Advances in conservation oceanography New tagging and tracking technologies and Their potential for transforming the science underlying Fisheries Management

Overexploitation of living resources and climate change are 
among the most obvious global-scale impacts of human society on marine 
ecosystems. In an age of such large-scale anthropogenic impacts, marine 
scientists, resource managers, and policymakers must rethink their approaches 
to protecting and managing marine populations and ecosystems. Conservation 
oceanography is an emerging field of science that incorporates the latest advances 
in ocean science and technology to provide resource managers and policymakers 
with the information they need to ensure the sustainability of the marine 
environment and its living resources. Here, we discuss the historical context of 
conservation oceanography as it applies to marine fisheries management. We 
then describe two projects, one focused on Atlantic bluefin tuna and the other 
on Pacific salmon, that illustrate the potential of new tagging and tracking 
technologies for transforming the science underlying fisheries management.

Abstr Act.Overexploitation of living resources and climate change are among the most obvious global-scale impacts of human society on marine ecosystems.In an age of such large-scale anthropogenic impacts, marine scientists, resource managers, and policymakers must rethink their approaches to protecting and managing marine populations and ecosystems.Conservation oceanography is an emerging field of science that incorporates the latest advances in ocean science and technology to provide resource managers and policymakers with the information they need to ensure the sustainability of the marine environment and its living resources.Here, we discuss the historical context of conservation oceanography as it applies to marine fisheries management.We then describe two projects, one focused on Atlantic bluefin tuna and the other on Pacific salmon, that illustrate the potential of new tagging and tracking technologies for transforming the science underlying fisheries management.

INtroductIoN
During the past half century, overfishing in many parts of the global ocean has had a substantial impact on predators at or near the top of the marine food chain (Myers and Worm, 2003;Sibert et al., 2006).In some regions, fisheries have responded by shifting their efforts to species further down the food chain (Pauly et al., 1998;Essington et al., 2006).
In addition, overfishing has blurred what the baseline conditions for many coastal ecosystems were like prior to human exploitation (Pauly, 1995;Jackson et al., 2001;Lotze et al., 2006), further complicating the setting of stock recovery goals.Recognition of these past impacts of overfishing, combined with the current overexploited status of many fish stocks, has encouraged a shift in fisheries management from a single-species approach, focused on maximizing yield, to an ecosystem-based approach, focused on rebuilding depleted fish stocks while protecting and restoring impacted ecosystems (Pauly et al., 2002;US Commission on Ocean Policy, 2004).In a global ocean responding to both fishing pressures from the top down and climate forcing from the bottom up (Frank et al., 2006;Cury et al., 2007;Greene and Pershing, 2007;Greene et al., 2008), such an ecosystem-based approach to fisheries management cannot continue to rely primarily on the same catch and survey data traditionally collected to support the single-species approach.Instead, a more dynamic framework for the science is needed, one that builds upon our  and Nielson, 2001;Block et al., 2003;Special Issue on Ocean Observations, Oceanography 16(4), 2003;Welch et al., 2003;Block, 2005;Block et al., 2005;Cooke et al., 2008).All of these technologies are in relatively early stages of development; hence, there is considerable scope for growth as the field of conservation oceanography matures.

hIstorIcAl coNtext
The roots of fisheries-associated conservation oceanography can be traced back to the first half of the twentieth century, when one of the principal goals of ocean scientists was to enhance the yield of commercially exploited fish stocks by improving our understanding of the ocean environment and its natural variability.To promote this agenda, scientific organizations, such as the International Council for the Exploration of the Sea (Rozwadowski, 2002), were founded, and ocean monitoring programs, such as the Continuous Plankton Recorder Survey (Reid et al., 2003), were initiated.Unfortunately, the ocean proved to vary greatly in both space and time, and the resources and technologies necessary to adequately sample and model such a highly variable environment were not available.
After the two world wars, it became clear that commercial harvesting was having an impact on fish population dynamics that was at least comparable to the effects of the ocean environment's natural variability.During the 1950s, fisheries management entered a period in which catch limits were set to maximize sustainable yields on the basis of single-species stock-recruitment models that largely ignored environmental variability (Schaefer, 1954;Ricker, 1954).
This approach to fisheries management avoided the problem of quantifying the ocean's environmental variability by assuming that it could be averaged out when determining the stock-recruitment relationship.The El Niño-related collapse of the Peruvian anchovy fishery during the early 1970s was the first of many failures in single-species fisheries management that suggested the fallacy of this assumption (Larkin, 1977).
As many fish stocks around the world since then to investigate the physical and biological processes affecting survivorship during the early life history stages of fish (Kendall and Duker, 1998;Govoni, 2005).Since the 1980s, advances in

