An Overview of EMAX: The Northeast U.S. Continental Shelf Ecological Network. ICES CM 2005/L:02

Notable changes have occurred in the Northeast U.S. Continental Shelf LME. To evaluate the response of this ecosystem to numerous human-induced perturbations, as well as to explore possible future scenarios, the Northeast Fisheries Science Center instituted the Energy Modeling and Analysis eXercise (EMAX). The primary goal of EMAX was to establish an ecological network model (i.e., a more nuanced energy budget) of the entire Northeast U.S. food web. The EMAX work focused on four subregions of the ecosystem from contemporary times (1996-2000), had 36 network “nodes” or biomass state variables across a broad range of the biological hierarchy, was highly interdisciplinary, and incorporated a wide range of key rate processes. The emphasis of EMAX was to explore the particular role of small pelagic fishes in the ecosystem. Various model configurations were constructed and psuedo-dynamic scenarios were evaluated to explore how potential changes to the small pelagic fishes can affect the rest of the food web. Our results show that small pelagic fishes are clearly keystone species in the ecosystem. There are some differences across the four EMAX regions reflective of the local biology, but major patterns of network properties are similar over space. Finally, EMAX will continue to play a critical role for the further development of EAF as: a catalogue of information and data; identifying major fluxes among biotic components of the ecosystem; a basis for further analytical models; a way to evaluate biomass tradeoffs; and a backdrop for a suite of other relevant, management and research questions. Link et al. ICES CM 2005/L:023 Introduction There have been numerous recent calls to adopt an ecosystem approach to fisheries (EAF or Ecosystem-based fisheries management, EBFM; here EAF and EBFM are used synonymously; e.g., Garcia et al., 2003, Garcia 2005, Link 2002). There many rationales for why EAF is an emerging approach. Central to these considerations are taking a more holistic look at an ecosystem and simultaneously evaluating tradeoffs among component biomass or user sectors (e.g., Larkin 1996, Link 2002, etc.). One approach to explore holistic ecosystem perspectives and to examine biomass tradeoffs is to use ecosystem models. Within the wide variety of possible models one could use (Hollowed et al. 2000, Whipple et al. 2000), energy budgets and network analyses provide useful tools to evaluate relative biomass, system properties, and fluxes within an ecosystem. Many of these models allow one to explore the fate and flux of production within a system by explicitly tracking how the energy flows from various components of the system to one another. Of the many possible network models, we chose to use EcoPath (Christensen and Pauly 1992, Walters et al. 1997) and EcoNetwrk (Ulanowicz 2004, Ulanowicz and Kay 1991) as two models to allow us to evaluate various spatial, temporal, and hypothetical scenarios in our ecosystem of interest. The Northeast United States continental shelf large marine ecosystem (hereafter, NEUS; Sherman 1991) is a highly productive ecosystem that has supported significant commercial fisheries for multiple centuries. The recent history of the component fish stocks has exhibited the classic cycles of excessive effort, stock declines, and iterations thereof until the point of sequential stock depletion (Fogarty and Murawski 1998, Murawski et al. 1997, Serchuk et al. 1994). Within NEUS, there are four main subLink et al. ICES CM 2005/L:024 regions (Figure 1) that have different ambient fauna and thus slightly different fishing histories and environmental considerations. In this context, the major fishery-related events in the NEUS over the past several decades can be characterized loosely as the following sequence: an increase in small pelagic catches by foreign fleets, a continued increase in demersal groundfish catches, a precipitous decline in small pelagic stocks, a decline of some groundfish stocks, an effective cessation of the small pelagic fisheries (and expulsion of foreign fleets), a continual series of overfishing on an ever-increasing array of groundfish, an increase in elasmobranch stocks, the beginnings of an increase in small pelagic stocks, an establishment of elasmobranch fisheries, an increase in benthic invertebrate fisheries and stocks, the persistence of groundfish stocks at moderate to low levels, the persistence of grounfish fisheries at suboptimal yields, the decline of elasmobranch stocks and subsequently their fisheries, and the effective explosion of small pelagic stocks to what are now record highs (Serchuk et al. 1994, Murawski and Fogarty 1998, Link and Brodziak 2002, Overholtz 2002). While the various fisheries and their effects were occurring, notable changes to protected, endangered and threatened species (PETS; e.g. many marine mammals) have occurred, with many in more critical condition than 50 years ago (Waring et al. 