Larval fish diversity on the Scotian Shelf - Canadian Science Publishing

Color profile: Disabled Composite Default screen

1747

Larval fish diversity on the Scotian Shelf N.L. Shackell and K.T. Frank

Abstract: We examined larval fish diversity on the Scotian Shelf using data, representing 91 genera, collected during the Scotian Shelf Ichthyoplankton Program from 1978 to 1982. Two diversity indices (genus richness (GR) and Shannon’s entropy (H)) were relatively lower from December to February–March and relatively higher and stable from April to September–October. Taxon composition changed seasonally. Total median log abundance (log10(number of individuals + 1)·1000 m–3) was low from December to February, increased in March, was stable from April to June, and declined from July to October. Our results suggest that the abundance trends of most taxa were not coincident with either a spring or fall bloom of calanoid copepods. Log GR was significantly positively related to H (r = 0.62, p < 0.001, n = 1853). A negative exponential best described the relationship between log GR and log abundance (R 2 = 0.77; log GR = 1.37(1 – e–(1.13)(log abundance)), p < 0.001, n = 2357). Shannon’s H was not related to log abundance in winter or in summer–fall and was negatively correlated in spring–summer (r = –0.12, p = 0.003, n = 593). Thus, diversity increases with abundance but the composition is dominated by relatively fewer genera at higher levels of abundance. Western – Sable Island banks had higher levels of GR and abundance in all seasons. Additional banks were diverse and productive during warmer months. Résumé : Nous avons étudié la diversité des larves de poisson sur la plate-forme néo-écossaise avec des données couvrant 91 genres et recueillies dans le cadre du programme d'étude de l'ichthyoplancton de cette plate-forme de 1978 à 1982. Deux indices de diversité (diversité des genres (GR, pour «genus richness») et entropie de Shannon (H)) étaient relativement plus bas de décembre à février–mars et relativement plus élevés et stable d'avril à septembre–octobre. La composition par taxons changeaient saisonnièrement. Le logarithme de l'abondance totale médiane (log10(nombre d'individus + 1)·1000 m–3) était bas de décembre à février, augmentait en mars, était stable d'avril à juin, et diminuait de juillet à octobre. Nos résultats laissent entendre que les tendances des abondances de la plupart des taxons n'étaient pas liées à une prolifération printanière ou automnale de copépodes calanoïdes. Le paramètre log GR était de façon significative positivement corrélé avec H (r = 0,62, p < 0,001, n = 1853). Une exponentielle négative décrivait le mieux la relation entre log GR et le logarithme de l'abondance (R2 = 0,77; log GR = 1,37(1 – e–(1,13)(log abondance)), p < 0,001, n = 2357). Le paramètre H de Shannon n'était pas relié au logarithme de l'abondance en hiver ou dans la période été– automne et était corrélé négativement dans la période printemps–été (r = –0,12, p = 0,003, n = 593). Ainsi, la diversité s'accroît avec l'abondance, mais la composition est dominée par un nombre relativement plus faible de genres quand les abondances sont plus élevées. Le banc occidental et le banc de l'île de Sable présentent une diversité de genres et une abondance plus élevées en toutes saisons. D'autres bancs montraient une diversité et une productivité importantes durant les mois les plus chauds. [Traduit par la Rédaction]

Shackell and Frank

Introduction The conservation of biological diversity has become a primary objective in various policy frameworks around the world, specifically those designed to incorporate an ecological approach to resource management (Grumbine 1997). The implementation of such an objective is proving problematic, in part because the patterns of biological diversity are not always known. The mechanisms or processes that maintain or cause changes in biological diversity, which can occur at a variety of scales, are even less clear. In this paper, we characterize broad-scale patterns of larval fish diversity and Received September 23, 1999. Accepted April 19, 2000. J15370 N.L. Shackell1 and K.T. Frank. Department of Fisheries and Oceans, Marine Fish Division, Bedford Institute of Oceanography, P.O. Box 1006, Dartmouth, NS B2Y 4A2, Canada. 1

Author to whom all correspondence should be addressed. e-mail: [email protected]

Can. J. Fish. Aquat. Sci. 57: 1747–1760 (2000)

J:\cjfas\cjfas57\cjfas-09\F00-111.vp Friday, August 18, 2000 8:44:16 AM

1760

abundance as an initial contribution towards discerning how such patterns are created, maintained, or changed on the Scotian Shelf. Processes responsible for patterns of diversity can occur at a variety of scales. We explore whether the observed patterns of diversity were related to broad-scale patterns of zooplankton abundance and (or) the climatological circulation. Larval fish assemblages have been related to hydrographic and physical oceanographic features in a wide range of systems (e.g., Frank and Leggett 1983; Cowen et al. 1993; Moser and Smith 1993). Larval assemblages have been further explained as a result of convergent spawning strategies among members of a given assemblage (e.g., Sherman et al. 1984; Doyle et al. 1993). Iles and Sinclair (1982) proposed that spawning has evolved to occur in areas that would enhance retention and that retention was positively related to survival (also see Werner et al. 1996; Shackell et al. 1999). In a review of life history strategies, Winemiller and Rose (1992) theorized that the response to seasonality in temperate zones is to spawn during a time when favorable conditions exist for survival. Cushing’s related “match–mismatch” hypothe© 2000 NRC Canada


1748

sis states that fish have evolved to spawn so that the peak in post-yolk-sac larval abundance coincides with a peak in prey availability (Cushing 1975, 1995). Although larval distribution may be a result of convergent spawning strategies, it is equally plausible that diverse strategies could result in the same distribution because certain areas or times favor survival. The common predictions of these varied hypotheses are that larvae that coincide with high food abundance and (or) occur in retention areas have a greater probability of survival. Our study gave us the opportunity to assess the generality of these predictions. If these hypotheses have general application, we would expect that larval fish diversity is greatest when and where physical retention (e.g., gyres) is highest and (or) when prey abundance is greatest.