" "
models that can be used to study the ecology of larval and juvenile fishes in their advective environment (Miller, 2007;Werner et al., 2007).Although that can be drawn from these studies (Gawarkiewicz et al., 2007).

receNt techNoloGIcAl AdVANces
Recent advances in animal tagging and tracking, using electronic tags and miniaturized sensor technologies, have introduced new capabilities for studying the ecological dynamics of marine animals in the wild.These technological advances, which have enabled scientists to track individual animals over vast areas of the ocean (Kitigawa et al., 2000;Gunn and Block, 2001;Sibert and Nielson, 2001;Wilson et al., 2001;Block et al., 2003;Welch, 2003;Block, 2005;Weng et al., 2005) while simultaneously sensing their physiological and local environmental conditions (Block et al., 2005;Biuw et al., 2007) 2003).The animals themselves will play an increasingly important role in this global ocean-observing effort as the amount of environmental data collected by tagged animals (Costa, 1993;Boehlert et al., 2001;Block et al., 2003;Biuw et al., 2007;Laidre and Heide-Jorgensen, 2007;Costa et al., 2008) begins to rival that collected by manmade, autonomous and Lagrangian platforms, such as autonomous underwater vehicles, drifters, and gliders (Perry and Rudnick, 2003).Models that can assimilate the four-dimensional data derived from satellites and ocean-observing systems will enable scientists to characterize the ocean's environmental variability in a manner few could have imagined only a decade ago.Although the ocean will continue to be undersampled in both space and time for many years to come, data-assimilation methods are allowing scientists to steadily bridge the gap between actual environmental conditions in the real world and those simulated in models of marine ecosystems (Rothstein et al., 2006).
Developing more sophisticated models that incorporate the movements of individual animals in their time-varying, three-dimensional environments will be vital to addressing many of the needs in ecosystem-based management.

FIsh tAles FroM tWo oceANs
To illustrate some of the recent advances   the implanted tags also record internal temperatures measured inside the tuna's peritoneal cavity (Block et al., 1998(Block et al., , 2001;;Teo et al., 2007b).Implanted tags can collect data for multiple years, but must be recovered from fish recaptured by the tuna fishery in order to obtain the archived data.A generous reward to fisherman for returning tags has resulted in over 20% of them being recovered to date (Block et al., 2005;Kurota et al., in press).In contrast to the implanted tags, PATs can collect data for one year, but do not require that fish be recaptured.

