Oceanic Ecosystem Time-series Programs Ten Lessons Learned

104 another hot day in hawaii. shown are a few of the many instruments used to collect samples during monthly hawaii ocean time-series (hot) cruises to station aloha. this page. Conductivity-temperature-Depth (CtD) rosette sampler equipped with 24 sampling bottles and additional sensors for oxygen, chlorophyll a, and nitrate. NeXt page, from top to bottom. free-drifting sediment traps, recovery of bottom moored instrumentation, recovery of spar buoy, recovery of in situ incubation experiment, deployment of bottom moored mclane sediment trap. Photographs by Paul Lethaby iNtroDUCtioN Since its creation within UNESCO a half-century ago, the Intergovernmental Oceanographic Commission (IOC) has been at the vanguard of ocean observation, serving to promote international cooperation, coordinate ocean research, and facilitate capacity development. Beginning with the International Indian Ocean Expedition in the early 1960s, and through meaningful partnerships with (POGO), and related organizations, IOC has provided invaluable leadership needed to help justify and promote large-scale ocean observation programs. A recent inter-than 600 participants from 36 nations to present and discuss ongoing and planned global ocean observation activities. These field efforts represented a diverse spectrum of time-series programs, including the use of satellite remote sensing, moored buoys, autonomous gliders, repeat hydrographic surveys, volunteer ships of opportunity, profiling floats, cabled seafloor observatories , and ship-supported time-series programs, to name a few examples. Each observation program is designed to address a specific set of scientific goals, and each has its own set of challenges to sustain and optimize data return. This article focuses on ecosystem-based, time-series programs that presently rely on ships to make observations , collect samples, and conduct experiments. These ecosystem investigations are an important subset of the much larger portfolio of research-based, ocean time-series programs that derive, in large part, from sustained IOC leadership (Valdés et al., 2010). 106 example, the formation of the ocean basins, the transition from a reducing to an oxidizing environment, and the rise of multicellular life. Superimposed on these long-term changes are higher-frequency variations, including epochs, regime shifts, cycles, stochastic habitat variability, and, in recent times, human-induced climate change. Based on results from the few long-term ecosystem studies that do exist, there is ample evidence that the ocean varies on a number of time scales, including multidecadal cycles as well as secular change. Long-term records are required to sort out the various signals from the noise (Overland et al., 2006) and the relevant time scales of interest will vary depending upon …

This article focuses on ecosystem-based, time-series programs that presently rely on ships to make observations, collect samples, and conduct experiments.These ecosystem investigations are an important subset of the much larger portfolio of research-based, ocean timeseries programs that derive, in large part, from sustained IOC leadership (Valdés et al., 2010).
example, the formation of the ocean basins, the transition from a reducing to an oxidizing environment, and the rise of multicellular life.Superimposed on these long-term changes are higherfrequency variations, including epochs, regime shifts, cycles, stochastic habitat variability, and, in recent times, humaninduced climate change.
Based on results from the few longterm ecosystem studies that do exist, there is ample evidence that the ocean varies on a number of time scales, including multidecadal cycles as well as secular change.Long-term records are required to sort out the various signals from the noise (Overland et al., 2006) and the relevant time scales of interest will vary depending upon the process under investigation.For example, habitat and ecosystem variability on time scales ranging from decades to millennia, or longer, can be obtained from natural archives, including deep-sea sediments, ice cores, and molecular biomarkers such as genes and genomes, to name a few resources.Instrument records of habitat variability are also invaluable for developing a comprehensive understanding, but cover a much shorter time period, at most a few centuries but more commonly years to a few decades.
In addition to temporal change, ecologists and oceanographers also need to measure and understand the spatial structure of the habitat and its inhabitants (Swanson and Sparks, 1990).Contemporary spatial variability is probably more easily documented, studied, and understood than temporal variability.Indeed, the differences are akin to viewing a snapshot versus a full-length motion picture.However, spatial and temporal variability are inextricably coupled (Stommel, 1963; Figure 1), even though often viewed as independent characteristics.The relevant scales for understanding ecosystem processes range from meters to hundreds of kilometers in space, and from hours to centuries in time.
Ocean surveys provide comprehensive, nearly synoptic maps of the distributions of properties and organisms, but are unable to reveal temporal changes unless they are repeated time and again.
On the other hand, observations made at geostationary ocean stations can resolve temporal variability over local scales, but cannot easily be extrapolated to regional or biome scales.Recently, it has been reported that the global extent of low-productivity oceanic regions is expanding (Polovina et al., 2008), so spatial scaling of marine ecosystems may also have a key temporal component.