Materials and methods Scotian Shelf Ichthyoplankton Program (SSIP) sampling protocol The data used in this study were collected on the Scotian Shelf from 1978 to 1982 as part of the SSIP (Fig. 1; Table 1). A primary goal of the SSIP was to collect temporal and spatial information about fish eggs and larvae and associated environmental data across the Scotian Shelf. A related goal of the SSIP was to characterize the distribution, abundance, mortality, and growth rates of various ichythyoplankton species. The survey design was based on standard sampling along a transect grid. Sampling protocol included standardization of various methods. An oblique Bongo tow was used typically with a 333-mm-mesh net deployed for a variable time, but most typically 30 min. Flowmeters were attached to the Bongo net to determine total volume filtered. Surface temperature and bottom depth were recorded for each sample. Bottom depth represented the distance from the surface to the bottom, whereas sampling depth represented the depth at which the sample was taken. Sampling tow depths ranged from 23 to 247 m, averaging 112 m, and the ranges of sampling depth were similar among months. Samples were stored in 5% formalin and then were identified and catalogued at the Huntsman Marine Laboratory, St. Andrew’s, New Brunswick, Canada. Further information on the SSIP survey design and a physiographic description of the Scotian Shelf are available in O’Boyle et al. (1984). We extracted only larval stage data from the original SSIP data set. Initially, our intent was to conduct a species-level analysis, but we eventually opted to conduct a genus-level analysis because of several limitations of the data. First, the identification to species level within some of the more common species, e.g., Sebastes, is difficult during the larval stage. It is also true that the taxonomic description of many of the common species is better developed than that of rare species, e.g., Melanogrammus aeglefinus and Gadus morhua, whereas rarer species would be identified only to higher taxonomic levels. For each genus, we itemized the possible number of species known to occur in the Scotian Shelf region (Scott and Scott 1988) and concluded that the majority of genera studied herein contain one to three species (Table 2). Genus diversity is, to some extent, representative of species diversity in our study because of the small number of species within many of the genera. Finally, in the extracted data set, 31.2% of the observations had been identified to the generic level, 58.9% to the species level, and 3.6% to subfamily and higher levels and 6.3% were unidentified. A genus-level analysis allowed us to use the majority (90.1%) of the data.

Can. J. Fish. Aquat. Sci. Vol. 57, 2000

Diversity and abundance indices We examined two measures of diversity, genus richness (GR) and Shannon’s entropy index (H) (Magurran 1988; Legendre and Legendre 1998). GR was measured as number of genera per 1000 m–3 and H was measured as q

H = å pi log10 pi i =1

where pi = ni /N, the proportion of individuals n of genus i of the total number of individuals N in a sample, and q is the total number of genera within a sample. Evenness, or the proportional abundances of each genus within a sample, is also an aspect of diversity. Shannon’s H accounts for both richness and abundance of each genus. A sample in which a given number of genera have equal abundance can be considered more diverse than the same number of genera dominated numerically by one or two genera. We also examined abundance patterns expressed as the number of individuals per 1000 m–3. Both GR and abundance were lognormally distributed and were log10(x + 1) transformed for various analyses. To ensure that patterns were not a function of sampling design, we examined the relationship among number of fish genera, number of individuals, volume towed, and sampling depth. The volume towed increased with sampling depth (r = 0.71, p < 0.001, n = 2260); the greater the depth, the more water that was filtered. Both volume towed and sampling depth were negatively related to number of fish genera and number of individuals; most individuals were observed at depths less than 110 m. When sampling depth is held constant, the number of genera and number of individuals were not related to volume towed. The number of genera and number of individuals were only weakly related to the depth of tow (r = –0.13, p < 0.001, n = 2257 and r = –0.11, p < 0.001, n = 2257, respectively), holding tow volume constant. Given that the majority of taxa and individuals occurred in a shallower depth range, that the range of sampling depths was similar among months, and that the relationships between sampled variables and depth were weak, we assumed that our results were not an artifact of the sampling procedure. Not all months or areas were sampled each year during the SSIP, resulting in a patchwork of samples that precluded a meaningful analysis of among-year effects for each month. It might be possible to extract a subset of the data in which the same area had been sampled during the same period each year to examine interannual variability. However, our intent was to examine broad-scale spatial and seasonal patterns. To create a data set containing information on all months of the year and over a broad spatial scale, a stepwise procedure was followed. We examined interannual variability in H and GR in each month among years. The number of years within months varied from 1 to 5. If there were no significant differences among years within a month (based on Kruskal–Wallace nonparametric analyses), the data were pooled and considered to be representative of that month. If there were significant differences within months among years, we selected those survey years that were similar and had the largest sample size and the broadest spatial sampling. The net result was a monthly time series of GR, H, and abundance representing samples from across the shelf.

Cluster analysis One of our goals was to examine larval assemblages. Among the logistic problems associated with creating assemblages was that the distribution of sampling ranged over seasons and different regions. That is, the same taxa could occur at different locations on the shelf but in similar environments or at different times of the year. One approach would be to identify a series of separate sets of assemblages of genera within seasons and similar regions. We © 2000 NRC Canada



Shackell and Frank Fig. 1. (a) Scotian Shelf and Bay of Fundy: BoF, Bay of Fundy; Br, Browns Bank; L, LaHave Bank; E, Emerald Bank; Wes, Western Bank; Sab, Sable Island Bank; Md, Middle Bank; Ban, Banquereau; Ms, Misaine Bank; Ca, Canso Bank. (b) Distribution of SSIP samples from 1978 to 1982.

1749 Table 1. Number of samples by year and month extracted from the SSIP database. Survey year Month Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. Total

1978

1979

1980

1981

1982

Total

46

90

87 68 44

160 158 44 117 243 149 131 545 423 180 125 83 2358

27

117 74

199 125

196 90

324

85 30 619

134 149 64 150 121 118 40 53 919

35 67 87 62

348

148

the distance between clusters as a guideline. When the distance between clusters increased, we chose that as a point to select cluster groups. There was no depth information recorded for single samples of four genera, Gonichthys, Gonostoma, Nessorhamphus, and Notolynchus, so they were not included in the cluster analysis.

