863
Biological oceanography and fisheries management:
perspective after 10 years
Trevor Platt, Shubha Sathyendranath, and César Fuentes-Yaco
Platt, T., Sathyendranath, S., and Fuentes-Yaco, C., 2007. Biological oceanography and fisheries management: perspective after 10 years. – ICES
Journal of Marine Science, 64: 863 – 869.
Despite 100 years of research into the relationship between oceanographic factors and fish recruitment, it has proved very difficult to
demonstrate causal connections between properties of the marine ecosystem and the success of fisheries. Some authors have been led
to conclude that such causal connections, therefore, do not exist: a corollary would be that biological oceanography is of limited
relevance to fisheries issues. However, it would be premature to dismiss biological oceanography as a tool in fisheries management.
If we have not been able to implicate ecosystem factors as a significant source of variance in fish recruitment, it may be because the
search has been conducted at an inappropriate scale, a consequence of the limitations of ships as oceanographic platforms. The advent
of remotely sensed data from satellites greatly extends the scales of time and space at which synoptic oceanography can be carried out.
Access to such data allows a wider range of hypotheses to be tested, than is possible with ships alone, on the relationship between
ecosystem factors and recruitment. Applications in both the Atlantic and Pacific have demonstrated strong fluctuations between years
in the timing and the intensity of phytoplankton dynamics, with implications for recruitment and growth of exploited populations of
fish and invertebrates. The results provide essential intelligence for those charged with stewardship of the marine environment.
Keywords: ecosystem variability, recruitment, remote sensing.
Received 30 October 2006; accepted 23 April 2007; advance access publication 12 June 2007.
T. Platt: Biological Oceanography Division, Bedford Institute of Oceanography, PO Box 1006, Dartmouth, Nova Scotia, Canada B2Y 4A2.
S. Sathyendranath: Plymouth Marine Laboratory, Prospect Place, The Hoe, Plymouth PL1 3DH, UK. C. Fuentes-Yaco: Department of
Oceanography, Dalhousie University, Halifax, Nova Scotia, Canada B3H 4J1. Correspondence to T. Platt: tel: þ1 902 4263793; fax: þ1 902
4269388; e-mail: [email protected]
Historical introduction
The inaugural meeting of the International Council for the
Exploration of the Sea (ICES) was held in 1902. At least since
then, the community of fisheries scientists has debated the relative
importance of fishing pressure and environmental conditions as
determinants of recruitment to exploited stocks (Hjort, 1914;
Cushing, 1974, 1990; Mills, 1989; Smith, 1994). In turn, these
debates have given rise to enormous expenditure on research
designed to elucidate and quantify the contribution of fluctuations
in environmental conditions to variance in recruitment. The term
“environmental conditions” is understood here to include both
physical properties, such as temperature, and ecological properties, such as the abundance and growth of plankton.
With respect to the physical properties of the environment, progress has been made in understanding, for example, the preferences
for temperature and salinity of exploited species. This new knowledge has been applied effectively in many harvest fisheries, such as
that for tuna (Polovina et al., 2001), and to understand the
migrations by populations of demersal fish, such as cod (Gadus
morhua). More direct examples include the demonstration in
gadoid fish of a dependence of recruitment on temperature
(Planque and Frédou, 1999), and on the North Atlantic
Oscillation (Brander and Mohn, 2004). Off West Greenland, recruitment to a variety of fishery resources, including invertebrates, has
been related to hydrographical variations (Buch et al., 2003).
However, with respect to our principal concern, variability in
the properties of the ecosystem (Hjort, 1914; Cushing, 1974,
1990), progress has been less easy to demonstrate. Although it is
true that the basis for present-day understanding of seasonal
changes in the pelagic ecosystem was laid down early in the life
of ICES, it has proved altogether more difficult either to apply
this knowledge to improve our understanding of variability in
recruitment or to integrate it into the emerging discipline of fisheries management.