tag-A-Giant project
Instead, each PAT is programmed to release at a set time via electrolysis of a metal wire attaching the tag to a dart embedded in the fish.Once released, the PAT floats to the surface where it transmits a summary of its archived data set to the ARGOS satellite system.
The ambient light and external temperature data from both types of archival tags are used to estimate geolocations, which in turn are used to reconstruct daily movement patterns (Teo et al., 2004;Nielsen et al., 2006).Longitudes and latitudes are estimated by using the light data, collected to reconstruct daily sunrise and sunset curves, in   (Carlsson et al., 2007;Boustany et al., 2008) and otolith (Rooker et al., 2008) studies to develop spatially and temporally explicit models for better assessing the numerical abundances, residence patterns, and mixing of the two stocks in each management unit (Kurota et al., in press).This line of research provides a clear example of how the results from fish tracking studies can have direct implications for managing international fisheries for highly migratory species and setting policy on an ocean-basin scale.
In addition to providing ambient light data for establishing migratory movements, electronic tags also archive internal and environmental data that are useful for inferring other aspects of tuna behavioral and physiological ecology (Block, 2005;Teo et al., 2007aTeo et al., , 2007b)).exploit the more abundant prey found there (Figure 3b).The unique cardiovascular system of bluefin tuna enables them to exploit foraging opportunities that would otherwise be thermally off limits to most subtropical predators (Blank et al., 2004).
Although Recent physiological studies show that the cardiac demands on bluefin tuna can lead to thermal stress at the higher temperatures typical of their breeding areas (Castilho et al., 2007;Teo et al., 2007a).The tuna exhibit deep-diving behavior in the Gulf, presumably for thermoregulatory cooling, prior to and after spawning.However, tagging data reveal that this behavior is altered during the spawning phase, with shallower dives occurring throughout the day and actual spawning thought to occur in the warm surface waters at night (Teo et al., 2007a(Teo et al., , 2007b)).Although there are clearly adaptive advantages to spawning at the surface, the potential for thermal stress may constrain the length of the spawning period as well as overall residence times in the Gulf.
Recent analyses of TAG tracking and satellite data sets demonstrate that certain remote-sensing data products, especially SST and sea surface height, can be used to predict bluefin breeding areas (Teo et al., 2007b).The combination of from climate-associated regime shifts (Mantua et al., 1996) to anthropogenic impacts ranging from dams to logging to pollution to overfishing (NRC, 1996;Ruckelshaus et al., 2002).
Prior to implementation of the permanent, year-round array, the POST project conducted studies with the pilot  et al., 2007, 2008;Melnychuk et al., 2007;Cooke et al., 2008;Welch et al., 2008).
One fundamental, but currently unresolved, question is whether juvenile salmon mortality is of greater importance in freshwater or in the marine environment.The conventional wisdom among most salmon biologists and managers has been that juvenile survival is greater in the marine environment than in freshwater, and therefore conservation efforts ignoring the role of the ocean can be successful.However, results from the POST pilot studies indicate that, although there were some differences among species and stocks and between years, survival estimates for tagged juveniles in freshwater were typically higher than those observed in the marine environment.This surprising finding, which runs counter to the conventional wisdom, suggests that some of our most basic assumptions about the ecology of salmon may need to be reconsidered.Fraser River, which has no dams.The results demonstrated that overall survival out of the Snake-Columbia River system, despite its dams, was comparable (Figure 6).Furthermore, when the survival estimates were scaled by distance traveled or time spent in each environment, survival was actually higher in the Snake-Columbia River system.Perhaps even more surprising, scaled survival estimates were actually higher for both species in the dammed stretch of the Snake-Columbia River system than in the Fraser River.These findings should raise important questions among fisheries managers about why endangered Snake-Columbia River stocks have such poor returns from the ocean despite reasonable juvenile survival when migrating out to sea.Prior to the POST project, it had been assumed that the poor returns in these stocks were a consequence of dam-associated mortality during the juvenile downstream migration (Welch has not been realized (Muir et al., 2006).