Results Genera composition The relative frequency of occurrence of the 91 genera extracted from the SSIP database is shown in Table 2. Merluccius, Sebastes, Urophycis, Glyptocephalus, and Ammodytes were the most frequently encountered genera. Their relative duration of spawning, larval phase or development rate, spatial distribution, and (or) abundance may be greater than that of other genera. The distributions of abundance for each genus were highly skewed. The highest median abundance was recorded for Merluccius, followed by Scophthalmus, Ammodytes, Urophycis, and Melanogrammus (Table 2).

opted for an approach that allowed us to examine a single set of assemblages that accounted for seasonal differences among regions. We identified assemblages of genera that co-occur in the same type of habitat as characterized by temperature and bottom depth (not sampling depth). Each combination of surface temperature and bottom depth could be considered a “habitat type,” and similar habitat types can co-occur across the shelf. Before clustering, we aggregated observations of each taxon by surface temperature and bottom depth, weighted by abundance. This resulted in a taxon’s average surface temperature and bottom depth of occurrence. These data were then subjected to cluster analysis. Members of a cluster did not necessarily occur at the same site but in similar combinations of surface temperature and bottom depth. We used a hierarchical (agglomerative) clustering method and Ward’s method as the distance formula (Legendre and Legendre 1998). Ward’s method uses the minimum variance summed over all variables as a measure of distance between two points. The groups with the smallest distance between them are joined, and then, groups are added successively using minimized distance as a criterion. We selected the number of clusters using the rate of change in

Seasonal and spatial patterns of larval fish diversity and abundance Median log GR was relatively low from December to February, increased in March, was fairly constant from April to September, and declined in October–November (Fig. 2a). Shannon’s H was relatively low from November to March and relatively higher from April to October (Fig. 2b). Total median log abundance was low from December to February, increased in March, was stable from April to June, and declined from July to October (Fig. 2). To present the spatial patterns and relationships among the diversity indices and abundance, months were grouped into three seasons based on the temporal and spatial similarity of diversity on a monthly scale: (i) December–March (winter), (ii) April–July (spring–summer), and (iii) August–November (summer–fall). If relationships were significant within seasons, we grouped seasons and reported the overall relationship. Log GR was significantly positively related to H (r = 0.62, p < 0.001, n = 1853) (Fig. 3a). A negative exponential best described the relationship between log GR and log abundance (R2 = 0.77; log GR = 1.37(1 – e–(1.13)(log abundance)), p < © 2000 NRC Canada



1750


Table 2. Genus, common name(s), possible number of species within genus in the Scotian Shelf area, frequency of occurrence, percentage of total count, and mean and median abundance of 91 genera. Genus

Common name(s)

No. of species

Frequency

% of total count

Mean abundance

Median abundance

Merluccius Sebastes Urophycis Glyptocephalus Ammodytes Enchelyopus Pleuronectes Gadus Benthosema Hippoglossoides Tautogolabrus Melanogrammus Clupea Citharichthys Lophius Scophthalmus Scomber Pollachius Liparis Ulvaria Lumpenus Myoxocephalus Brosme Bothus Notolepis Ceratoscopelus Triglops Aspidophoroides Lampanyctus Mallotus Pleuronectes Cryptacanthodes Protomyctophum Stomias Anarhichas Myctophum Stichaeus Symbolophorus Cyclothone Ophichthus Pholis Howella Maurolicus Anguilla Lobianchia Peprilus Echiodon Hygophum Scomberesox Reinhardtius Vinciguerria Argentina Hildebrandia Paraconger Bathylagus Chauliodus

Offshore, silver hakes Redfish Longfin, red, white hakes Witch flounder Sandlance Fourbeard rockling Yellowtail flounder Atlantic cod Lanternfish American plaice Cunner Haddock Atlantic herring Gulf Stream flounder Monkfish Windowpane Atlantic mackerel Pollock Snailfish Radiated shanny Shanny, eelblenny Sculpin Cusk Eyed flounder Barracudina Lanternfish Moustache sculpin Alligatorfish Lanternfish Capelin Winter flounder Wrymouth Protomyctophum Dragonfish Wolffish Lanternfish Arctic shanny Largescale lanternfish Anglemouth Eels Rock gunnel Howella Mullers pearlsides American eel Lanternfish Butterfish Chain pearlfish Lanternfish Saury Greenland halibut Lightfish Argentine Conger Conger Deepsea smelts Viperfish

2 3 3 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 3 3 1 1 1 2 1 1 3–6 1 1 1 1 1–2 3 2 1 1 4 3 1 1 1 1 2 1 1 2 1 1 3 2 ? ? 2 2

768 568 554 413 412 274 258 237 180 161 154 136 135 130 123 90 75 66 65 57 55 45 38 37 34 25 25 17 15 15 14 13 13 12 10 10 10 10 9 9 9 7 7 6 6 6 5 5 5 4 4 3 3 3 2 2

14.24 10.53 10.27 7.66 7.64 5.08 4.78 4.39 3.34 2.99 2.86 2.52 2.50 2.41 2.28 1.67 1.39 1.22 1.21 1.06 1.02 0.83 0.70 0.69 0.63 0.46 0.46 0.32 0.28 0.28 0.26 0.24 0.24 0.22 0.19 0.19 0.19 0.19 0.17 0.17 0.17 0.13 0.13 0.11 0.11 0.11 0.09 0.09 0.09 0.07 0.07 0.06 0.06 0.06 0.04 0.04