An important consequence of the inability to translate advances
in understanding plankton dynamics into operational bases for
fisheries management has been that fisheries science and biological
oceanography have developed more independently of each other
than would have seemed desirable in 1902. Therefore, generally
speaking, the theory of fish populations has been based on the
premise that predictions about future stock size could be made
using as independent variables only the population structure of
the present stock (Beverton and Holt, 1957), that is, to say
without consideration, implicit or explicit, of the requirement of
the stock (in all its life stages) to nourish itself.
Dissent has arisen, from time to time, against this way of thinking (e.g. Paloheimo and Dickie, 1970). Beverton (1998) himself
subsequently acknowledged the shortcomings of the method.
Dissenters pursued further research devoted to building an
ecological context for fisheries management (Cushing, 1975).
Crown Copyright # 2007. Published by Oxford Journals on behalf of the International Council for the Exploration of the Sea.
All rights reserved
864
However, although much elegant work was done, it proved difficult to merge the results into the body of principles upon which
fisheries management is based. If a causal thread did indeed
exist, it was slipping through scientists’ fingers.
During the 1990s, collapse of a number of major fish stocks
around the world brought into focus the fragility of resources
upon which so many depended for a livelihood and for a source
of protein. More than that, however, it emphasized once again
our inability to account for the fluctuations in fish stocks by partitioning the variance into contributions arising from fishing
pressure and contributions from the ecosystem.
In an ICES symposium paper, we argued (Platt and
Sathyendranath, 1996) that the failure to demonstrate the influence of ecosystem fluctuations on recruitment of exploited
species was a consequence, not necessarily of the lack of such a
link, but rather of the inadequacy of the tools available until
then, in particular to the limitations of ships as oceanographic
platforms. The nature of the tools available controls the way
research questions can be posed. The advent of instruments for
remote sensing of an important ecosystem property (autotrophic
biomass) expanded the range of questions that could be addressed,
and offered the possibility that biological oceanography could
develop an operational component to complement its research
achievements. Here, we discuss the extent to which this promise
has been fulfilled.
T. Platt et al.
context. At sea, the conventional observation platform (a research
vessel) can deliver data of very high quality, but is too slow, and
cuts too narrow a swath, for any pretension to synopticity.
Remote sensing by visible spectral radiometry provides the only
synoptic-scale window we have into the pelagic ecosystem. It is
likely to remain for the foreseeable future the only such option,
as far as we are able to judge at present.
Compared with the data that can be collected by ship, the data
collected by remote sensing are of lower precision and lack information on vertical structure. However, they make up in sample
size what they lack in precision. For example, a ship might be
able to sample 50 points on the perimeter of a 100 km box in
1 d. At a local resolution of 1 km, a satellite will provide 104 simultaneous observations inside the box. Given that, generally speaking, sampling variance is inversely proportional to the number of
observations, this is a powerful attribute indeed. Moreover, remote
sensing will show how conditions inside the box relate to those
outside, even as far as the edges of the ocean, and it will do so
again tomorrow, and the next day, for as long as we remain interested in the data. Finally, sea surface temperature can be collected
by remote sensing on the same scales of time and space, opening
the possibility of a multidisciplinary, synoptic approach to the
marine ecosystem.
Application of remotely sensed information
The advent and advantages of ocean-colour
remote sensing
One of the recurring problems in fieldwork with ships at sea is the
gross undersampling that we are obliged to accept, a simple consequence of the speed of ships and the size of the ocean.
Therefore, we are faced continually with the task of extrapolation
of sparse datasets if we strive to produce conclusions at the
regional and larger scales of interest to fisheries scientists and climatologists. Creation of a well-resolved serial record of ecosystem
properties for a large region using ships alone is excluded on the
grounds of expense. A practical corollary, for example, was the
lack of a synoptic and serial record describing the history of
the ecosystem in the eastern North Atlantic Ocean in the
20 years or so before the collapse of the groundfish stocks there
during the 1990s.