2004). Additionally, shifts in non-targeted fauna (NEPA species; e.g. some benthos, some non-targeted fishes) also occurred (Link and Brodziak 2002), with some actually persisting at relatively stable levels or even increasing (Link and Brodziak 2002, Link 2005). Overall, the general observation is that the ecosystem has undergone a shift from a vertical to a horizontal system (Link 1999) due to the resurgence of herring and mackerel. With this resurgence of herring and mackerel stocks, the question begs: how Link et al. ICES CM 2005/L:025 important have these small pelagics become to the success of other commercial fish stocks, PETS, NEPA species, and ultimately the overall functioning of the ecosystem. This issue has become increasingly important as multiple stakeholders have begun exploring potential tradeoffs in the NEUS ecosystem. Our objectives in this work were fourfold. First, we wanted to balance a contemporary energy budget of NEUS. Second, we wanted to construct an energy budget/network for each of the four sub-regions of NEUS. Third, we wanted to compare key network properties spatially across these four regions. Finally, we wanted to evaluate the relative importance of the small pelagics in this ecosystem. To achieve these and related objectives, the Northeast Fisheries Science Center (NEFSC) instituted the Energy Modeling and Analysis eXercise (EMAX). In some respects this document serves as an initial but partial report of that exercise. Materials and Methods Nodes & Data Sources After a long series of iterations, we settled on 36 major nodes for the EMAX network (Figure 2). These represent the full range of the biological hierarchy, with organisms spanning < 10 μm to > 30 m. Most network nodes represent a broad range of functionally similar taxa, but also integrate across a wide range of diversity and ecological functionality. The species in each node can be found in a more detailed set of documentation (EMAX 2005). We particularly chose to include elements of the microbial loop (Azam et al. 1983) to help in the balancing, to better reflect reality, and to acknowledge some of the fundamental shifts in views of how the oceans function. Again, Link et al. ICES CM 2005/L:026 our emphasis was on small pelagics so some of the nodes that may not greatly interact with those species were not necessarily grouped as they might have been with a different focus. We limited our data usage, parameter estimation, and temporal coverage to the 1996-2000 time period. We did so for several reasons, chief of which was data availability, commonality of units, and spatio-temporal overlap for as many of the nodes as possible. We used an appropriate level of seasonality for each node integrated into an annual average for each estimate during this period. The units we chose to use were g wet weight m. There were five main elements critical to the construction of each node for the four NEUS networks. We estimated biomass, production, consumption, respiration, and diet composition for all nodes. Additionally, for some nodes it was germane to estimate other sources of removalsnamely fisheries. Various approaches were used for all the nodes, ranging from literature bounding of values for some of the globally underdetermined groups (e.g. bacteria, microzooplankton) to probabilistic estimates from multiple sampling regimes (e.g. some of the fishes). Further details of our estimation protocols and methods, for biomass, the various rates, and diet compositions, can be found in EMAX (2005). We noted a qualitative pedigree of data “certainty” for each parameter of each node. An important distinction of our approach relative to many other energy budget and network modeling exercises is that we started the balancing protocol with all parameters having an initial estimate and did not use the models to estimate any parameter values. Certainly the models modified these values in the balancing protocols, but the point is Link et al. ICES CM 2005/L:027 that, factually speaking without any hubris, we started with a degree of completeness where most other studies end up. Models We explored a broad myriad of energy budget and network models (and their associated software packages). After several iterations, we settled on using two models as our primary tools: EcoPath (Christensen and Pauly 1992, Walters et al. 1997) and EcoNetwrk (Ulanowicz 2004, Ulanowicz and Kay 1991). The pros and cons of each have been variously noted elsewhere (e.g., Allesina and Bondavalli 2003, Heymans and Baird 2000, Hollowed et al. 2000, Kavanagh et al. 2004, Ulanowicz 2004, Walters et al. 1997, Whipple et al. 2000). One reason we chose these two is the subtle differences in underlying philosophy and numerical solving approaches between the two programs (Heymans and Baird 2000). Other models we evaluated are also potentially very useful but were either redundant with the two we chose, focused on more qualitative network properties, were less user friendly, or obfuscated their underlying model structure. As