The design of the time-series program including both spatial coverage and the arrow of time
The accretion and subduction of oceanic plates, the rise and fall of sea level, the evolution and extinction of species, El Niño-La Niña climate oscillations, the vernal blooming of phytoplankton, and diel vertical migrations of mesozooplankton all share the common element of time, albeit on very different scales.Time is so fundamental to our understanding of Earth system processes that we sometimes take this important variable for granted, or worse, ignore it, in the development of conceptual models of how ocean habitats are structured and how they function.One undeniable fact of oceanic ecosystems is that they are complex, time variable, nonsteady state, nonlinear features, and need to be studied as such.However, chronic undersampling is a fact of life in oceanography (Platt et al., 1989) and still constrains the interpretation of available field data.Indeed, there are many examples in the scientific literature where interpretations from short-term ecological studies are at odds with similar data sets collected over much longer time scales (Strayer et al., 1986).It is difficult to observe slow or abrupt environmental changes directly, much less to understand the fundamental cause-and-effect relationships of these changes.As data accumulate in a long-term ecological study, new phenomena become apparent and new understanding is achieved.
Much of the temporal change that has occurred in oceanic ecosystems over the 4.5-billion-year history of Earth can be characterized as directional change-for David m. Karl (dkarl@hawaii.edu)(1984)(1985)(1986)(1987)(1988) and the Canadian Joint Global Flux Study (JGOFS) program (1992-1997).for climate studies due to its proximity to the site of bottom water formation and its role in sustaining global thermohaline circulation (Østerhus et al., 2009).Initially, only temperature and salinity data were collected, but later dissolved oxygen, nutrients, and chlorophyll a (chl a) concentrations were added as core parameters.
In 3 Lesson 3: Due to the existence of historical, long-term data and the independent collections of contemporary "background" hydrographical and biogeochemical data, there is a logical and meaningful attraction to these special time-series sites of scientific importance.In essence, if you build them, they will come, and PAPA is an excellent example.
4  (1969,1972,1975,1978,1981).Unfortunately, and tragically, this sampling design imposed serious limitation on data analysis and on the detection of climate-related impacts on ecosystem dynamics (Chelton et al., 1982) (Franklin et al., 1990).The There is an enormous need for standard reference materials (SRM) that can be used for analytical comparability among otherwise independent time-series measurement programs.The advent of reference materials for dissolved inorganic (Dickson et al., 2003) and organic (Hansell, 2005) (Allen and Harvey, 1928).
In 1894, E.J. Allen was appointed the director of the laboratory, and shortly thereafter-in 1896-an 18.5-m vessel was purchased and outfitted to support local fieldwork (Southward and Roberts, 1987).In addition to his leadership of the laboratory, Allen also conducted research on marine phytoplankton.
Among other achievements, Allen and his colleagues were able to establish laboratory cultures of selected phytoplankton that provided an opportunity to study their growth requirements, growth rates, and related functions.in seawater used a pre-concentration step that involved co-precipitation with iron hydroxide (Matthews, 1916(Matthews, , 1917)), a predecessor of the modern "MAGIC" method (Karl and Tien, 1992) that was introduced to oceanography 75 years later.
In a series of publications on "phos- to study many variables related to the fertility of the sea, including the lability of dissolved organic matter (Atkins, 1923), the importance of thermal stratification and mixing for phytoplankton growth (Atkins, 1924a,b), and the proton balance in relation to net photosynthesis (Atkins, 1922a,b).The plankton production in the sea (Mills, 1989).Coincident with the re-establishment of the time-series program, Atkins prepared a two-part synthesis of what was known at that time about the controls of phytoplankton growth in the sea, including both physical (Atkins, 1926a) and chemical (Atkins, 1926b) factors.This body of extant knowledge been observed previously by Matthews (Atkins, 1926c(Atkins, , 1930; Figure 2), the first ever time series of nitrate using a novel colorimetric method that had been perfected for seawater analysis (Harvey, 1926(Harvey, , 1928)), observations on the penetration of sunlight into seawater using a new photoelectric instrument (Poole and Atkins, 1937), and use of a novel silk net sampler to concentrate and quantify the amount of phytoplankton in the water column (Harvey, 1934).These time-series data, combined with the routine measurements of water temperature and salinity, provided a foundation for understanding plankton production and its control.This new knowledge, now part of the enduring legacy of biological oceanography, was reported in an article by the same title, "Plankton Production and Its Control" (Harvey et al., 1935), and later in key review articles (Harvey, 1942(Harvey, , 1950) ) and in an influential monograph (Harvey, 1955).It of DIP in the water column (Cooper, 1938;Harvey, 1942).This observation not only suggested a strong link between nutrient delivery-which was ultimately tied to ocean circulation-and organic matter production at higher trophic levels, but also revealed large year-toyear variations in ecosystem structure and function (Figure 3).(Russell et al., 1971;Cushing and Dickson, 1976;Southward, 1974Southward, , 1980Southward, , 1995) ) [1969][1970][1971][1972][1973][1974][1975][1976][1977] near Station E1 (Butler et al., 1979).