623.35 22.25 116.00 23.11 235.13 9.32 66.01 33.05 19.22 24.47 14.02 43.28 24.76 21.95 5.96 96.33 70.86 25.95 21.45 26.55 5.85 10.53 14.67 3.30 4.80 20.75 6.56 4.93 3.21 4.94 12.20 3.49 2.49 1.93 5.55 3.11 562.80 2.63 6.37 2.78 11.44 4.19 3.89 2.44 2.59 15.09 2.27 1.76 4.99 3.06 3.91 2.12 2.32 1.70 1.86 1.48

27.66 8.13 17.99 7.71 22.20 5.16 12.98 7.43 7.38 9.00 5.79 16.71 5.30 9.75 4.27 24.48 6.88 5.32 11.11 11.82 3.74 6.52 6.20 2.82 3.49 4.72 4.34 4.17 2.34 3.27 9.59 3.12 1.89 1.85 3.81 2.07 5.72 2.28 3.52 2.52 11.85 2.04 4.72 2.05 2.19 5.33 2.02 1.72 4.70 2.82 1.93 1.97 1.80 1.70 1.86 1.48 © 2000 NRC Canada



Shackell and Frank

1751

Table 2 (concluded). Genus Diaphus Gigantactis Icelus Nemichthys Paralepis Peristedion Pontinus Scorpaena Syacium Apogon Argyropelecus Aristostomias Artediellus Bascanichthys Bolinichthys Diplogrammus, Foetorepus, and Paradiplogrammus Gasterosteus Gonichthys Gonostoma Gymnothorax Hippocampus Hoplunnis Leptagonus Myrophis Nessorhamphus Nezumia Notolychnus Notoscopelus Phycis Prionotus Psenes Svetovidovia Synodus Uroconger Valenciennellus

Common name(s) Lanternfish Anglerfish Sculpin Snipe eel Barracudina Armored searobin Scorpionfish Smoothhead scorpionfish Channel flounder Cardinalfish Hatchetfish Loosejaw Hookear sculpin Sand-eel Lanternfish Dragonets

No. of species 8 2 2 1 1 1 1 1 1 2 5 4 2 ? 1 3

Stickleback Gonichthys Longtooth anglemouth Green moray Lined seahorse Hoplunnis Atlantic poacher Worm eel Duckbill oceanic eel Marlin-spike Notolychnus Notoscopelus Hake Northern searobin Driftfish Svetovidovia Red lizardfish Threadtail conger Valenciennellus

2 1 1 1 1 ? 1 2 1 1 1 1 ? 1 2 ? 1 1 1

Frequency 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

% of total count 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.02 0.02 0.02 0.02 0.02 0.02 0.02

0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02

Mean abundance 3.44 3.94 4.03 1.52 4.45 2.75 2.25 1.49 5.76 2.52 1.72 4.30 5.70 1.89 1.72 0.75

6.17 6.61 1.95 1.21 4.10 3.20 15.65 1.86 1.71 1.28 1.95 1.94 4.28 5.16 1.87 5.16 4.30 1.44 1.72

Median abundance 3.44 3.94 4.03 1.52 4.45 2.75 2.25 1.49 5.76 2.52 1.72 4.30 5.70 1.89 1.72 0.75

6.17 6.61 1.95 1.21 4.10 3.20 15.65 1.86 1.71 1.28 1.95 1.94 4.28 5.16 1.87 5.16 4.30 1.44 1.72

Note: Genera are ordered by frequency of occurrence.

0.001) (Fig. 3b). Shannon’s H was not related to log abundance in winter and in summer–fall and was negatively correlated in spring–summer (r = –0.12, p = 0.003, n = 593). A preliminary interpretation would be that log GR and log abundance were correlated because they were positively related to area sampled or to the number of individuals sampled. It is common to observe that the number of taxa encountered increases with sample size up to a given level, after which it stays constant. Sample size is one of various causes of a “species–area” curve (MacArthur 1965; Magurran 1988). In our study, the volume towed (sample size) was not significantly related to the number of either genera or individuals, holding sampling depth constant. Because samples were standardized, uniform, and small relative to the area of the shelf, we would not expect log GR or H to be related to sample size. In areas of high GR, abundance is high, whereas in areas of high H, abundance is low (Fig. 4). Given that GR and H

are positively related, how is it that they show different patterns with abundance? The difference is that GR is simply a measure of the number of genera, whereas H is a measure of the number of genera while accounting for varying abundances. Although the number of genera can result in an increase in abundance, a corresponding decrease in H reflects that an increase in abundance is increasingly dominated by fewer genera. This explanation is borne out by the shape of frequency of abundance distributions. Skewness was lowest in winter (skewness = 1.3, SE = 0.12), higher in summer– fall (skewness = 0.65, SE = 0.069), and highest in spring– summer (skewness = 0.4, SE = 0.097). We analysed the spatial distribution of diversity (GR and H) and abundance. Data were plotted into 0.031 decimal degree grids for the purposes of contouring. In general, the distribution of empty sets was greater in winter (Figs. 4a–4c), reflecting both a lower diversity and a more limited distribution. GR, H, and abundance were high in the Western – Sable © 2000 NRC Canada



1752 Fig. 2. (a) Monthly log GR (log10(number of genera + 1)·1000 m–3), (b) H, and (c) log abundance (log10(number of individuals + 1)·1000 m–3). Boxes represent interquartile range (50% of data), the horizontal line within the box represents the median, circles represent 1.5–3 box lengths from the edge of the box, and asterisks represent >3 box lengths from the outer edge of the box.

Can. J. Fish. Aquat. Sci. Vol. 57, 2000 Fig. 3. Relationships between (a) log GR (log10(number of genera + 1)·1000 m–3) and H, (b) log GR and log abundance (log10(number of individuals + 1)·1000 m–3), and (c) H and log abundance. In Fig. 3c, seasons are represented by plus signs (winter), open circles (spring–summer), and solid circles (summer–fall).

