Climate impacts on zooplankton population dynamics in coastal marine ecosystems

. The 20-year US GLOBEC (Global Ocean Ecosystem Dynamics) program examined zooplankton populations and their predators in four coastal marine ecosystems. Program scientists learned that environmental controls on zooplankton vital rates, especially the timing and magnitude of reproduction, growth, life-cycle progression, and mortality, determine species population dynamics, seasonal and spatial distributions, and abundances. Improved knowledge of spatial-temporal abundance and distribution of individual zooplankton taxa coupled with new information linking higher trophic level predators (salmon, cod, haddock, penguins, seals) to their prey yielded mechanistic descriptions of how climate variation impacts regionally important marine resources. Coupled ecological models driven by improved regional-scale climate scenario models developed during GLOBEC enable forecasts of plausible future conditions in coastal ecosystems, and will aid and inform decision makers and communities as they assess, respond, and adapt to the effects of environmental change. Multi-region synthesis revealed that conditions in winter, before upwelling, or seasonal stratification, or ice melt (depending on region) had significant and important effects that primed the systems for greater zooplankton population abundance and the following spring-summer,

. In each, understanding the spatial and temporal abundances of zooplankton species required understanding of the species population dynamics (vital rates) and the physical processes affecting the habitat.
In continental shelf ecosystems, environmental conditions, such as temperature, stratification, and current velocity, vary over both small and large spatial ranges and temporal scales, from daily to interdecadal. Species life histories interact with temporal-spatial environmental variability, often in a nonlinear manner. For example, short-term temporal variability in ocean conditions may be important for some species but not ABSTR ACT. The 20-year US GLOBEC (Global Ocean Ecosystem Dynamics) program examined zooplankton populations and their predators in four coastal marine ecosystems. Program scientists learned that environmental controls on zooplankton vital rates, especially the timing and magnitude of reproduction, growth, life-cycle progression, and mortality, determine species population dynamics, seasonal and spatial distributions, and abundances. Improved knowledge of spatialtemporal abundance and distribution of individual zooplankton taxa coupled with new information linking higher trophic level predators (salmon, cod, haddock, penguins, seals) to their prey yielded mechanistic descriptions of how climate variation impacts regionally important marine resources. Coupled ecological models driven by improved regional-scale climate scenario models developed during GLOBEC enable forecasts of plausible future conditions in coastal ecosystems, and will aid and inform decision makers and communities as they assess, respond, and adapt to the effects of environmental change. Multi-region synthesis revealed that conditions in winter, before upwelling, or seasonal stratification, or ice melt (depending on region) had significant and important effects that primed the systems for greater zooplankton population abundance and productivity the following springsummer, with effects that propagated to higher trophic levels.
The GLOBEC approach is to develop fundamental information about the basic mechanisms that determine the abundance and distribution of marine animal populations and, most importantly, the variability of these populations about their average values. (GLOBEC, 1991a, p. 1) Moreover, the characteristics of a region that determine its suitability for any given organism depend not only on the availability of food and the abundance of predators but also upon the dynamic physical features of the local environment that influence the success of recruitment, the efficiency of feeding, and the susceptibility to predation. (GLOBEC, 1991b, p. 5) Oceanography for others, and a species with a diverse repertoire of behaviors may respond differently to climate change than species with less flexible life-history strategies.
These differences among zooplankton may have dramatic effects on marine ecosystem structure (Peterson, 2009;Johnson et al., 2011).