Although During the interim period (1988)(1989)(1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002) several key, short-term time-series studies were conducted.For example, in May 1997, the Plymouth Marine Bio-Optical Databuoy (PlyMBODy) was deployed in the western English Channel to serve as a local, vicarious validation of the NASA SeaWiFS ocean color satellite (Pinkerton et al., 2003).Data from this buoy provided an invaluable time-space assessment of phytoplankton dynamics that improved interpretations of data collected at Station E1.Another specialized study included a 15-month plankton time series (viruses to mesozooplankton) to provide data to inform conceptual trophodynamic models, including the microbial loop component (Rodriguez et al., 2000).Finally, a detailed study of the annual cycle (2001) of photosynthetic quantum efficiency and bio-optical characteristics of the western English Channel provided new insights into the process and controls of phytoplankton production (Aiken et al., 2004).
A long-lasting legacy of this, and all other, oceanic ecosystem time-series programs is the data and any sample archives that result from the intensive field efforts.If properly managed and maintained, they can be readily used by future investigators.An excellent example is the re-analysis of Station E1 nutrient data (Jordan and Joint, 1998).
A summary of over 1000 paired nitrate   (1969)(1970)(1971)(1972)(1973)(1974)(1975)(1976)(1977).From Butler et al. (1979)   the North Pacific Ocean (Haury et al., 1994;Karl and Tien, 1997) and Walker, 1985).This great age and relative isolation were primary factors leading to the "climax community" hypothesis: an ecosystem in its final stages of succession is time and space invariable.A test of this general hypothesis was the primary motivation behind the establishment of a multiyear observational program centered near 28°N, 155°W in an area dubbed the "CLIMAX region" (Venrick, 1990).However, once thought to be a homogeneous and static habitat, there is now increasing evidence,  (Hayward, 1987).
Despite the fact that this extensive time series was biased with respect to season (70% of the cruises were in summer, June-September, and 35% in August alone) and was discontinuous over the 17-year period of observation (there were no cruises in 1970, 1975, 1978-79, 1981, or 1984) of a coordinated international effort (Karl et al., 2003).
The Hawaii Ocean Time-series (HOT) program's deep ocean station was established in 1988 to address WOCE and JGOFS scientific goals (Karl and Winn, 1991;Karl and Lukas, 1996).However, no direct measurements of nutrient loading were presented.
A subsequent study, which also included a decade of HOT observations, confirmed the high chl a concentrations and provided evidence for decreases in silicate and phosphate; primary production was also much higher than in the period prior to approximately 1980 (Karl et al., 2001).This observation period coincided with improvements in the measurement of in situ primary production, especially the use of tracemetal-free incubation techniques, so it is difficult to separate methodological changes from habitat differences.However, the observed increases in chl a are independent of the primary 9 Lesson 9: Time-series measurement programs should use consistent methods for sample analysis until it can be shown that any "new and improved" methodology provides similar, if not identical, results.Any deviations in protocol, instrumentation, standardization, or data analysis need to be carefully documented and made available to data users.If possible or practical, a sample archive should be established.
production measurements, so at the very least one can conclude that there have been changes in phytoplankton biomass, efficiency of photosynthesis, or both.Karl et al. (2001) hypothesized that there had been a "domain shift" in the photoautotrophic communities from a eukaryote-dominated ecosystem to one dominated by Prochlorococcus.This fundamental change would have significant ecological implications, including altered food web structure and changes in new, export, and fish production (Karl, 1999;Karl et al., 2001).
Like all other complex processes in nature, a single climate indicator such as PDO, or any other one, may not be adequate to characterize climate variability (Bond et al., 2003).Corno et al. (2007) and Saba et al. (2010) reported significant two-decade-long changes in NPSG ecosystem processes, including an approximately 30-50% increase in primary production, and variations in mixed-layer depth, nutrient fluxes, microbial abundances, and pigment inventories.These variations were attributed to the magnitude, duration, and synchrony of the El Niño/Southern Oscillation (ENSO) and PDO (Corno et al., 2007).Because these two climate  Karl et al. (2001) phosphorus, and its relationship to the coupled cycles of carbon, nitrogen, and, possibly, iron (Karl et al., 2001(Karl et al., , 2003;;Karl, 2002Karl, , 2007b)) (Karl, 1999(Karl, , 2002)).If true, this situation could eventually lead to a 20-to 30-year cycle of N vs.
Science is a world of ideas, and progress in science is limited ultimately by the emergence of new hypotheses and by the ability to test them (Pomeroy, 1981).