Within the past 20 years, a tool of enormous power has been
introduced that simplifies (but does not eliminate) the task of
extrapolation: visible spectral radiometry, also known as remote
sensing of ocean colour (Esaias et al., 1986). Notwithstanding
the limitations of this tool, with its aid, we can see, for the first
time, the synoptic distribution of a biologically important quantity
(autotrophic biomass indexed as chlorophyll concentration) at the
regional scale of the ocean. These data can be applied to derive estimates of phytoplankton production, another basic property of the
pelagic ecosystem, at the same scales of time and space. Remotely
sensed data should be seen as complementary to data acquired
using ships. As a platform for studying processes in the marine
ecosystem, and for the development of prototype instrumentation,
a ship is ideal. As a platform for synoptic coverage of the ocean,
however, a ship is limited by slow speed and lack of peripheral
vision. In spatially distributed systems, such as the pelagic ecosystem, the value of synoptic-scale observations is widely recognized.
For example, the quality of local weather forecasts is greatly
enhanced if local observations can be set in a broader synoptic
Shortly after first-generation, ocean-colour data were released into
the public domain (Esaias et al., 1986), a 10-year hiatus in the data
stream commenced. The SeaWiFS sensor was launched towards
the end of 1997 and continues to provide high-quality data. On
the eve of this launch, we discussed ways in which the new data
might be applied with advantage (Platt and Sathyendranath,
1996). It was recognized that with the new data, the pelagic ecosystem could be characterized in ways that would be completely inaccessible to ships, even with unlimited availability of ship-days.
One of the key phenomena that could be characterized in this
way was the spring bloom of phytoplankton, the strongest signal in
the seasonal cycle of the marine ecosystem. The spring bloom featured prominently in the early ideas of Hjort (1914) and Cushing
(1974, 1990), the success of recruitment in exploited stocks being
linked to the quality of the bloom. Cushing saw the quality of the
bloom as related mainly to its timing, such that when the timing of
the bloom occurred at a suitable lag from the spawning of the
exploited stocks, the two phenomena were said to be matched.
Otherwise they were mismatched. Blooms that were matched to
occurrence of larvae led to superior recruitment.
It was recognized that remotely sensed data offered the opportunity to make a test at appropriate scales of time and space. We
proposed (Platt and Sathyendranath, 1996) to characterize the
spring bloom with respect to the timing of initiation, the amplitude of the peak, the width (duration) of the peak, and the
timing (phase) of the peak (Figure 1). These properties could be
calculated for all pixels in the region of interest, such that all
spatial structure would be preserved. The statistical moments of
all of these properties, and their variation between years, could
be calculated and the results used as tools to analyse the effect of
ecosystem fluctuations on exploited stocks. The bloom properties
could be applied singly, or in linear combination, to provide an
index of ecosystem function for comparison against fluctuations
in recruitment of exploited stocks. This index, or indices, and its
statistical moments could be computed with a spatial resolution
Biological oceanography and fisheries management
Figure 1. Characteristics of the phytoplankton spring bloom: timing
of initiation, and timing and duration of maximum amplitude.
of 1 km over a spatial domain commensurate with the lifetime
range of the exploited stock of interest.
Further, because sea surface temperature could be estimated
from satellites at the same scales of time and space as ocean
colour, we could begin to treat the dependence of ecosystem function on physical forcing at a scale that is relevant to the fish.
Whereas before (using ships), it had not been possible to
address the possible connection between ecosystem fluctuations
and variability in recruitment at an appropriate scale, we could
now begin to practice holistic operational oceanography. Such
was the prospectus laid out in Platt and Sathyendranath (1996).
Next, we evaluate the prospectus in the light of developments
over the past 10 years.
Developments over the past decade
Analysis of the effect of ecosystem fluctuation on exploited stocks
is a problem in time-series, and the first requirement is to construct a suitable series. The region of interest to our group is the
so-called Canadian Atlantic Zone, the eastern seaboard of
Canada set in its oceanographic context (Figure 2). For this
region, we construct weekly composite images for chlorophyll
and temperature.
The data have a nominal spatial resolution of 1 km. For all years
in the time-series, we can characterize the spring phytoplankton
bloom climatologically according to the properties already mentioned, on every pixel in the composite image covering the
Canadian Atlantic Zone. Thus, the indices are resolved at 1 km
and all spatial structure in the fields is preserved (see Figure 2
for an example). Further, the indices of the bloom can be calculated for individual years and compared with the climatological
means. In this way, we can assess interannual variations in
properties and also evaluate whether in a particular year events
are retarded or advanced compared with the mean. In other
words, we can calculate the anomalies in properties for particular
years. For example, we found that the timing (peak) of the spring
bloom could vary by at least six weeks between years in a given
region (Figure 3).