Population ecology focuses on population abundance, how it varies temporally, and the biotic and physical processes that determine it. Fundamental to zooplankton population dynamics are the vital rates of birth, development, growth, and mortality that lead to changes of body size and numbers (or biomass), as well as the environmental factors that influence individual vital rates ( Figure 1). Key zooplankton and fish or other higher trophic level species were targeted for detailed study based on their importance to the ecological dynamics or fisheries of a region (Turner et al., 2013, in this issue Figure 2). Standard net tow and pump sampling methods were used to collect zooplankton in the Northwest Atlantic (Durbin and Casas, 2006), Northeast Pacific Pinchuk, 2003, 2005;Peterson and Keister, 2003;Lavaniegos and Ohman, 2007), and Southern Ocean (Marrari et al., 2011a,b;Wiebe et al., 2011). These regions provided the samples needed to characterize the detailed species and life stage information critical for population and life-history studies (Daly, 2004;Reese et al., 2005;Marrari et al., 2011a,b), such as recruitment into krill and copepod populations (Feinberg and Peterson, 2003;Runge et al., 2006;Feinberg et al., 2010), stage-specific rates of mortality , and timing of dormancy Maps et al., 2012). Bioacoustic and optical technologies were developed and employed to describe finer-scale distributions over broad areas. The Video Plankton Recorder enabled specific identification of species and stage (or size) of zooplankton that could be directly related to concurrent physical measurements at similar spatial-temporal scales (Benfield et al., 1996;Norrbin et al., 1996;Davis et al., 2005). Processes that influence vital rates of individuals and abundance and distribution of zooplankton populations. Environmental (temperature, turbulence, light, food) and individual (behavior) factors that control these processes are shown. Modified from Figure 3 in GLOBEC (1992) to include diapause and lipid storage impacts on individuals (Jaffe et al., 1999) up to 50 m or even coast-wide were characterized using a variety of bioacoustic or optical instruments (Ressler et al., 2005;Swartzman et al., 2005;Wiebe et al., 1996;Lawson et al., 2004Lawson et al., , 2008. Such data are necessary for assessing the spatial patchiness of prey composition and evaluating its impact on the feeding dynamics of zooplankton predators (Young et al., 2009).
GLOBEC observations and understanding of population dynamics provide insight into the mechanisms of bottomup physical forcing that determine biological production at lower trophic levels (phytoplankton and zooplankton), which in turn influence production of upper trophic level species subject to resource management (Fogarty et al., 2013, in this issue). Coupled bio-physical population models provided predictions of spatio-temporal distribution and abundance of key zooplankton species in the North Atlantic (e.g., Ji et al., 2009;Stegert et al., 2012), Northeast Pacific (Dorman et al., 2011;Lindsey, 2014), and Southern Ocean (Piñones et al., 2013). These models allow a dynamic description of interactions among life history strategies and the physical environment at many scales simultaneously ( Figure 2). Examples of these models are described in greater detail in a review of the advancements in coupled modeling achieved by GLOBEC (Curchitser et al., 2013, in this issue).  (Caswell, 2001), population surface (Wood, 1994), vertical life table (Aksnes and Ohman, 1996), and an  . Note the general tendency for the correlation to be positive when there is higher variability between spatial and temporal scales, with the greatest variability at daily (DVM = diel vertical migration; all scales > 1 m), intraseasonal (mesoscale), annual (basin), multidecadal (basin-global), and glacial-interglacial (global) time (space) scales. Haury et al. (1978) advection-differencing method (Li et al., 2006). A fifth approach for estimating mortality of the total plankton community using biomass spectrum theory (Zhou and Huntley, 1997;Edvardsen et al., 2002) was applied in the California Current System (Wu, 2008)   . Mortality rates, which vary not only in time, but also in space, were found to covary with predators in both the California Current System (Ohman and Hsieh, 2008) and the Georges Bank region (Li et al., 2006;. Notably, regions of elevated food availability to zooplankton in the coastal upwelling region were also associated with elevated predation mortality, confirming Bakun's (2006, p. 117) assertion that "for planktonic organisms…food heaven almost invariably equates to predation hell. " A comparison of mortality rates of Calanus finmarchicus in five locations across the North Atlantic revealed regional differences in stagespecific patterns of mortality (Ohman et al., 2004), together with evidence for density-dependent egg mortality related to cannibalism in both the open ocean (Ohman and Hirche, 2001) and on the Northwest Atlantic continental shelf (Ohman et al., 2004). The findings of density-dependent mortality in GLOBEC field studies corroborated the inference of the importance of this process deduced from pelagic ecosystem models (Steele and Henderson, 1992;Fasham, 1995).

Modified and redrafted with inspiration from an original graphic in
In addition to zooplankton mortality,
Other investigations examined egg production of copepods (Napp et al., 2005) and euphausiids (Pinchuk and Hopcroft, 2006). Growth, development, and reproduction rates applied to stage-specific field abundances provided estimates of seasonal production of prey for larval fishes in the northern Gulf of Alaska.
These zooplankton vital rate measurements will contribute to future coupled bio-physical models directed at understanding climate impacts on population dynamics (Pinchuk et al., 2008).