© 2000 NRC Canada



Shackell and Frank

1753

Table 3. Cluster groups as recognized by hierarchical agglomerative cluster analysis. Cluster

Genus

Egg type

1: shallow, cool

Leptagonus Hippoglossoides Clupea Melanogrammus Liparis Gadus Pleuronectes (winter flounder) Phycis Apogon Hoplunnis Ophichthus Psenes Scomberesox Aristostomias Synodus Citharichthys Glyptocephalus Prionotus Svetovidovia Urophycis Pleuronectes (yellowtail flounder) Merluccius Scophthalmus Gasterosteus Peprilus Scomber Bascanichthys Enchelyopus Lophius Tautogolabrus Syacium Argyropelecus Bathylagus Diplogrammus, Foetorepus, and Paradiplogrammus Hildebrandia Lobianchia Hygophum Gymnothorax Notoscopelus Valenciennellus Ammodytes Triglops Aspidophoroides Icelus

ProbD P D P D P D P D Unknown Unknown P P Unknown P Unknown P P ProbP P P P P D P P Unknown P P P P P P P

2: shallow, warmest

3: deep, cold 4: deep, warmest

5: shallow, cold

Unknown ProbP ProbP Unknown Unknown P D ProbD ProbD ProbD

Cluster

6: middepth, warmer

7: middepth, cool/warm

8: deep, warmest

9: deep, warm

10: middepth, warmest

11: shallow, warm

Genus

Egg type

Myoxocephalus Stichaeus Pholis Pollachius Anarhichas Cryptacanthodes Lumpenus Artediellus Anguilla Reinhardtius Argentina Maurolicus Bolinichthys Protomyctophum Echiodon Paraconger Vinciguerria Uroconger Benthosema Chauliodus Nemichthys Paralepis Ceratoscopelus Scorpaena Myctophum Symbolophorus Gigantactis Howella Lampanyctus Myrophis Nezumia Notolepis Stomias Bothus

D D D P D ProbD D D Unknown P P P ProbP Unknown P Unknown P Unknown Unknown O Unknown Unknown ProbP P Unknown Unknown Unknown P Unknown Unknown ProbP Unknown O ProbP

Pontinus Diaphus Cyclothone Peristedion Brosme Hippocampus Ulvaria Mallotus Sebastes

Unknown ProbP Unknown P P ProbO D D O

Note: Egg type: P, pelagic; D, demersal; O, oviviparous (ProbO, probably oviviparous; ProbD is probably demersal and ProbP is probably pelagic, as inferred by Lou Van Guelpen (Atlantic Reference Centre, St. Andrew’s by the Sea, NB EJB 2L7, Canada, personal communication; [email protected])). Range of depths observed: shallow (600 m) water. Within depth ranges, assemblages occurred in a range of temperatures roughly categorized as cold (5°C and 9°C and 12.4°C and 15°C).

Island banks area compared with the rest of the shelf (Figs. 4a–4c). During spring–summer, GR and H were both higher on LaHave, Emerald, Western – Sable Island, and Georges banks (Figs. 4d and 4e). GR was also high on Browns Bank and Banquereau (Fig. 4d). Abundance was relatively higher in the Middle, Canso, and Misaine banks and Banquereau areas and to a lesser extent in Browns and Sable Island bank areas (Fig. 4f). During summer–fall, GR and

abundance remained high on Western – Sable Island banks, whereas H did not (Figs. 4g–4i). Larval fish assemblages The five most frequently occurring genera during winter were Ammodytes (26.8%), Clupea (19.0%), Pollachius (12.8%), Gadus (10.6%), and Lumpenus (5.6%). During spring–summer, the most frequently encountered genera © 2000 NRC Canada



1754


Fig. 4. Seasonal distribution of (a, d, g) log GR, (b, e, h) H, and (c, f, i) log abundance: (a–c) winter, (d–f) spring–summer, and (g–i) summer–fall. Contour shades ranging from light to dark represent the 0, 25, 50, 75, and 100 percentiles for each metric.

were Sebastes (18.6%), Ammodytes (16.6%), Gadus (8.8%), Hippoglossoides (8.2%), and Melanogrammus (7.6%). During summer–fall, Merluccius (21.9%), Urophycis (16.1%),

Glyptocephalus (11.3%), Enchelyopus (7.6%), and Sebastes (7.2%) were the most frequently encountered genera. Associations among taxa based on cluster analysis re© 2000 NRC Canada



Shackell and Frank Fig. 4 (concluded).

1755

curred in a range of temperatures roughly categorized as cold (5°C and 9°C and 12.4°C and 15°C). These loosely defined temperature–depth categories have been used to describe the habitat type of each of the 11 assemblages (Table 3). To some extent, the assemblages aligned along the temperature axis (Fig. 6) represent seasonally occurring assemblages. For example, cluster 5 represents larvae occurring in cold water. Larvae that originated from demersal eggs dominated cluster 5 (Table 3) and their peak abundance occurred from November to June (Fig. 7a). Cluster 1 represents larvae that occurred in cool water. The peak larval abundance for cluster 1 occurred from April to June and October to December; the latter peak reflects the influence of the fall-spawning Clupea and Gadus components (Fig. 7b). Cluster 11 occurred in warmer water; peak larval abundance occurred from June to July (Fig. 7c). Cluster 2 represents larvae that occurred in warm water collected over shallow depths, the hakes and flounders. Abundance peaked in July (Fig. 7d). Clusters 4, 9, 8, 6, and 10 occurred in relatively warmer, deeper water and represent larvae of mesopelagic and deepwater fishes. Clusters 7 and 3 occurred in cooler water over greater depths and represent bathypelagic or deepwater fishes.