After the time-series had accumulated for more than 5 years, we
embarked on a trial application. The idea was to make an operational test of the Cushing– Hjort match/mismatch hypothesis.
Under the null hypothesis that between-year variance in recruitment is independent of fluctuations in the properties of the
spring bloom, we applied the chlorophyll time-series to an independent series of data on recruitment of haddock (Melanogrammus
aeglefinus) collected on the continental shelf off Nova Scotia. This
is a 30-year series in which 2 years stand out as having produced
865
exceptional year classes. For both those years, remotely sensed
data are available.
We found (Platt et al., 2003) that the null hypothesis should be
rejected. The recruitment of haddock, normalized to biomass of
the spawning stock, was highly correlated with the timing of the
bloom. Early blooms were associated with better recruitment.
The two exceptional year classes were in years with unusually
early spring blooms. It is difficult to see how this result could
have been established other than by use of remotely sensed data.
Anomalies of bloom timing accounted for 95% of the variance
in normalized recruitment of haddock under a quadratic model
(Figure 4). The results argue for the importance of the trophic
link between phytoplankton and fish stocks, especially the importance of fluctuations between years at the autotrophic level. The
tentative explanation advanced for the results was that, for
species with protracted spawning (such as haddock), an early
spring bloom would confer enhanced survival on early larvae. In
a modelling study, Brander et al. (2001) also concluded (tentatively) that local meteorological forcing affects gadoid survival,
through its control on the match between plankton production
and larval feeding.
Another application of remotely sensed data for the same
Atlantic Zone dealt with the northern shrimp Pandalus borealis
(Fuentes-Yaco et al., 2007; Koeller et al., 2007). The fishery for
this species, covering the entire Labrador Shelf, is divided into subareas according to latitude. Carapace length of the adults in a particular subarea correlates positively with the intensity of the bloom
in that subarea. Timing of bloom onset was also important
(Fuentes-Yaco et al., 2007). Within subareas (e.g. subarea 6,
Hamilton Bank), the minimum size of males was greater in
years when the bloom was early than in years when the bloom
was late.
Research in areas other than the North Atlantic Ocean has also
shown the utility of remote sensing as a tool for understanding
ecosystem dynamics in a fisheries context. In the Japan Sea,
Yamada and Ishizaka (2006) studied the timing of the spring
bloom, in particular its long-term variation. Bloom initiation
was correlated with average winds in spring, such that light
winds favoured early blooms. Early occurrence of the peak was
associated with strong density gradients over the upper 30 m.
These results, consistent with the theory of Sverdrup (1953) on
bloom development, were used to reconstruct the characteristics
of the bloom in earlier years when satellite data were lacking.
During the 30-year period beginning in 1972, the reconstructed
times of bloom initiation and peak occurrence varied, with the
mid-1980s characterized by early blooms of short duration. The
changes in bloom timings were also associated with changes in
the community structure of the phytoplankton (Chiba and
Saino, 2002). Oligotrophic summer-like conditions were established early in years when the bloom was early.
On the west coast of Canada, using remotely sensed imagery of
ocean colour, Ware and Thomson (2005) found evidence for ecosystem control mediated by fluctuations in phytoplankton production. They recognized that in any ecosystem, removal of
predatory fish at increasingly high rates would eventually result
in a major perturbation of ecosystem structure. The threshold
removal rate at which this may take place is, of course,
unknown, but it is believed not to have been reached in the
Northeast Pacific Ocean. Rather, Ware and Thomson (2005)
found that, from Alaska to California, latitudinal variation in
fish yield correlates strongly with that in primary production, as
866
T. Platt et al.
Figure 2. (a) Climatology of the timing of the spring bloom maximum in chlorophyll concentration for the Northwest Atlantic Zone for the
period 1998– 2004. (b – f) Anomalies for specific years (1998 – 2000, 2003, and 2004) with respect to the climatology. The timing of bloom
maximum is computed on the basis of time-series of available data for individual pixels. Gaps in the time-series (because of cloud cover, for
example) are ignored in these computations.