In a thoughtful review entitled "The
Partnerships between modelers and data originators have proven to be of great shifts to predict climate impacts on ecosystem processes (Carpenter et al., 2008;Biggs et al., 2009).Warning signs, including wide swings in ecosystem dynamics, slower return rates following physical perturbations, and substantial changes in standard deviations of key variables, may hold predictive value for future ecosystem change (Carpenter et al., 2008).While this exciting research holds great promise and obvious soci- to facilitate this important mission (Glover et al., 2010).
Currently, there is a critical need to aCkNowleDgemeNts

iNtroDUCtioN
Since its creation within UNESCO a half-century ago, the Intergovernmental Oceanographic Commission (IOC) has been at the vanguard of ocean observation, serving to promote international cooperation, coordinate ocean research, and facilitate capacity development.Beginning with the International Indian Ocean Expedition in the early 1960s, and through meaningful partnerships with the Scientific Committee of Ocean Research (SCOR), the International Geosphere-Biosphere Program (IGBP), the Partnership for Observation of the Global Ocean (POGO), and related organizations, IOC has provided invaluable leadership needed to help justify and promote large-scale ocean observation programs.A recent international meeting, co-sponsored by IOC, OceanObs'09: Ocean Information for Society -Sustaining the Benefits, Realizing the Potentials, was held in Venice, Italy, in September 2009.The conference was attended by more than 600 participants from 36 nations to present and discuss ongoing and planned global ocean observation activities.These field efforts represented a diverse spectrum of time-series programs, including the use of satellite remote sensing, moored buoys, autonomous gliders, repeat hydrographic surveys, volunteer ships of opportunity, profiling floats, cabled seafloor observatories, and ship-supported time-series programs, to name a few examples.Each observation program is designed to address a specific set of scientific goals, and each has its own set of challenges to sustain and optimize data return.

Following
Canada, continued to provide leadership and funding to sustain three to four cruises per annum along "Line P" (a transect of stations from the coast to the open sea) and at Station PAPA.This Line P/PAPA time-series program has attracted numerous international partners, and has grown into a rich multidisciplinary enterprise.The second major root of oceanic ecosystem time-series programs was the need to understand and manage commercial fisheries.In the United States, Spencer F. Baird helped convince Congress to establish the US Commission of Fish and Fisheries in 1871.He was later appointed the inaugural Commissioner and established a laboratory in Woods Hole, MA.The 72-m, steel-built steamer, Albatross, became the first US government oceanographic research vessel.In 1903, the organization became part of the Department of Commerce, and after the National Oceanic and Atmospheric Administration (NOAA) was created in the early 1970s, it was renamed the National Marine Fisheries Service (NMFS).As part of the federal government, this organization conducts research on life histories, physiology, breeding, and ecology of many commercially important fish and shellfish, including stock assessments.
-see Vs. see-aND-DoSome have argued that time-series science is akin to routine monitoring (i.e., "look-and-see") and has no place in the portfolio of scholarly academic activities.Even the pioneering, and now invaluable, efforts of Charles D. Keeling's atmospheric CO 2 measurement program at the Mauna Loa Observatory, Hawaii, were criticized by the science funding agencies early on as routine monitoring that "strayed from basic science"(Keeling, 1998).I strongly disagree with this assessment; the legacy and scientific value of Keeling's time series and related programs has already been clearly demonstrated.There is, however, a fundamental, but perhaps subtle, difference between monitoring and observing.Both imply serial measurement; however, the latter is designed-at least in science-to lead to hypothesis generation, prediction, and experimentation (i.e., "see-and-do").Whereas both types of time-series programs capture environmental variability and change, only the latter has a required mission for new scientific understanding.The use of time-series data to generate testable hypotheses demands that the measurement program be properly designed and that the observations are analytically consistent, accurate, and relevant.A hallmark of most oceanic ecosystem time-series programs is a rigorous hypothesis-testing component, either built in from the start or added later as the novel data sets are made available.In this way, time-series programs contribute greatly to the scientific method of creating new knowledge and lead to better understanding and prediction.The unique value of a time-series observation program is the ability to detect change (stochastic disturbance, cyclic behavior, long-term trends) in one or more of the measured parameters.The main mission is then to interpret that change in the context of existing models of how we believe ocean ecosystems are structured and how they function (or to improve