This finding was used to argue that barg-  b) survival estimates as above but scaled by migration day (1/t).survival probabilities in the upper and lower river were estimated with the cormack Jolly seber mark-recapture model, which jointly estimates survival and detection probabilities.This model could not be used to estimate coastal shelf survival probability directly due to limited detections in Alaska.Therefore, we estimated the long-shelf survival probability using the following three assumptions: (1) that the detection probability of the lippy point subarray was equal to the mean detection probability (MeAN) observed at Willapa bay (where detection efficiency could be estimated accurately), (2) that the detection probability of the lippy point subarray was equal to the lower 95% confidence limit on detection probability (MIN) observed at Willapa bay, and (3) that the detection probability of the lippy point subarray was equal to the upper 95% confidence limit on detection probability (MAx) observed at Willapa bay.Irrespective of the detection probability assumed, the conventional wisdom that survival per day is better in the coastal ocean than in freshwater is not supported.Figure 5 shows the locations of the lippy point and Willapa bay subarrays.(see recent work of author rechisky and colleagues for further details.) s c I e N c e A N d p o l I c y F e At u r e Advances in conservation oceanography New tagging and tracking technologies and Their potential for transforming the science underlying Fisheries Management b y c h A r l e s h .G r e e N e , b A r b A r A A .b l o c k , d AV I d W e l c h , G e o r G e J A c k s o N , G A r e t h l .l AW s o N , A N d e r I N l .r e c h I s k y technological capacity to monitor marine living resources in their natural environment.Conservation oceanography can provide the new concepts and tools for such a framework, enabling scientists to quantify and forecast the behavioral, physiological, and ecological responses of fish populations to changes in their ocean environment.The technological bases for this transformation can be found in recent advances in miniaturized physiological and environmental sensors, animal tracking techniques, ocean-observing systems, and computer models that assimilate biological and environmental data rapidly after collection (Sibert computational methods have enabled scientists to develop individual-based coNserVAtIoN oceANoGrAphy Is AN eMerGING FIeld oF scIeNce thAt INcorporAtes the lAtest AdVANces IN oceAN scIeNce ANd techNoloGy to proVIde resource MANAGers ANd polIcyMAkers WIth the INForMAtIoN they Need to eNsure the sustAINAbIlIty oF the MArINe eNVIroNMeNt ANd Its lIVING resources.
such individual-based models are capable of simulating and exploring the range of theoretical possibilities arising from different scenarios of interacting physical and biological processes, the limited ability of scientists to track the movements as well as behavioral and physiological responses of individual animals in situ has constrained the conclusions in conservation oceanography, we describe two ongoing projects, Tag-A-Giant (TAG) in the North Atlantic and Pacific Ocean Shelf Tracking (POST) in the North Pacific.The TAG project focuses on tagging and tracking northern bluefin tuna (Thunnus thynnus), while the POST project focuses primarily, but not exclusively, on tagging and tracking a variety of Pacific salmonid species (Oncorhynchus spp.).
1982, this species continues to be harvested at high rates throughout the North Atlantic despite dramatic declines in the adult breeding populations on both sides of the basin.The Atlantic bluefin tuna fishery is regulated by the International Commission for the Conservation of Atlantic Tunas (ICCAT).ICCAT manages bluefin tuna as two separate management units, one in the western Atlantic with a breeding area in the Gulf of Mexico, and the other in the eastern Atlantic with a breeding area in the Mediterranean Sea.The western Atlantic Gulf of Mexico breeding population has declined by over 80%, which led to a reduction in fishing quotas off North America beginning in 1981 and the implementation of a fishery recovery plan in 1998.To date, the population has not recovered, and calls have arisen for listing the Atlantic bluefin tuna as an endangered species and placing a moratorium on its harvest (Safina and Klinger, 2008).Quotas in the eastern Atlantic have been considerably higher, with actual catches far exceeding allocated quotas.No credible recovery plan has been implemented in the eastern Atlantic to assure the longterm survival of this species.The TAG project uses a combination of surgically implanted archival tags and externally attached pop-up satellite archival tags (PATs) to study the movements, behavior, population structure, and physiological ecology of Atlantic bluefin tuna (Figure 1a-d).Both types of tags sense and archive the pressure, external temperature, and ambient light experienced by individual bluefin tuna;