Biological oceanography and fisheries management
867
Figure 3. Anomalies in weeks (mean + s.d.) for the timing of the bloom on the Scotian Shelf along a line from 44.408N 63.468W to 42.538N
61.408W (the Halifax Line) for the period 1998– 2005. The mean and the standard deviations are computed for 118 pixels along the line.
inferred loosely from remotely sensed data on ocean colour.
Similar results were found for resident fish and for highly
migratory fish. Spatial variations in ocean-colour fields explained
some 87% of the variance in yields of resident fish. It made no
difference whether the area analysed was dominated by coastal
upwelling or by coastal downwelling. When the analysis was
applied to the resident fish yield and chlorophyll fields for the
restricted area of the British Columbian coast, a very high proportion of the variance in fish yield could be accounted for.
The principal causes of spatial structure in the chlorophyll
fields were thought to be differential strength of coastal upwelling,
run-off from major rivers, and width of the continental shelf.
These are the physical origins of ecosystem modulation mediated
by phytoplankton, and their aggregate effect accounts for the
Figure 4. Relationship between survival index of haddock
(normalized to spawning biomass) and local anomalies in bloom
timing. A quadratic fit explains 95% of the variation in recruitment.
structure observed in ocean-colour imagery. A time-series of
ocean-colour data may then be used as an aid to understand variations in fish production along the grand sweep of the western seaboard of North America.
Also for the North Pacific Ocean, Polovina and Howell (2005)
discussed the extraction of ecosystem indicators from remotely
sensed data on sea-surface height, temperature, wind, and ocean
colour. They characterized large-scale structure using the principal
empirical –orthogonal function modes of monthly averaged seasurface height. These modes carry information on the vertical
structure of the ocean, an important determinant of primary production. They are also relevant to the occurrence and strength of
the El Niño perturbation. Ocean-colour data were used to describe
interannual changes in the large-scale chlorophyll fields. King and
McFarlane (2006) argue for a management framework that would
take such information into account.
In their review of clupeoid populations in coastal upwelling
systems, Cole and McGlade (1998) emphasized the utility of remotely sensed sea-surface temperature as an aid to elucidation of
environmental effects on population variability. Cole (1999)
pursued this approach in the Benguela upwelling system to discover strong dependence of clupeoid recruitment on environmental conditions.
Another application of the chlorophyll fields is to recognize the
boundaries between ecological provinces (sensu Longhurst, 1998).
For example, Polovina and Howell (2005) discuss the Transition
Zone Chlorophyll Front separating the Subarctic and subtropical
gyres: it is an important feeding area for large pelagic fish. The
salient point here is that the front (easily detected in chlorophyll
fields) has a seasonal excursion north and south of some
1000 km. In particular, the southern extent of the front in
winter is variable between years, with important implications for
ecosystem dynamics. In the Northwest Atlantic, the seasonal readjustment of boundaries between ecological provinces has also been
studied (Platt et al., 2005). The location of the boundaries can be
868
monitored objectively in real time using remotely sensed chlorophyll fields. The results are important for ecosystem analysis, for
example computation of phytoplankton production, because
usually, different ecological rate parameters are assigned to different provinces. The instantaneous boundaries provide the template
on which the assignment can proceed.
Concluding remarks
In fisheries ecology, the advent of remote sensing has greatly
expanded the scope of research questions that can be addressed
compared with what is possible using ships alone. A much
broader range of scales of time and space is available for consideration. Issues such as the connection between fluctuations in the
pelagic ecosystem and variability in recruitment of exploited
stocks can now be addressed at the relevant scales. Application
of remotely sensed data will become progressively more important
as we learn, collectively, how better to interpret them. An increasing reliance on satellites does not mean that ships will become
superfluous to the conduct of oceanography. The best work will
combine the excellence of data that can be achieved using ships
with the spatial and temporal coverage that can be achieved with
satellites. Ships will always be required to calibrate satellite
signals. If we are to get the most out of satellite data, we must continue to have access to oceanographic vessels of the highest quality.