them), including model predictions of how ecosystems might change in the future.To succeed in this mission, the measured parameters must be carefully selected and well calibrated, using consistent methodology and instrumentation.Those individuals who are committed to time-series observations are constantly challenged with external criticisms regarding the sometimes limited scope of core measurements, the reluctance of the lead investigators to change sampling or measurement protocols in spite of new advances in technology, and an intellectual struggle between a desire to chart the course of oceanic variability and an equally strong desire to understand the underlying cause of the observed temporal change.The struggle between observation and understanding has recently been highlighted with the marine genomics revolution (DeLong and Karl, 2005), which has led to a re-evaluation of some of the most basic principles of ecology and evolution.Because this new information is rapidly changing the way we view the most fundamental processes of solar energy capture and dissipation, the probable impacts of climate-induced changes to the oceanic carbon cycle, and the very nature of life in the sea, it is leading to paradigm shifts in marine ecology and biogeochemistry such that existing ocean observation programs may not be well positioned to observe these novel and arguably important properties.In many ways, the microbial genomics revolution is one of the most exciting aspects of contemporary oceanography-but, at the same time, the worst nightmare for ocean time-series managers and those who use these key data sets for global ocean modeling and prediction (see Role of Oceanic Ecosystem Modeling section).he would later go on to serve as the Director of the Laboratory for two decades (1945-1965), and in 1965 he was knighted for his career scientific contributions.The Russell Cycle describes a set of interrelated physical, chemical, and biological features that led to a significant and fairly rapid change in both the species and abundance of fish 5 Lesson 5: The merger of laboratory research on algal growth and fieldwork in the English Channel off Plymouth can be traced to the pioneering measurements of dissolved phosphorus by D.J. Matthews.His improved method for the measurement of low concentrations of dissolved inorganic phosphorus (DIP) research at Plymouth was built on the pioneering work of Victor Hansen and his team, especially Karl Brandt and Hans Lohmann, at Kiel who developed most of our modern concepts regarding 6 Lesson 6: It is important to periodically review overall progress toward the stated goals, to reassess the design of the field sampling program in light of new knowledge, and to disseminate comprehensive, integrated interpretations to the scientific community for their evaluation and comment.Time-series programs should themselves be subject to change, whenever it is necessary to do so.
served as a blueprint to focus the new phase of time-series research in the western English Channel.As they began this second phase of fieldwork, the team also established a survey line of stations from Plymouth to Ushant.These sites were periodically occupied, but eventually the team concentrated their sampling efforts at the International Council for the Exploration of the Sea (ICES) Hydrographic Station England number 1 (E1), located 22 nautical miles southwest of Plymouth at a water depth of 72 m (50°0.2'Nand 4°22'W).Station E1 soon became the proving ground for many new analytical techniques, and once confidence was established, these parameters were added to the list of core measurements to track seasonal and interannual ecosystem variability.A series of very influential papers appeared during the period 1925-1935, including a confirmation of the seasonal variation in DIP that had

7
figure 2. seasonal and interannual variations in mean water column phosphate (expressed as mg p 2 o 5 per m 3 ) at station e1 in the western english Channel for the period1923-1929.From Atkins (1930)  reproduced with permission from Cambridge University Press.
As the length of the record grew and new understanding was gained, the temporal variations in the structure and efficiency of the marine food web leading to fish were explicitly tied to climate change, specifically a general warming.Indeed, this warming trend had initially gone unnoticed in part due to methods used for data and trend analysis.Upon re-analysis, Southward (1960) discovered a significant warming in the period of 1930-1960, especially in near-surface waters.However, the most fundamental knowledge of the relationships between climate and fish would not emerge until there was a reversal in the general warming trend that began in the 1930s and a reversion of the abundance and type of fish; this change began in the early 1960s (Figure 3).The identification and interpretation of the multidecadal cycle in ecosystem structure that emerged as a result of this important measurement program figure 3. The first report(russell, 1971)  of what would later be dubbed "The russell Cycle."This graph shows data on: (•) numbers of young fish, and (ο) winter maxima in concentration of phosphate.FromRussell et al. (1971) reproduced with permission from Nature Publishing Group

(
Apr May Jun Jul Aug Sep Oct Nov Dec

figure 4 .