Discussion

vealed 11 assemblages (Table 3; Figs. 5 and 6). Assemblages occurred over a range of depths roughly categorized as shallow (600 m) water. Within these depth ranges, assemblages oc-

Diversity, abundance, and prey To examine the temporal correspondence between food availability and diversity, we used published information on zooplankton collected during the SSIP developed by Tremblay and Roff (1983), O’Boyle et al. (1984), Sameoto and Herman (1992), and Brander and Hurley (1992). In general, the zooplankton community composition can be related to hydrographic variables and varies spatially and temporally from inshore to offshore (Tremblay and Roff 1983) and by latitude (Sameoto and Herman 1992). During the SSIP 1979–1980 surveys, the zooplankton biomass peaked in May (O’Boyle et al. 1984). As reported by Sameoto and Herman (1992), the average concentration of Calanus finmarchicus, C. glacialis, and C. hyperboreus peaked from May to June and there was suggestion of a lesser peak in September. In our study, there was no obvious “peak” in larval diversity. Rather, diversity was relatively higher over a prolonged period (April to September–October). Because we might have obscured smaller-scale patterns at the shelf-scale level of analysis, we examined temporal patterns in diversity separately for the eastern and western Scotian Shelf (as defined by longitude 63.49°). There were no apparent peaks in diversity within regions. In general, the Scotian Shelf supports a changing array of larval fish throughout the year. The taxon composition is dynamic, but the level of diversity is fairly constant through spring–summer and early fall. There is a seasonal component to taxon composition and this is apparent in how the cluster groups are aligned on a surface temperature axis, as well as the seasonal progression of abundance. Interestingly, the peak abundance of members of cluster 1 occurs in spring and fall. This cluster contains taxa, such as Gadus and Melanogrammus, that conform to the conventional notion of a co-occurrence between larval fish and calanoid copepods. © 2000 NRC Canada



1756

Although Brander and Hurley (1992) and Sherman et al. (1984) suggested that various genera may be taking advantage of a spring and fall bloom, Leggett and Deblois (1994) could report only limited evidence to support the match– mismatch hypothesis. Further, Mousseau et al. (1998) suggested that, in general, larvae are not dependent on peak zooplankton blooms. They showed that larval fish production on Sable Island Bank can be partially supported by the microbial food web and may not be entirely dependent on spring and fall blooms. Both calanoid and cyclopoid copepods can prey on members of the microbial food web. A diversity of food webs may exist and different genera may use different food resources, and this may account for the fairly constant levels of diversity and the variation in larval abundance observed in this study throughout spring–summer and early fall. Our results suggest that there is substantial ecological latitude to support a variety of larval fish types throughout most of the year, and they support the assertion of Mousseau et al. (1998) that the match–mismatch hypothesis has only limited application. A limitation of our analysis is that we combined data from different years to create a seasonal cycle. Indeed, a lack of correspondence between larval diversity and zooplankton abundance could result from our methodology or because we are comparing larval and zooplankton abundance at too large a scale. We favor the interpretation that the stability of GR and H through spring–summer and early fall reflects that calanoid copepods do not have to be the direct or sole food source for the majority of larval fish. This is supported by the observation that peak abundances of various assemblages vary throughout the season and that the peak abundance of cluster 1 occurs during spring and fall, as would be predicted given the spawning characteristics of the cluster members (e.g., Gadus, Melanogrammus, Clupea). In several plant, terrestrial, and deep-ocean studies, diversity exhibits a dome-shaped relationship with productivity (reviewed in Rosenzweig and Abramsky 1993). Assuming that abundance is an indirect measure of productivity, our results support that observation; GR increased with abundance, reaching an asymptote at higher levels of abundance. Regions with high abundance have sufficient resources to support a higher richness, but diversity is dominated by fewer genera as abundance increases. Interestingly, the spatial pattern of zooplankton collected during the SSIP accords well with the spatial pattern of larval fish diversity that we observed. O’Boyle et al. (1984) described zooplankton biomass to be highly concentrated on Emerald and Western banks and the eastern end of Sable Island Bank and highest over Misaine Bank. This would support the notion that a sufficient resource base characterizes areas of relatively high richness and abundance. Diversity, abundance, and oceanographic circulation Does the surface circulation correspond to patterns of diversity and abundance? The most striking feature was that a partial gyre encircling the Western – Sable Island banks area in winter and summer (Fig. 8) (C. Hannah, Department of Fisheries and Oceans, P.O. Box 1006, Dartmouth, NS B2Y 4A2, Canada, personal communication; Cong et al. 1996) corresponded to high levels of diversity and abundance in that area. Gyres may promote retention and survival in


gadoid larvae (Werner et al. 1996; Shackell et al. 1999). It is possible that gyres may do so in other taxa, resulting in higher richness and (or) abundance associated with gyres. However, higher diversity is not exclusively observed in areas that promote retention, nor is it consistently higher in areas where there are gyres. The very general correspondence between circulation and diversity may reflect the scale of our study. Certainly, a drawback of many studies of larval assemblages and hydrography is the common attempt to relate average, or climatological, conditions (Cowen et al. 1993). We have compared the climatological circulation with larval diversity as determined from the SSIP years. A smaller-scale study may have revealed larval associations with fronts and more highly resolved gyres. The alternative is that despite the governing hypothesis that retention favors survival, many genera collected herein have an alternative strategy: maximum dispersal of larvae, instead of maximum retention, is more favorable to survival. We observed high levels of GR and abundance on Western – Sable Island banks. What are the mechanisms that support richness and abundance in that area? A species–area relationship describes the common observation that the number of species increases as area increases (MacArthur 1965). There are various causes of a species–area relationship including an increase in within- and (or) between-habitat diversity concomitant with an area increase until habitats are saturated. Certainly, the Western – Sable Island banks region is larger than other banks. It could also be considered a transitional zone between two or three distinct ecosystems, the cooler eastern Scotian Shelf, the warmer western Scotian Shelf, and the shelf-slope region. Merluccius and Urophycis, for example, occur primarily on Western – Sable Island banks and west, whereas Pleuronectes (yellowtail flounder) occurs primarily on Western – Sable Island banks and east. Western – Sable Island banks may contain more heterogeneous habitat and that may sustain a greater variety of fish. In addition, as noted above, zooplankton biomass on Western – Sable Island banks has been observed to be relatively high (O’Boyle et al. 1984). If this feature were consistent over time, it may itself be governed by the circulation and support a relatively greater number of species. We propose that GR is relatively high in the Western – Sable Island banks area because it is a larger bank with relatively higher habitat heterogeneity and a large partial gyre that serves as a retentive mechanism and to concentrate both larval fish and their prey. Diversity and conservation strategies We chose a shelf-scale level of analysis to describe larval diversity in part because there is little shelf-wide information on larval fish assemblages and because our results may have application to monitoring programs, to the delineation of shelf-scale ecosystems, and to the identification of potential protected areas. Diversity is often used as one of a suite of criteria to describe ecosystems or to identify potential protected areas. There is considerable debate as to the value of diversity and its role as an ecological process. In the marine environment, there is some information on patterns of diversity (Ormond et al. 1997), yet there exists little empirical information on the relationship between marine diversity, stability, and (or) © 2000 NRC Canada