The future lies in the optimal blending of the best features of data
derived from satellites and ships.
Other types of remote sensing, conducted from ships or otherwise, will be important in helping to elucidate the mechanisms
that underlie the results we have described. For example, Pepin
(2004) emphasized the value of acoustic surveys of predatory
fish for interpreting data on survival of larvae. We should be
alert to the spatial and temporal scales that different methods
can access, and learn to combine the results from various
methods, at their appropriate scales, in reaching a synthesis.
We conclude, therefore, that the prospectus laid out in Platt
and Sathyendranath (1996) has been fulfilled, with new results
relating to the spatial structure of the pelagic ecosystem, the intensity of its cycles and the variations between years in their phase
relationships (timing), having important consequences for
exploited stocks. Visible spectral radiometry (ocean colour) gives
us the possibility to examine the spatial gradients of primary production, their seasonal variation, and their fluctuations between
years. It is the best option we have to establish the long-term
history of the pelagic ecosystem in any region. By its use, we can
contribute intelligently to discussions about the influence of ecosystem trends on trends in exploitable stocks. We can supplement
the information with remotely sensed data on sea surface height
and temperature to achieve a holistic oceanographic synthesis. It
is an opportunity not to be discarded, in that it provides the
basis for a new partnership between fisheries managers and biological oceanographers. In his retrospective synthesis published
posthumously, Beverton (1998) recognized two main sources of
uncertainty in fisheries assessment. One is physical, deriving
from climate, and the other biological, deriving from species interactions. It is clear that physically forced ecosystem fluctuations can
be characterized and kept under surveillance using remote sensing,
removing, to some extent, one of the two sources of uncertainty.
Therefore, questions of mutual interest to oceanographers and
fisheries scientists can be addressed at the correct scales for the
first time. In this way, the two fields of research may finally
T. Platt et al.
achieve the coordinated development hoped for by the founders
of ICES.
Acknowledgements
We thank Alida Bundy, Alan Longhurst, and Venetia Stuart for
constructive comments. The work was supported by the
Canadian Space Agency (GRIP Programme), and satellite data
were provided courtesy of NASA and Orbimage.
References
Beverton, R. 1998. Fish, fact and fantasy: a long view. Reviews in Fish
Biology and Fisheries, 8: 229– 249.
Beverton, R. J. H., and Holt, S. J. 1957. On the Dynamics of Exploited
Fish Populations, Series II, XIX. Fishery Investigations, London.
533 pp.
Brander, K. M., Dickson, R. R., and Shepherd, J. G. 2001. Modelling
the timing of plankton production and its effect on recruitment
of cod (Gadus morhua). ICES Journal of Marine Science, 58:
962– 966.
Brander, K., and Mohn, R. 2004. Effect of the North Atlantic
Oscillation on recruitment of Atlantic cod (Gadus morhua).
Canadian Journal of Fisheries and Aquatic Sciences, 61:
1558– 1564.
Buch, E., Nielsen, M. H., and Pedersen, S. A. 2003. On the coupling
between climate, hydrography and recruitment variability of fisheries resources off West Greenland. ICES Marine Science
Symposia, 219: 231 – 240.
Chiba, S., and Saino, T. 2002. Interdecadal change in the upper water
column environment and spring diatom community structure in
the Japan Sea: an early summer hypothesis. Marine Ecology
Progress Series, 231: 23 – 35.
Cole, J. 1999. Environmental conditions, satellite imagery, and clupeoid recruitment in the northern Benguela upwelling system.
Fisheries Oceanography, 8: 25 – 38.
Cole, J., and McGlade, J. 1998. Clupeoid population variability, the
environment and satellite imagery in coastal upwelling systems.
Reviews in Fish Biology and Fisheries, 8: 445 –471.
Cushing, D. H. 1974. The natural regulation of fish populations. In Sea
Fisheries Research, pp. 399– 412. Ed. by F. R. Harden Jones. Elek
Science, London.
Cushing, D. H. 1975. Marine Ecology and Fisheries. Cambridge
University Press, Cambridge, UK.