figure 4. relationships between dissolved inorganic and organic nutrients in the english Channel.The seasonal climatologies are based on observations made during an 11-year study period(1969)(1970)(1971)(1972)(1973)(1974)(1975)(1976)(1977).FromButler et al. (1979) reproduced with permission from Cambridge University Press lower values in summer (Figure 5); the overall mean molar N:P ratio was 11.6, lower than the predicted Cooper Ratio of 15:1 or the Redfield Ratio of 16:1 (Jordan and Joint, 1998).The authors also reported evidence for transient increases in P that did not show up as corresponding increases in N, as might be expected from deep-water intrusions or mixing.Similar near-surface enrichments have been reported for based largely on high-frequency timeseries observations, that the NPSG exhibits substantial physical, chemical, and biological variability on a variety of time scales, from months to decades.The NPSG ecosystem is characterized by a relatively deep permanent pycnocline (and nutricline) and fairly shallow (≤ 100 m) mixed-layer depths throughout the year.Consequently, the mixed layer is chronically nutrient starved, and the near-zero nutrient concentration gradient routinely observed in the upper 100 m of the water column suggests that continuous vertical nutrient flux cannot be the primary source of dissolved inorganic nutrients (e.g., nitrate and phosphate) to the upper euphotic zone (Hayward, 1991).The observed separation of light in the surface waters from inorganic nutrients beneath the euphotic zone predicts environmental conditions of extreme oligotrophy (low standing stocks of nutrients and biomass), low rates of primary production of organic matter, and low rates of carbon export to the deep sea.Because subtropical ocean gyres are dominant habitats of the world's ocean, accurate estimation of global ocean production and export will require reliable estimation of NPSG ecosystem processes.Although the NPSG was sampled during the Challenger Expedition (1872-1876), and several later "voyages of discovery" (e.g., Albatross in 1903, Carnegie in 1928), it was not until after World War II that extensive and systematic ship-based observations were initiated.These projects included the NORPAC expedition, which deployed 19 ships from Canada, Japan, and the United States to survey the entire Pacific Ocean north of 20°N during summer 1955, and Weather Ship Station November (30°N, 140°W), which was occupied during 121 cruises between July 1966 and May 1974.In 1968, the CLIMAX I expedition from Scripps Institution of Oceanography occupied a series of stations near 28°N, 155°W during August and September; CLIMAX II reoccupied the site during September the following year.Thus began an important oceanic ecosystem timeseries study in the NPSG; an additional 18 major cruises would be conducted between 1971 and 1985 , the CLIMAX program observations provided an unprecedented view of ecosystem structure and function.From January 1969 to June 1970, a deep ocean hydrostation (Gollum) was established by scientists at the University of Hawaii at a location 47 km north of Oahu (22°10'N, 158°00'W;Gordon, 1970).At approximately monthly intervals, 13 research cruises were conducted to observe and interpret variations in particulate organic matter distributions in the water column in addition to other chemical and physical parameters(Gordon, 1970).Unfortunately, this pioneering NPSG time series was terminated due to lack of funding and interest to sustain the intensive field effort.With the abandonment of the central North Pacific Ocean Weather Ship and time-series programs such as Gollum, there was no location where comprehensive, seasonally resolved measurements of the ecosystem variability of the NPSG were available.IOC and the World Climate Research Programme (WCRP) Committee on Climate Change in the Ocean (CCCO) recognized this deficiency and, in 1981, endorsed the initiation of new ocean observation programs.Reactivation of Gollum was an explicit Committee recommendation.In response to the growing awareness of the ocean's role in climate and global environmental change, and the need for additional and more comprehensive oceanic time-series measurements, the Board on Ocean Science and Policy of the National Research Council (NRC) sponsored a workshop on "Global Observations and Understanding of the General Circulation of the Oceans" in August 1983.The proceedings of this workshop served as a prospectus for the development of the US component of the World Ocean Circulation Experiment (US-WOCE).Shortly thereafter, in September 1984, NRC's Board on Ocean Science and Policy sponsored a second workshop on "Global Ocean Flux Study, " which served as an eventual blueprint for the JGOFS program.In 1986, the International Council of Scientific Unions (ICSU) established IGBP: A Study of Global Change, and the following year, JGOFS was designated as a core project of IGBP.US-JGOFS research efforts centered on the oceanic carbon cycle, its sensitivity to change, and the regulation of the atmosphereocean CO 2 balance (Brewer et al., 1986; Sabine et al., 2010).In 1988, two open-ocean time-series programs were established-one in the North Atlantic near Bermuda and the other in the North Pacific near Hawaii (Karl and Michaels, 1996).The Bermuda Atlantic Time-series Study (BATS) was an extension of the Hydrostation S program started by Henry Stommel and co-workers in 1954 (dubbed Panularis Station after the beloved 19-m research vessel that was initially used to support the approximately biweekly occupations; Michaels and Knap, 1996).During the more than half-century of ocean timeseries sampling at Bermuda, pioneering research on several important physical, chemical, and biological processes has been supported mostly through logistical and intellectual support of scientists and staff of the Bermuda Biological Station for Research (now Bermuda Institute of Ocean Sciences).Additional JGOFS-sponsored, IOC-inspired