) fishing mortality fluctuated around a stable level between 0.5 and 0.8.

DISCUSSION

Feeding on larger copepods doubled the food consumption in lesser sandeel in the Dogger area.

Further calculations indicated that in a scenario where only small copepods are available, the

individual sandeel have on average roughly 10 seconds (one order of magnitude lower than for

large copepods) available for handling each prey item to achieve a daily ration equal to that

observed for sandeels feeding on large copepods (group-2). The upper limit of realistic handling

times reported for similar sized zooplanktivorous fish is also 10 seconds (Hjelm and Persson 2001),

indicating that very little time is left in this scenario for prey searching, unsuccessful attacks, and

escaping predators (either by burying into the sand or by forming temporary denser schools). In

consequence this suggests that consumption rate is inclined to becoming handling time limited

when only small copepods are available.

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

According to the strict conceptual understanding of handling time limitation (Jeschke et al. 2002)

only predators that digest prey items faster than they are handled are likely to be handling time

limited. From the evacuation rate applied in the present study (around 5% per hour) (van Deurs et

al. 2010) and a handling time of maximum 10 seconds it follows that sandeels handle prey faster

than they are digested. Hence, sandeels may not be handling time limited in a strict sense. However,

the longer sandeels spend foraging away from their sand refuge, the higher the risk of predation.

This leads to the prediction that as prey size decreases (and the reward per unit handling time

decreases) the sandeel thus become increasingly inclined to choose between exposing itself to a

high level of predation risk or consent to retreat to their sand refuge with less in their stomach (the

energy-predation trade off, see van Deurs et al. (2010)).

The weight specific daily ration of large sandeels was significantly less than for small sandeels.

This is in line with our calculations of maximum allowed handling time, which indicated that large

sandeels are more likely (lower MAH) to be handling time limited than small sandeels. However,

the interaction-effect between fish size and prey size on weight specific consumption was not

significant, which would be reasonable to expect in a situation where handling time constraints is

intensifying with fish size.

C. finmarchicus is the only abundant zooplankton genus in the North Sea in spring that reaches

prosome lengths >1.3 mm (Pitois et al. 2009). We therefore suggest that sandeels with mean food

item prosome length of >1.3 mm (group-2) were feeding mainly on Calanus by the time they were

caught. The springtime mesozooplankton data revealed that availability of Calanus is positively

correlated to the availability of smaller copepods, yet stomachs from sandeels of all sizes, contained

either small or large copepods, but never a mix. The simplest explanation for this observation is that

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

sandeels deliberately ignore smaller copepods when larger copepods are sufficiently abundant,

which is also in line with traditional optimal foraging theory stating that an optimal forager rank

prey preference according to energy gain per handling time (Stephens and Krebs 1986; Hart and

Ison 1991). However, this conclusion requires that sandeels containing only small copepods were

experiencing a prey field nearly deprived of Calanus (we will return to this later in the discussion).

Acting to blur the picture, the stomach content of the sandeels may also reflect the copepod

assemblage in its horizontal or vertical position. The zooplankton samples used in the present study

did not originate from the same location or date as the sandeel catch, and were integrated over the

entire water column. The relationship between large and small copepods could thus potentially be

different between areas, seasons or depths. Copepods are known to be patchily distributed, occupy

different layers of the water column and have different vertical migration behavior (Daro 1988). For

instance, in some areas of the North Sea small copepods with limited reserves have been observed

to mainly reside in the proximity of the chlorophyll maximum layer (Daro 1988; Koski et al. 2011),

while C. finmarchicus migrates between different layers (Daro 1988) or resides deeper down in the

water column (Jónasdóttir and Koski 2011). Although typically one would expect both small and

large copepods to be abundant at similar productive areas, we cannot exclude the possibility that

sandeel stomachs reflect the prey availability at their exact location rather than selective feeding on

larger copepods. Nevertheless, the presence of exclusively large or small copepods in the stomachs

suggest some degree of selective feeding, and according to optimal foraging theory (i.e. Stephens

and Krebs 1986; Hart and Ison 1991) it would be unexpected if that selective feeding would mainly

focus on small prey.

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

Why did some stomachs contain only small copepod species? Access to large copepods may be

determined by the overlap between sandeel habitats and the distribution of C. finmarchicus. C.

finmarchicus protrudes into the North Sea from the north during spring. The distribution finds a

southern boundary at latitude 55 to 56oN (Planque and Fromentin 1996) and thereby partially

overlaps with the important sandeel areas on Dogger Bank (Fig. 1). However, while the exact

position of the front in that area is dynamic and change over time (Munk et al. 1999), sandeels are

bound to their habitats (Jensen et al. 2011). Consequently, variation in the degree of overlap could

be an important factor affecting access to larger copepods, and may act independent of the overall

climatic changes and the general C. finmarchicus distribution in the North East Atlantic.