Shackell and Frank

1757

Fig. 5. Dendrogram of hierarchical relationships among larval fish genera, as determined by hierarchical agglomerative cluster analysis. Common symbols denote members of the same cluster. Di, Fo, and Pa, Diplogrammus, Foetorepus, and Paradiplogrammus.

© 2000 NRC Canada



1758 Fig. 6. Cluster groups as per surface temperature and bottom depth. Symbols correspond to closest cluster labels (1–11), which correspond to cluster groups in Table 3. Solid circles, cluster 1; open triangles, cluster 2; open diamonds, cluster 3; open squares, cluster 4; solid triangles, cluster 5; solid squares, cluster 6; open circles, cluster 7; solid diamonds, cluster 8; ´’s, cluster 9; asterisks, cluster 10; plus signs, cluster 11.

resilience (Allison et al. 1996). Further to that is the observation that the pattern of diversity of one group does not necessarily coincide spatially or temporally with that of another (Nott and Pimm 1997). Where does our study fit in the rapidly evolving field of conservation? Larval assemblages are dynamic and may have changed since the SSIP, particularly on the eastern Scotian Shelf where Atlantic cod have collapsed and the fish community composition has changed (Frank et al. 1996). Biodiversity in general is changeable and is a function of biophysical processes. Because of potential change, conservation measures ideally would be based on those processes that govern its level and composition. Because processes are difficult to identify in the short term, the use of landscape- or habitatlevel features as descriptors of community diversity has been proposed (Angermeier and Winston 1999; Ward et al. 1999). We have identified seasons and areas that were favorable to larval finfish survival on the Scotian Shelf during the SSIP. A further question is which habitat features/descriptors and what scales are appropriate for conservation objectives. We showed that (i) diversity and abundance were relatively higher from April to September–October and that (ii) the level of diversity from April to September–October was stable but the taxon composition changed. From this, we inferred that the majority of taxa occurring on the Scotian Shelf were not dependent on a peak spring or fall bloom of calanoid copepods. Instead, as Mousseau et al. (1998) have proposed, there is a diversity of food webs supporting larval fish production. We also showed that (iii) GR was positively related to abundance but that the relationship asymptotes at higher levels of abundance and that (iv) H, which measures both diversity and evenness, was either unrelated or negatively related to abundance. Thus, diversity increases with abundance but the composition is dominated by relatively fewer genera at higher levels of abundance. We also showed

Can. J. Fish. Aquat. Sci. Vol. 57, 2000 Fig. 7. Monthly mean and SEM of log abundance of four clusters, (a) 5, (b) 1, (c) 11, and (d) 2, showing seasonal progression of abundance.

that (v) various banks were diverse and (or) productive areas, especially Western – Sable Island banks, and that the latter was associated with a large gyre that could act as a retentive © 2000 NRC Canada



Shackell and Frank

1759

Fig. 8. Mean velocities, averaged between 20 and 50 m, during (a) January–February and (b) July–August as estimated using a physical circulation model (reprinted with permission from the Journal of Physical Oceanography (Hannah et al. 2000)).

mechanism and could concentrate both larval fish and their prey. In this initial contribution, we did not address those smaller-scale associations, e.g., whether larvae were concentrated in fronts coincident with their food (see McLaren et al. 1997). Smaller-scale research should reveal whether circulation in high-diversity areas enhances larval survival and whether the temporal and spatial variability in larval finfish diversity reflects an abundant and (or) diverse food supply.

Acknowledgements We thank Mark Fowler for his singular understanding of the SSIP archival database. We also thank Lou Van Guelpen of the Atlantic Reference Centre for information on the life history of a vast array of marine fish taxa. Ian Perry reviewed an earlier version of the manuscript and provided generous advice. We are grateful to Dave Brickman, Susanna Fuller, Inka Milewski, and Daphne Themelis for thoughtful discussion. This is a GLOBEC Canada contribution.

References Allison, G.W., Menge, B.A., Lubchenco, J., and Navarrete, S.A. 1996. Predictability and uncertainty in community regulation:

consequences of reduced consumer diversity in coastal rock ecosystems. In SCOPE 55: functional roles of biodiversity: a global perspective. Edited by H.A.Mooney, J.H. Cushman, E. Medina, O.E. Sala, and E.-D. Schulze. ICSU UNEP, John Wiley & Sons, Toronto, Ont. pp. 371–392. Angermeier, P.L., and Winston, M.R. 1999. Characterizing fish community diversity across Virginia landscapes: prerequisite for conservation. Ecol. Appl. 9: 335–349. Brander, K., and Hurley, P.C.F. 1992. Distribution of early-stage Atlantic cod (Gadus morhua), haddock (Melanogrammus aeglefinus), and witch flounder (Glyptocephalus cynoglossus) eggs on the Scotian Shelf: a reappraisal of evidence on the coupling of cod spawning and plankton production. Can. J. Fish. Aquat. Sci. 49: 238–251. Cong, L., Sheng, J., and Thompson, K.T. 1996. A retrospective study of particle retention on the outer banks of the Scotian Shelf 1956–1993. Can. Tech. Rep. Hydrogr. Ocean Sci. No. 170. Cowen, R.K., Hare, J.A., and Fahay, M.P. 1993. Beyond hydrography: can physical processes explain larval fish assemblages within the Middle Atlantic Bight? Bull. Mar. Sci. 53: 567–587. Cushing, D.H. 1975. Marine ecology and fisheries. Cambridge University Press, London, U.K. Cushing, D.H. 1995. Population production and regulation in the sea: a fisheries perspective. Cambridge University Press, Cambridge, U.K. Doyle, M.J., Morse, W.W., and Kendall, A.W., Jr. 1993. A compar© 2000 NRC Canada