biogeochemical time-series programs were eventually established at key locations in the Ligurian Sea (DYFAMED), near Gran Canaria (ESTOC), southwest of Kerguelen Island (KERFIX), northwest of Hokkaido Island (KNOT), and southwest of Taiwan (SEATS) as part The NPSG benchmark location, dubbed Station ALOHA (A Long-term Oligotrophic Habitat Assessment), is located at 22°45'N, 158°W, approximately 100 km north of Oahu, Hawaii, in deep water (4800 m), outside any biogeochemical influence of the Hawaiian Ridge, yet still close enough to facilitate approximately monthly sampling from shore-based facilities (Karl and Lukas, 1996).Now, after 20 years of intensive sampling, the NPSG has become one of the most well-studied open-ocean ecosystems, providing a global reference point for tracking the health of the ocean, including the rate of CO 2 sequestration and ocean acidification (Dore et al., 2009), and an experimental framework for studying seasonal and interannual ecosystem dynamics.One of the major goals of the NPSG time series (1968-present) is to link oceanic ecosystem changes to climate variability.Just as the CLIMAX study was winding down, and prior to the start of HOT, a significant paper was published that reported a major change in the plankton community-and the presumed productivity-in the NPSG.Venrick et al. (1987) reported that the average euphotic zone (0-200 m) chl a concentration in the oligotrophic North Pacific Ocean during summer (May-October) had nearly doubled from 1968 to 1985.The sampling frequency was insufficient to determine whether the chl a increase had been continuous over time or whether there had been a "step-function" increase between 1973 and 1980(Venrick et al., 1987; Figure6).An abrupt shift in climate, beginning in the mid 1970s, appears to be one example of a recurring pattern of interdecadal climate variability referred to as the PDO(Mantua et al., 1997).Venrick et al. (1987) attributed their field observations to the North Pacific regime shift that caused an enhanced nutrient flux, and resulted in a significant long-term change in the carrying capacity of the NPSG ecosystem.
suggested that ecosystem changes in the subtropical North Pacific following the 1976 climate step were a result of increased mixedlayer depths and a higher frequency of deep mixing events due especially to an intensification of the Aleutian low-pressure system in late winter.This vigorous mixing might be expected to enhance nutrient input to the euphotic zone and stimulate ecosystem productivity.However, it is important to emphasize that the Aleutian low-pressure system returned to its normal, pre-1976, condition in 1988 just at the start of the HOT program.The concentrations of chl a and primary production, on the other hand, have remained at the elevated "regimeshift" values (Figure 6).Given the rapid doubling times of microorganisms in the NPSG (one to a few days) and the relatively rapid turnover of particulate organic matter pools (10-20 days), it is unlikely that the oceanic biogeochemical response to climate forcing or relaxation would have a two-decade-long time lag.It is conceivable, even possible, that the NPSG has alternative or multiple stable states and once a new state is established it is resilient to change until some new environmental threshold is achieved figure 6. Composite time-series analysis of phytoplankton community abundance, measured as total euphotic zone integrated chlorophyll a (chl) concentration (mg m -2 ), and productivity, measured using the 14 C-radiotracer technique.The green symbols in the upper plot are chl a derived from hplC analyses and the red symbols are based on fluorometric analyses.The letters in the lower panel represent individual research expeditions or programs: C-i = Climax i, g = gollum, C = Climax time-series, f = fiona, V = VerteX, p = prpoos, a-i = aDios i, a-iii = aDios iii, aloha-hawaii ocean time-series.The data-relevant metadata and data source credits are available from the author.The pre-1988 data were collected in the ClimaX region (see text) and the post-1988 data were collected at station aloha.Redrawn and updated fromKarl et al. (2001) recharge the upper water column with dissolved organic matter and oxygen that support post-bloom heterotrophic metabolism.More importantly, blooms contribute to the seascape mosaic that is essential for maintaining genetic diversity in these expansive habitats.Specifically, mesoscale physical forcing may be an important control on the abundance, diversity, and activity of N 2 -fixing microorganisms and, hence, on the nutrification of the gyre(Church et al., 2009).Even the approximately monthly sampling schedule adopted in the HOT program may be too infrequent to resolve these and related, intermittent physical processes that may impact nutrient budgets in the NPSG(Johnson et al., 2010).In the past few years, the scale and scope of ocean observation at Station ALOHA has been enhanced by several new programs, including: (1) an ocean mooring for meteorological and physical oceanographic observations (WHOTS), (2) deployment of a fleet of APEX profiling floats equipped with sensors to measure oxygen and nitrate concentrations (Riser and Johnson, 2008; Johnson et al., 2010), (3) deployment of a fleet of Seagliders equipped with sensors to detect a variety of environmental variables, including chl, colored dissolved organic matter, and particle scattering, and (4) placement of a commercial fiberoptic telecommunications cable at Station ALOHA.The latter project, dubbed the ALOHA Cabled Observatory (ACO), provides an "extension cord" from shore into the deep sea for interactive instrumentation, delivery of power, and highspeed/optical transmission of data, video, and related information.The ongoing 22-year-long HOT study is changing the way we view the NPSG.Numerous unexpected discoveries, some serendipitous, have already been made including new microbes, new metabolic processes, and new paradigms, and several