The time-series analysis indicated changing growth rates for Lesser sandeel in the Dogger area

around 1987, where also C. finmarchicus began to decline (Beaugrand et al. 2004). However, this

change was not characterized by a shift from a high plateau to a low plateau, as reported by

Beaugrand et al. (2003) for the abundance of C. finmarchicus. Instead it conformed to a turning-

point with gradually increasing length at age prior to 1987 and declining length at age in the period

after, consistent with reports for Lesser sandeel off the Scottish coast (Wanless et al. 2005;

Frederiksen et al. 2011). The gradual decline in length at age corresponds to an earlier CPR time-

series analysis (Planque and Fromentin 1996), showing a similar gradual decline in C. finmarchicus

after the mid-1980s rather than an abrupt change as suggested by Beaugrand et al. (2004). However,

the gradual increase in length at age prior to the mid-1980s is not represented in the trends in C.

finmarchicus development. It should however be noted that the CPR survey does not cover small-

scale changes in vertical distribution of different copepod species: for instance, C. finmarchicus

often resides much deeper than the CPR sampling depth (Jónasdóttir and Koski 2011), It is thus

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

unclear how evident long-term trends could one expect to observe between different trophic levels

occupying different and varying water layers using the present day sampling strategies.

In addition to the connection to Calanus finmarchicus, we searched for alternative explanations for

the changes in sandeel size. Based on long term patterns in total stock biomass and fishing

mortality, we investigated the role of density dependence and fishery induced evolution as drivers

of the observed decreasing length at age (Fig. 6). We found no reason to believe that either of these

were the responsible factor. There were, however, indications that a period of increasing length

after 2003 may be explained by relaxed food competition (density dependence) after a pronounced

stock decline in 2003.

Conclusion: (1) Food consumption is higher for Lesser sandeel dieting on larger copepods

compared to smaller copepods. (2) Handling time limitation provides a plausible causal explanation.

(3) Lesser sandeel actively target large copepods. (4) Length at age of Lesser sandeel began to

decrease in the late 1980s, simultaneously with the C. finmarchicus decline. Although, during other

periods sandeel growth-trends are inconsistent with patterns of C. finmarchicus concentrations

reported in the literature. Overall we found no strong arguments for rejecting the hypothesis that

climate driven reductions in copepod size will lead to diminished growth rates of Lesser sandeel.

In the present study we focused on food consumption measured as biomass. Energy density of

Calanus may on certain times of the year be double that of smaller short-lived non-overwintering

copepods in the North Sea (i.e. Corner and O´Hara 1986). Certainly a difference this large in food

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

quality will most likely also affect fish growth. However, the seasonal dynamics in lipid content and

energy density of relevant North Sea copepod species, experiencing the same environmental

conditions, is currently not available in the scientific literature. We suggest that a food quality study

represents a natural next step from the study presented here.

ACKNOWLEDGEMENT

This research has been supported by EU FP7 grant FACTS (Forage Fish Interactions), Grant

Agreement No. 244966. We like to acknowledge DMI and people (in particular Peter Munk) behind

the LIFECO data for providing modelled temperature data and mesozooplankton samples. We also

like to thank Peter Munk for reading and constructively commenting on the manuscript.

REFERENCES

Arendt KE, Jonasdottir SH, Hansen PJ, Gartner S (2005) Effects of dietary fatty acids on the reproductive success of the calanoid copepod Temora longicornis. Mar Biol 146:513-530

Beaugrand G (2004) The North Sea regime shift: evidence, causes, mechanisms and consequences. Prog Oceanogr 60:245-262

Beaugrand G, Brander KM, Lindley JA, Souissi S, Reid PC (2003) Plankton effect on cod recruitment in the North Sea. Nature 426:661-664

Beaugrand G, Ibanez F, Lindley JA, Reid PC (2002) Diversity of calanoid copepods in the North Atlantic and adjacent seas: species associations and biogeography. Marine Ecology-Progress Series 232:179-195

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398399

400401

402403

404405406

Breteler WCM, Fransz HG, Gonzalez SR (1982) Growth and development of four calanoid copepod species under experimental and natural conditions. Netherland J Sea Res 16:195-207

Broekhuizen N and McKenzie E (1995) Patterns of abundance for Calanus and smaller copepods in the north-sea - time-series decomposition of 2 cpr data sets. Marine Ecology-Progress Series 118:103-120