Cushing, D. H. 1990. Plankton production and year-class strength in
fish populations: an update of the match/mismatch hypothesis.
Advances in Marine Biology, 26: 249– 294.
Esaias, W. E., Feldman, G. C., McClain, C. R., and Elrod, J. A. 1986.
Monthly satellite-derived phytoplankton pigment distribution for
the North Atlantic ocean basin. EOS, Transactions of the
American Geophysical Union, 67: 835– 837.
Fuentes-Yaco, C., Koeller, P. A., Sathyendranath, S., and Platt, T. 2007.
Shrimp (Pandalus borealis) growth and timing of the spring phytoplankton bloom on the Newfoundland – Labrador Shelf. Fisheries
Oceanography, 16: 116– 129.
Hjort, J. 1914. Fluctuations in the great fisheries of northern Europe,
viewed in the light of biological research. Rapports et ProcèsVerbaux des Réunions du Conseil Permanent International pour
l’Exploration de la Mer, 20: 1 – 228.
King, J. R., and McFarlane, G. A. 2006. A framework for incorporating
climate regime shifts into the management of marine resources.
Fisheries Management and Ecology, 13: 93 – 102.
Koeller, P. A., Fuentes-Yaco, C., and Platt, T. 2007. Decreasing shrimp
(Pandalus borealis) sizes off Newfoundland and Labrador—
environment or fishing? Fisheries Oceanography, 16: 105– 115.
Longhurst, A. 1998. Ecological Geography of the Sea. Academic Press,
San Diego.
Biological oceanography and fisheries management
Mills, E. L. 1989. Biological Oceanography. An Early History,
1870– 1960. Cornell University Press.
Paloheimo, J. E., and Dickie, L. M. 1970. Production and food supply.
In Marine Food Chains, pp. 499– 527. Ed. by J. H. Steele. Oliver
and Boyd, Edinburgh.
Pepin, P. 2004. Early life history studies of prey-predator interactions:
quantifying stochastic individual responses to environmental
variability. Canadian Journal of Fisheries and Aquatic Sciences,
61: 659– 671.
Planque, B., and Frédou, T. 1999. Temperature and the recruitment of
Atlantic cod (Gadus morhua). Canadian Journal of Fisheries and
Aquatic Sciences, 56: 2069– 2077.
Platt, T., Bouman, H., Devred, E., Fuentes-Yaco, C., and
Sathyendranath, S. 2005. Physical forcing and phytoplankton
distributions. Scientia Marina, 69: 55 –73.
Platt, T., Fuentes-Yaco, C., and Frank, K. T. 2003. Spring algal bloom
and larval fish survival. Nature, 423: 398 – 399.
Platt, T., and Sathyendranath, S. 1996. Biological oceanography and
fisheries management. ICES Document CM 1996/O: 3.
869
Polovina, J. J., and Howell, E. A. 2005. Ecosystem indicators derived
from satellite remotely sensed oceanographic data for the North
Pacific. ICES Journal of Marine Science, 62: 319 – 327.
Polovina, J. J., Howell, E., Kobayashi, D. R., and Seki, M. P. 2001. The
transition zone chlorophyll front, a dynamic global feature defining
migration and forage habitat for marine resources. Progress in
Oceanography, 49: 469– 483.
Smith, T. D. 1994. Scaling Fisheries: The Science of Measuring the
Effects of Fishing, 1885– 1955. Cambridge University Press,
Cambridge, UK.
Sverdrup, H. U. 1953. On conditions for the vernal blooming of
phytoplankton. Journal du Conseil International Permanent pour
l’Exploration de la Mer, 18: 287– 295.
Ware, D. M., and Thomson, R. E. 2005. Bottom-up ecosystem trophic
dynamics determine fish production in the Northeast Pacific.
Science, 308: 1280– 1284.
Yamada, K., and Ishizaka, J. 2006. Estimation of interdecadal change
of spring bloom timing, in the case of the Japan Sea. Geophysical
Research Letters, 33: L02608, doi:10.1029/2005GL024792.
doi:10.1093/icesjms/fsm072