California Current System
Euphausiids and copepods are important prey for higher trophic levels in the California Current. Copepods are indicators of transport processes and the bioenergetic content of the food chain.
The California Current program included biweekly monitoring of physical and biological properties along a cross-shelf transect (Newport Hydrographic Line) that had been intensively sampled in the 1960s and early 1970s (Huyer, 1977;Peterson and Miller, 1977 (Peterson and Keister, 2003).
Horizontal transport was suggested as the process linking the PDO with coastal ecosystem structure in the Northern California Current (Peterson and Hooff, 2005;Hooff and Peterson, 2006). In a modeling study, Keister et al. (2011) showed that northward and onshore transport of warmer waters during the positive phases of the PDO introduced smaller, subtropical copepods to the shelf waters off central Oregon (Figure 4; see also Di Lorenzo et al., 2013, in (Lavaniegos and Ohman, 2003). Year PDO

X-Axis Anomaly
Year PDO

X-Axis Anomaly
Year PDO X-Axis Anomaly  Strub et al. (2002) went into reproduction rather than somatic growth. Seasonal growth rates of E. pacifica from short-term incubations were similar to cohort analysis estimates from the Oregon shelf (e.g., Smiles and Pearcy, 1971). These results were complemented by observations of ontogenetic behavioral differences in diel vertical migration, and modeling of cross-and along-shelf transport of eggs, larvae, and adults (Dorman et al., 2011;Lindsey and Batchelder, 2011;Lindsey, 2014).

THE IMPORTANCE OF WINTER CONDITIONS TO ZOOPL ANKTON DYNAMICS
The  (Durbin et al., 1997(Durbin et al., , 2003Feinberg et al., 2010). Wind-driven coastal upwelling of nutrients supports primary production in the California Current (Checkley and Barth, 2009). In the Northern California Current (Oregon), production is concentrated in spring and summer when upwelling-favorable winds dominate, and both nutrients and light are favorable for phytoplankton growth. Indeed, the conventional view that production depends almost entirely upon local coastal upwelling processes during the so-called "upwelling season" is reflected in the design of GLOBEC  and other large interdisciplinary studies of ecosystem processes and productivity in the California Current (Barth and Wheeler, 2005;Largier et al., 2006 (Legaard and Thomas, 2006), the early diatom production nonetheless fuels elevated egg production by C. marshallae and C. pacificus and an early burst of egg production by the coastal euphausiid, Thysanoessa spinifera   et al., 2011), rockfish growth (Black et al., 2010(Black et al., , 2011 and recruitment (Laidig, 2010), and initiation of seabird nesting (Schroeder et al., 2009)  Bank, which depend on annual resupply from the Gulf of Maine (Durbin et al., 1997(Durbin et al., , 2003Miller et al., 1998). The fallwinter changes in phytoplankton production have also been hypothesized to impact subsequent reproduction by cod and haddock (Friedland et al., 2008). In the Northwest Atlantic GLOBEC program, hydrodynamic and trophodynamic processes were related to growth and survival of larvae of Atlantic cod and haddock using field observations, experiments, and models (Werner et al., 1996(Werner et al., , 2001Mountain et al., 2003Mountain et al., , 2008Kristiansen et al., 2009).

ZOOPL ANKTON AND L ARVAL FISH
Individual-based modeling, often linked with physical and/or ecosystem models, is a common approach for investigating the importance of advection, starvation, and predation on survival of larval fish (Peck and Hufnagel, 2012). Cod and haddock larvae in the Northwest Atlantic rely extensively on zooplankton (especially copepods) as prey (Broughton and Lough, 2010), and larval survival is positively related to the abundance of suitable zooplankton prey (Buckley and Durbin, 2006;Mountain et al., 2008). Mountain et al. (2008) linked egg mortality to wind-driven Ekman (e.g., transport) losses from Georges Bank, which has also been shown using coupled trophodynamics-transport models (Werner et al., 2001).
In the Northern California Current, salmon survival and climate are linked through zooplankton transport and population dynamics. Francis and Hare (1994) and Mantua et al. (1997) showed that salmon survival was correlated with shifts in the North Pacific Index and the PDO, respectively, but neither suggested a mechanism for this correlation.