Cohen RE, Lough RG (1981) Length-Weight Relationships for Several Copepods Dominant in the Georges Bank-Gulf of Maine Area. J Northw Atl Fish Sci 2:47-52

Corner EDS and O´Hara SCM. (1986) The biological chemistry of marine copepods. Clarendon press, Oxford

Corten A (2000) A possible adaptation of herring feeding migrations to a change in timing of the Calanus finmarchicus season in the eastern North Sea. ICES J Mar Sci 57:1261-1270

Daro MH (1988) Migratory and Grazing Behavior of Copepods and Vertical-Distribution of Phytoplankton. Bull Mar Sci 43:710-729

De Silva SS, Balbontin F (1974) Laboratory studies on food intake, growth and food conversion of young herring, Clupea harengus (L.). J Fish Biol 6:645-658

Dragesund O, Johannessen A, Ulltang O (1997) Variation in migration and abundance of norwegian spring spawning herring (clupea harengus L.). Sarsia 82:97-105

Frederiksen M, Edwards M, Richardson AJ, Halliday NC, Wanless S (2006) From plankton to top predators: bottom-up control of a marine food web across four trophic levels. J Anim Ecol 75:1259-1268

Frederiksen M, Elston DA, Edwards M, Mann AD, Wanless S (2011) Mechanisms of long-term decline in size of lesser sandeels in the North Sea explored using a growth and phenology model. Mar Ecol Prog Ser 432:137-147

Hart P J.B, Ison S (1991). The influence of prey size and abundance, and individual phenotype on prey choice by the three-spined stickleback, Gasterosteus aculeatus L.. J Fish Biol 38: 359-372 

Helaouët, P., Beaugrand, G., 2009. Physiology, ecological niches and species distribution. Ecosystems 12, 1235 - 1245.

Hjelm J, Persson L (2001) Size-dependent attack rate and handling capacity: inter-cohort competition in a zooplanktivorous fish. Oikos 95:520-532

ICES (2012) ICES WGNSSK Working Group on the assessment of demersal stocks in the North Sea and Skagerrak (www.ices.dk)

407408

409410411

412413

414415

416417

418419

420421

422423

424425426

427428429

430

431432433

434435

436437

438439

Jensen H, Rindorf A, Wright PJ, Mosegaard H (2011) Inferring the location and scale of mixing between habitat areas of lesser sandeel through information from the fishery. ICES J Mar Sci 68:43-51

Jeschke JM, Kopp M, Tollrian R (2002) Predator functional responses: Discriminating between handling and digesting prey. Ecol Monogr 72:95-112

Jonasdottir SH, Koski M (2011) Biological processes in the North Sea: comparison of Calanus helgolandicus and Calanus finmarchicus vertical distribution and production. J Plankton Res 33:85-103

Kjellerup S, Dunweber M, Swalethorp R, Nielsen TG, Moller EF, Markager S, Hansen BW (2012) Effects of a future warmer ocean on the coexisting copepods Calanus finmarchicus and C-glacialis in Disko Bay, western Greenland. Marine Ecology-Progress Series 447:87-U132

Koski M, Jonasdottir SH, Bagoien E (2011) Biological processes in the North Sea: vertical distribution and reproduction of neritic copepods in relation to environmental factors. J Plankton Res 33:63-84

Last JM (1987) The food of immature sprat (Sprattus sprattus (L.)) and herring (Clupea harengus L.) in coastal waters of the North Sea. Journal du Conseil - Conseil International pour l'Exploration de la Mer 44:73-79

Macer CT (1966) Sandeels (Ammodytidae) in the south-western North Sea: Their biology and fishery. MAFF Fishery Invest London ser II 24:1-55

MacLeod CD, Santos MB, Reid RJ, Scott BE, Pierce GJ (2007) Linking sandeel consumption and the likelihood of starvation in harbour porpoises in the Scottish north sea: Could climate change mean more starving porpoises? Biology Letters 3:185-188

Maes J, Tackx M, Soetaert K (2005) The predation impact of juvenile herring Clupea harengus and sprat Sprattus sprattus on estuarine zooplankton. Hydrobiologia 540:225-235

McKenzie N, Broekhuizen E (1995) Patterns of abundance for Calanus and smaller copepods in the North Sea: time series decomposition of two CPR data sets. Mar Ecol Prog Ser 118: 103-120

Misund O, Vilhjalmsson H, Jakupsstovu S, Rottingen I, Belikov S, Asthorsson O, Blindheim J, Jonsson J, Krysov A, Malmberg S, and others (1998) Distribution, migration and abundance of norwegian spring spawning herring in relation to the temperature and zooplankton biomass in the norwegian sea as recorded by coordinated surveys in spring and summer 1996. Sarsia 83:117-127