Figure 1 .
Figure 1.(a) Implantable archival tag.(b) pop-up satellite archival tag (pAt) used in the study of wide-ranging pelagic vertebrates.(c) surgical implantation of archival tag in a bluefin tuna.(d) bluefin tuna swimming in situ with pAt tag attached.

Figure 2 .
Figure 2. (a) positions of bluefin tuna classified as western breeders tagged at three western Atlantic locations during 1996-2008.orange circles represent locations based on deployment positions, light-and sst-based geolocation estimates, and pAt tag satellite endpoint positions.yellow triangles represent recapture locations of western Atlantic-tagged fish.(b) positions of bluefin tuna classified as eastern breeders tagged at three western Atlantic locations during 1996-2008.White circles represent locations, determined as before, with green triangles representing recapture locations.
Figure 3. (a) track for an Atlantic bluefin tuna fitted with an archival tag during January 2003 and recaptured during december 2004.(b) Ambient water temperatures experienced by the fish during the January-August 2003 period as a function of depth and calendar day.Vertical bars show daily tag-collected temperature-depth profiles, with the bottom of each bar indicating the maximum depth reached that day.during the early portion of the deployment, the fish occupied warm and well-mixed waters near the North carolina tagging grounds and farther south.during late March, the fish moved northward and spent late April-June in the vicinity of the north wall of the Gulf stream, in association with the 15-20°c sst isotherms.during this period, the temperature-depth record also shows evidence of the fish making excursions of one or more days into cold waters beyond the 15°c sst isotherm (examples of such excursions are indicated by arrows), before returning to warmer waters.during late June, the fish moved south into the warmer waters of the Gulf stream, and then it moved up into the highly stratified waters of the Gulf of Maine during July.
physiological specializations may permit exploitation of a broader thermal niche, thereby providing bluefin tuna with obvious advantages for foraging, such enhancements may not be sufficient to buffer them from future climate-associated changes in oceanographic conditions.It has been reported that natural modes of climate variability, such as the North Atlantic Oscillation (NAO), are associated with shifts in the location where the Gulf Stream separates from the continental shelf and becomes a freely meandering jet (Gangopadhay et al., 1992).Furthermore, changes in the trajectory of the Gulf Stream's north wall and in the hydrographic properties of the slope water sea and Gulf of Maine are also associated with modal shifts in the NAO (Taylor and Stephens, 1998; MERCINA, 2001).Under certain climate-change scenarios, the separation point and trajectory of the Gulf Stream's north wall are predicted to shift to the south of their climatological mean positions.Under other scenarios, changes in Arctic climate are expected to alter circulation patterns in the Northwest Atlantic, leading to a cooling and freshening of shelf and slope waters from the tail of the Grand Banks to the Gulf of Maine (Greene et al., 2008).Such changes could alter prey distributions or place foraging grounds currently exploited by bluefin tuna outside of their thermal tolerance range.With climate-associated constraints on their ecology, bluefin tuna might be expected to alter their migration patterns and use of important foraging areas.Stock assessment methods that rely on repeated sampling year after year at specific locations would be compromised by such changes in bluefin tuna distributions.In addition to foraging excursions into colder environments that test the lower limits of their thermal niche, bluefin tuna also enter warmer environments to spawn that test their upper thermal limits.During their spring breeding season in the Gulf of Mexico, spawning adult tuna seek out relatively warm breeding areas (SST ≥ 24° C), which are typically found in the slope waters and frontal regions of mesoscale eddies spun off the Loop Current(Teo et al.,   2007a,b).When these large, endothermic fish, which can weigh over 600 kg, enter the Gulf, they face thermoregulatory challenges much greater than those experienced in the North Atlantic.
TAG and remote-sensing data are being used to develop models for forecasting the locations of breeding areas, spawning-period lengths under different environmental conditions, and overall residence times in the Gulf (recent work of author Block and Steven Teo, University of California, Davis).Such forecasting models have the potential to contribute to both operational and impactassessment conservation oceanography.throuGhout the hIstory oF huMAN cIVIlIzA-tIoN, socIety hAs NeVer posed As GreAt A threAt As It does NoW to the sustAINAbIlIty oF lIVING resources IN the seA." " to fishing mortality later in life and few other anthropogenic threats earlier on, salmon run a gauntlet of anthropogenic challenges throughout their lifetimes.Threats to salmon survival in freshwater involve primarily habitat alteration and destruction, including the damming of rivers for hydroelectric power generation.Hydroelectric dams can impede the downstream migration of juveniles to the coastal ocean as well as the upstream return of mature fish to their spawning grounds.The ecology of salmon in the ocean is poorly understood (Brodeur et al., 2000), and there is considerable controversy about which is more important to recruitment variability, mortality in freshwater or mortality in the marine environment.Finding the answer to this question has critical policy implications as two decades of declining salmon runs in California, Oregon, and Washington have been blamed on everything