1760 ison of larval fish assemblages in the temperate zone of the Northeast Pacific and Northwest Atlantic oceans. Bull. Mar. Sci. 53: 588–644. Frank, K.T., and Leggett, W.C. 1983. Multispecies larval fish associations: accident or adaptation? Can. J. Fish. Aquat. Sci. 40: 754–762. Frank, K.T., Carscadden, J.E., and Simon, J.E. 1996. Recent excursions of capelin Mallotus villosus to the Scotian Shelf and Flemish Cap during anomalous hydrographic conditions. Can. J. Fish. Aquat. Sci. 53: 1473–1489. Grumbine, R.E. 1997. Reflections on “What is ecosystem management?” Conserv. Biol. 11: 41–47. Hannah, C.G., Shore, J.A., Loder, J.N., and Naimie, C.E. 2000. Seasonal circulation on the western and central Scotian Shelf. J. Phys. Ocean. In press. Iles, T.D., and Sinclair, M. 1982. Atlantic herring: stock discreteness and abundance. Science (Washington, D.C.), 215: 627–633. Legendre, P., and Legendre, L. 1998. Numerical ecology. Developments in environmental modelling 20. 2nd English ed. Elsevier Science Publishers, Amsterdam, The Netherlands. Leggett, W.C., and DeBlois, E. 1994. Recruitment in marine fishes: is it regulated by starvation and predation in the egg and larval stages? Neth. J. Sea Res. 32: 119–134. MacArthur, R.H. 1965. Patterns of species diversity. Biol. Rev. 40: 510–533. Magurran, A.E. 1988. Ecological diversity and its measurement. Princeton University Press, Princeton, N.J. McLaren, I.A., Avendano, P., Taggart, C.T., and Lochmann, S.E. 1997. Feeding by larval cod in different water-masses on Western Bank, Scotian Shelf. Fish. Oceanogr. 6: 250–265. Moser, G.H., and Smith, P.E. 1993. Larval fish assemblages and oceanic boundaries. Bull. Mar. Sci. 53: 283–289. Mousseau, L., Fortier, L., and Legendre, L. 1998. Annual production of fish larvae and their prey in relation to size-fractionated primary production (Scotian Shelf, NW Atlantic). ICES J. Mar. Sci. 55: 44–57. Nott, M.P., and Pimm, S.L. 1997. The evaluation of biodiversity as a target for conservation. In The ecological basis of conservation: heterogeneity, ecosystems, and biodiversity. Edited by S.T.A. Pickett, R.S. Ostfeld, M. Shachak, and G.E. Likens. Chapman and Hall. International Thomson Publishing, London, U.K. pp. 125–135.

Can. J. Fish. Aquat. Sci. Vol. 57, 2000 O’Boyle, R.N., Sinclair, M., Conover, R.J., Mann, K.H., and Kohler, A.C. 1984. Temporal and spatial distribution of ichthyplankton communities of the Scotian Shelf in relation to biological, hydrological and physiographic features. Rapp. P.-v. Réun. Cons. Int. Explor. Mer, 183: 27–40. Ormond, R.F.G., Gage, J.D., and Angel, M.V. (Editors). 1997. Marine biodiversity: patterns and processes. Cambridge University Press, Cambridge, U.K. Rosenzweig, M.L., and Abramsky, Z. 1993. How are diversity and productivity related? In Species diversity in ecological communities: historical and geographical perspectives. Edited by R.E. Ricklefs and D. Schluter. University of Chicago Press, Chicago, Ill. pp. 52–65. Sameoto, D.D., and Herman, A.W. 1992. Effect of the outflow from the Gulf of St. Lawrence on Nova Scotia Shelf zooplankton. Can. J. Fish. Aquat. Sci. 49: 859–869. Scott, W.B., and Scott, M.G. 1988. Atlantic fishes of Canada. Can. Bull. Fish. Aquat. Sci. No. 219. Shackell, N.L., Frank, K.T., Petrie, B., Brickman, D., and Shore, J. 1999. Dispersal of early life stage haddock (Melanogrammus aeglefinus) as inferred from the spatial distribution and variability in length-at-age of juveniles. Can. J. Fish. Aquat. Sci. 56: 2350–2361. Sherman, K., Smith, W., Morse, W., Berman, M., Green, J., and Ejsymont, L. 1984. Spawning strategies of fishes in relation to circulation, phytoplankton production, and pulse in zooplankton off the northeastern United States. Mar. Ecol. Prog. Ser. 18: 1–19. Tremblay, M.J., and Roff, J.C. 1983. Community gradients in the Scotian Shelf zooplankton. Can. J. Fish. Aquat. Sci. 40: 598–611. Ward, T.J., Vanderklift, M.A., Nicholls, A.O., and Kenchington, R.A. 1999. Selecting marine reserves using habitat and species assemblages as surrogates for biological diversity. Ecol. Appl. 9: 691–698. Werner, F.E., Perry, R.I., Lough, R.G., and Naimie, C.E. 1996. Trophodynamic and advective influences on Georges Bank larval cod and haddock. Deep-Sea Res. II, 43: 1793–1822. Winemiller, K.O., and Rose, K.A. 1992. Patterns of life-history diversification in North American fishes: implications for population regulation. Can. J. Fish. Aquat. Sci. 49: 2196–2218.

© 2000 NRC Canada