Moellmann C, Kornilovs G, Fetter M, Koester FW, Hinrichsen H- (2003) The marine copepod, Pseudocalanus elongatus, as a mediator between climate variability and fisheries in the Central Baltic Sea. Fish Oceanogr 12:360-368

440441442

443444

445446447

448449450

451452453

454455456

457458

459460461

462463

464465

466467468469470

471472473

Munk P, Larsson PO, Danielssen DS, Moksness E (1999) Variability in frontal zone formation and distribution of gadoid fish larvae at the shelf break in the northeastern North Sea. Marine Ecology-Progress Series 177:221-233

Nielsen TG, Lokkegaard B, Richardson K, Pedersen FB, Hansen L (1993) Structure of Plankton Communities in the Dogger Bank Area (North-Sea) during a Stratified Situation. Marine Ecology-Progress Series 95:115-131

Olsen EM, Heino M, Lilly GR, Morgan MJ, Brattey J, Ernande B, Dieckmann U (2004) Maturation trends indicative of rapid evolution preceded the collapse of northern cod. Nature 428:932-935

Pitois SG, Shaw M, Fox CJ, Frid CLJ (2009) A new fine-mesh zooplankton time series from the dove sampling station (north sea). J Plankton Res 31:337-343

Planque B, Fromentin JM (1996) Calanus and environment in the eastern North Atlantic .1. Spatial and temporal patterns of C-finmarchicus and C-helgolandicus. Marine Ecology-Progress Series 134:101-109

Planque B, Hays GC, Ibanez F, Gamble JC (1997) Large scale spatial variation in the seasonal abundance of Calanus finmarchicus. Deep-Sea Res I 44:315-326

Planque and taylor (1998) Long-term changes in zooplankton and the climate of the North Atlantic. ICES J Mar Sci 55:644-654

Raab K, Nagelkerke LAJ, Boerée C, Rijnsdorp AD, Temming A, Dickey-Collas M (2012) Dietary overlap between the potential competitors herring, sprat and anchovy in the North Sea. Marine Ecology Progress Series: In press DOI:10.3354/meps09919

Stephens DW, Krebs JR (1986). Foraging theory. Princeton University Press (1986), pp. 13-24

 Uye S (1982) Length-weight relationships of important zooplankton from the inland sea Japan. Journal of the Oceanographical Society of Japan 38:149-158

van Deurs M, Christensen A, Frisk C, Mosegaard H (2010) Overwintering strategy of sandeel ecotypes from an energy/predation trade-off perspective. Marine Ecology-Progress Series 416:201-U215

van Deurs M, van Hal R, Tomczak MT, Jonasdottir SH, Dolmer P (2009) Recruitment of lesser sandeel Ammodytes marinus in relation to density dependence and zooplankton composition. Marine Ecology-Progress Series 381:249-258

van Deurs M, Behrens JW, Warnar T, Steffensen JF (2011) Primary versus secondary drivers of foraging activity in sandeel schools (ammodytes tobianus). Mar Biol 158:1781-1789

Wanless S, Harris MP, Redman P, Speakman JR (2005) Low energy values of fish as a probable cause of a major seabird breeding failure in the North Sea. Marine Ecology-Progress Series 294:1-8

474475476

477478479

480481

482483

484485486

487488

489490

491492493

494495

496497

498499500

501502503

504505

506507

Winslade P (1974) Behavioral-Studies on Lesser Sandeel Ammodytes-Marinus (Raitt) .3. Effect of Temperature on Activity and Environmental-Control of Annual Cycle of Activity. J Fish Biol 6:587-599

Yamaguchi A, Ikeda T (2000) Vertical distribution, life cycle, and developmental characteristics of the mesopelagic calanoid copepod Gaidius variabilis (Aetideidae) in the Oyashio region, western North Pacific Ocean. Mar Biol 137:99-109