By 2010 ,
POST plans to deploy a permanent array of cross-shelf receiver lines at strategic locations along the North American west coast from California to the Aleutian Islands.The pilot array has been deployed each year since 2004 and extends over 2,500 km along the coasts

Figure 4 .
Figure 4. (a) relative sizes of acoustic tags and a juvenile salmon.(b) surgical implantation of acoustic tag.
array from 2004 to 2006 to establish both the technical performance of a largescale telemetry system and to document freshwater and marine migration and survival patterns of both juvenile and adult salmon.Results from these pilot studies provided the first opportunity to compare relative freshwater and marine survival rates of juveniles and returning adults from a number of different Pacific salmon species and stocks (Crossin If the conventional wisdom about juvenile survival in freshwater is called into question, then what other surprises might the POST telemetry system reveal?Welch et al. (2008) recently reported results from a study comparing juvenile survival of chinook and steelhead out of the Snake-Columbia River system, which has eight dams, with juvenile survival of these same species out of the

Figure 5 .
Figure 5.The current post (pacific ocean shelf tracking) acoustic array extending over 2,500 km along the coasts of southeast Alaska, british columbia, Washington, and oregon.

Figure 6 .
Figure 6.survival estimates for Thompson and snake river spring chinook and steelhead.(a) estimated survival.(b) survival scaled by100 km traveled (100/l).(c) survival scaled by migration day (1/t).For each species, the left panel shows the survival of different Thompson river stocks released to migrate down to the Fraser river mouth; the right panel shows the survival of snake river stocks migrating down three sections of the columbia river hydropower system: impounded (upper river; eight dams), unimpounded (lower river; undammed), and entire river.The vertical bar to the right of each group of individual survival estimates (circles) shows the mean survival and 95% confidence interval for the group, averaged across all available data.The blue band and dashed red line shows the mean and 95% confidence interval for the Fraser river survival in comparison to the columbia.In all survival comparisons, snake-columbia river estimates were significantly different from Thompson-Fraser estimates at the 95% confidence level unless indicated by Ns.For further details, see the original Figure3fromWelch et al. (2008), from which this figure was modified.because the post array was deployed only in the lower Fraser river, it is not possible to break out survival by river section, as can be done for the columbia river data.

Figure 7 .
Figure7.(a) 2006 survival probability estimates in the dammed upper portion of the columbia river, the undammed lower portion of the river, and the coastal shelf.(b) survival estimates as above but scaled by migration day (1/t).survival probabilities in the upper and lower river were estimated with the cormack Jolly seber mark-recapture model, which jointly estimates survival and detection probabilities.This model could not be used to estimate coastal shelf survival probability directly due to limited detections in Alaska.Therefore, we estimated the long-shelf survival probability using the following three assumptions: (1) that the detection probability of the lippy point subarray was equal to the mean detection probability (MeAN) observed at Willapa bay (where detection efficiency could be estimated accurately), (2) that the detection probability of the lippy point subarray was equal to the lower 95% confidence limit on detection probability (MIN) observed at Willapa bay, and (3) that the detection probability of the lippy point subarray was equal to the upper 95% confidence limit on detection probability (MAx) observed at Willapa bay.Irrespective of the detection probability assumed, the conventional wisdom that survival per day is better in the coastal ocean than in freshwater is not supported.Figure5shows the locations of the lippy point and Willapa bay subarrays.(see recent work of author rechisky and colleagues for further details.)