emerging climate-ecosystem connections have been value in climate and ecosystem research, and should be further encouraged.The predictive skill of current models can be tested using time-series observations and results derived from at-sea experiments.An interdisciplinary team of marine scientists was recently established to examine some of the contemporary challenges; the PARtnership for Advancing Interdisciplinary Global Modeling (PARADIGM) program has already made substantial progress toward the goal of predictive marine microbial ecology (Rothstein et al., 2006).Two of the main challenges confronting marine ecosystem modelers are: (1) current limitations of mechanistically based parameterizations of key processes, including solar energy capture and dissipation, nutrient flux processes, and controls on primary and secondary production, and (2) the ability to model mesoscale space and time variabilities that are now known to control carbon and related bioelemental fluxes in many open-ocean ecosystems(Doney, 1999).More recently, the marine microbial genomics revolution has fundamentally altered our views of many basic aspects of ecology and metabolism, and these discoveries must also be included in ocean simulations to provide the most accurate climate change projections(Doney et al., 2004).Environmental genomics, transcriptomics, and proteomics are beginning to provide new views of an old ocean; improved ecosystem models, and possibly new ecological theory, will be needed to fully capitalize on the new discoveries.CoNtempor arY ChalleNges aND opportUNitYOcean time-series programs promote discovery, ignite hypothesis generation and testing, and provide unique opportunity for enhancing our understanding of natural ecosystems.All successful ocean time-series programs to date have been hybrids of routine observations and exciting cutting-edge, hypothesisdriven research.Indeed, this is a logical linkage because the serial observations feed directly into hypothesis generation, prediction, and experimentation, oftentimes leading to new observations.In this way, time-series programs act as intellectual flywheels that create and sustain ever larger, complementary programs where the scientific outcome of the integrated effort is much larger than the sum of its parts.The teams of collaborating scientists that emerge and the unique partnerships that are created add to the overall value of the time-series effort.All successful time-series programs also have strong leadership, especially initially, but the truly impressive ones have multigenerational intellectual leadership that was passed on from senior to younger scientists without negative impact on the program as a whole.This program model is very important because the relevant time scale of many key ecological processes is longer than one academic lifetime.Because many modern time-series programs are transdisciplinary in their scientific missions, the leadership is often shared among disciplinary experts who collectively manage and promote the program among diverse science communities.Human influence, mostly during the past century, on coastal and open-ocean ecosystems has been profound.Foremost among the deleterious impacts has been the accumulation of greenhouse gases in the atmosphere and surface ocean due to the accelerated use of carbon-based fuels mostly for energy generation.This has led to warming, enhanced stratification, acidification, and nutrient limitation of the surface waters of the open sea.The impact of these habitat changes on primary production, fish stocks, and the ability of the ocean to further sequester CO 2 or to perform other key ecosystem services is largely unknown.These are societal matters of great concern, and comprise the main justification for investments in ocean time-series programs.One relatively new area of research involves the search for leading indicators of probable regime figure 7. map showing the geographical distribution of current ocean time-series sites, including moorings, observatories, and transport stations coordinated by the oceansites science and data management team.additional activities are in various stages of planning so the global network is likely to expand in the near future.From: http://www.oceansites.org establish additional time-series measurement programs if we ever hope to achieve the ultimate goal of predicting the probable impacts of human-induced climate change.Ocean time-series programs of the future must embrace new methods, sensors, instruments, and other improvements in engineering and computation.Not unlike the movement away from an annual physical examination to track one's general health to a personalized, full genome sequence and computer-based surveillance of vital signs, ocean ecosystem programs are also on the verge of embracing novel, high-frequency observation systems.It should be a thrilling and productive next decade.
Lesson 1: Site selection, sampling design, and frequency, as well as spatial scale of observations (single station vs. survey grid) are all relevant criteria in planning and implementing an oceanic ecosystem time-series program.Although it is impossible to know for sure whether a preselected site is representative of the larger region of interest, it is prudent to use any and all environmental data available at the time, as well as judgment concerning program logistics and costs.Although it is possible to relocate a time-series station that has later been shown to be inappropriately sited, it does mean that the initial time series will end and a new one will begin.
is Professor, Department of Oceanography, University of Hawaii, Honolulu, HI, USA.1 figure 1. schematic representation of the relevant temporal and spatial scales for key physical and ecological processes in the sea.The arrows and shaded regions define the approximate boundaries of at-sea time-series observations using current technologies.Courtesy of Tommy Dickey, University of California, Santa Barbara