508509510

511512513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

Number of fishDate Number of included in the

Sample ID dd/mm/yy Latitude Longitude fish per sample diet analysis

1 18/04/2006 54.633 1.183 19 32 21/04/2006 54.584 1.264 22 43 22/04/2006 55.030 1.337 24 54 23/04/2006 54.617 1.200 24 35 23/04/2006 55.317 1.648 21 36 28/05/2006 54.533 1.283 22 37 28/05/2006 54.327 0.872 24 38 29/05/2006 54.327 0.872 23 39 29/05/2006 54.327 0.872 26 210 30/05/2006 55.317 1.648 24 411 31/05/2006 54.584 1.264 19 212 01/06/2006 54.584 1.264 29 213 03/06/2006 54.584 1.264 29 714 04/06/2006 54.145 2.044 22 315 06/06/2006 54.145 2.044 22 416 10/06/2006 54.145 2.044 21 417 10/06/2006 53.907 1.926 24 418 14/06/2006 53.907 1.926 24 419 14/06/2006 54.195 1.529 29 320 16/06/2006 54.277 0.735 35 221 14/04/2007 55.350 1.300 34 422 14/06/2007 55.400 1.300 29 723 22/05/2009 54.200 2.000 66 724 24/05/2009 53.500 1.000 68 425 30/05/2009 54.100 0.900 24 1526 07/04/2010 54.567 1.217 25 2227 12/04/2010 55.317 0.883 26 1028 18/04/2010 55.347 1.512 17 1329 22/05/2010 55.100 1.350 35 1530 07/06/2010 55.067 1.083 16 16

Total 823 181

Table 1. Overview of sandeel samples.

529

530

531

532

           2006 2007 2009 2010Small copepods in stomachs (group-1) 0.011 no data 0.009 0.011

(± 0.0011) (± 0.0009) (± 0.0009)Large copepods in stomachs (group-2) 0.022 0.023 no data no data  (± 0.0017) (± 0.0029)    

Table 2. Geometric mean relative consumption C (g d-1 g-1 fish) by year and copepod size category

(± standard error).

533

534

535

536

Figure 1. Study area. Sandeel habitats in relation to the inflow (arrows) and southerly boarder of

the Calanus finmarchicus protrusion into the North Sea (solid line). The square window defines the

Dogger area. Sandeel habitats were re-produced from (Christensen et al. 2009) and the southerly

boarder of C. finmarchicus was adopted from Planque and Fromentin (1996). Stars represent the

positions of LIFECO copepod samples (n = 62).

537

538

539

540

541

542

543

544

Figure 2. Mean prosome length of copepods found in the sandeel stomachs calculated separately

for each sample. Samples are sorted according to the mean prosome length. Solid whiskers

represent standard errors and dashed whiskers represent standard deviations. Sample identifiers

correspond to sample ID in Table 1. The horizontal dashed line represents the threshold used to

divide samples into group-1 and group-2.

545

546

547

548

549

550

551

552

Figure 3. The effect of copepod size (prosome length) on consumption rate C [g (g fish) -1 d-1].

Graph displays the geometric mean food consumption (+/- s.e.) for small copepods (group-1) and

large copepods (group-2) and for small (black) and large (white) sandeels. n = 173, 293, 183, and

174 for group-1/small sandeels, group-1/large sandeels, group-2/small sandeels, and group-2/large

sandeels respectively. Mean size of small sandeels: 10.65 cm +/- 1.26 (s.d.) and 3.3 g; and large

sandeels: 15.14 cm +/- 1.28 (s.d.) and 10.3 g.

553

554

555

556

557

558

559

560

Figure 4. The effect of copepod size on maximum allowed handling time MAH [s prey-1]. Graph

displays the geometric mean MAH assuming a daily effective foraging period τ of 13 hours (± 25%)

for each of three scenarios of daily ration and copepod size. A: Daily ration as observed in group-1

(see Fig. 3) and prosome length = 0.85 mm, B: Daily ration as observed in group-2 (see Fig. 3) and

prosome length = 0.85 mm, and C: Daily ration as observed in group-2 and prosome length = 1.75

mm. Black symbols represent small sandeels (< 12 cm) and white symbols represent large sandeels

(> 12 cm). Horizontal dashed line defines the upper limit of realistic handling times reported for

zooplanktivorous fish of similar size (e.g. Hjelm and Persson 2001).

561

562

563

564

565

566

567

568

569

Figure 5. Co-occurrence of small and large copepods in the vicinity of North Sea sandeel areas.

Graphs show the concentration of Calanus versus the concentration of small copepods (A: numbers

m-2; B: g m-2) and each point represents a given sampling station. Small copepods include Acartia,

Temora, Centropages, Pseudocalanus, Oithona, Microcalanus, Oncaea, and Microsetella. p-values

were derived using the spearman correlation test.

570

571

572

573

574

575

576

577

Figure 6. Trends in total length at age for Lesser sandeel in the Dogger area in relation to the

regime shift in the late 1980s, total stock biomass and fishing mortality. Upper graph displays

length at age for age 1 (grey) and age 2 (black) together with trend lines (dashed lines) fitted to the

period before and after 1987. Lower graph depicts total stock biomass (bars) and fishing mortality

(lines) back to 1983 as provided by the ICES working group on Lesser sandeel in the North Sea.

578

579

580

581

582

583

584

585

586

587

588

589