Project Overview

Summary:

Dr. Frank Hernandez at the University of Southern Mississippi’s Division of Coastal Sciences, School of Ocean Science and Technology, was awarded an RFP-VI grant at $900,821 to conduct the RFP-VI project titled, “Deep-Pelagic Plankton Communities of the Northern Gulf of Mexico: Trophic Ecology, Assemblage Dynamics, and Connectivity with the Upper Ocean”. The project consisted of 2 other institutions (the University of Louisiana at Lafayette and Oregon State University); 1 principal investigator (Hernandez); 3 co-PIs (Drs. Kevin Dillon, Kelly Robinson, James Ruzicka); 1 postdoc (Verena Wang); 1 PhD student (Stacy Calhoun); 1 Masters student (Emily Gipson); 1 research staff member (Carley Zapfe).

     Much of what is known about the Gulf of Mexico (GOM) ecosystem is limited to coastal and upper ocean regions, even though > 90% of the GOM's volume occurs at depths > 200 m. The Deepwater Horizon oil spill (DWHOS) occurred in the deep GOM, and the lack of baseline data for this region was a major impediment to the damage assessment efforts. Although remote and understudied, the deep-pelagic environment plays a vital role, as many deep-pelagic organisms (including fish larvae) undergo wide-ranging and daily vertical migrations that drive a "biological pump" by actively transporting nutrients from the epipelagic zone to the deep GOM. To date, the Deep-Pelagic Nekton Dynamics (DEEPEND) Consortium is the only science team researching the deep-pelagic ecosystem in the GOM. However, the focus of DEEPEND is on micronekton and nekton, and not planktonic communities, which include the early life stages of deep-pelagic fishes and invertebrates, as well as prey resources for micronekton and nekton predators.

     Our goal is to address major knowledge gaps for the GOM by describing the community structure and trophic ecology of deep-pelagic plankton assemblages, and their connectivity with the upper ocean. Our specific objectives are to: 1) describe diets and trophic linkages for dominant, deep-pelagic larval and juvenile fishes using gut content analysis and stable carbon and nitrogen isotope analysis; 2) identify environmental drivers that structure deep-pelagic planktonic assemblages and vertical migration patterns; 3) develop a coupled plankton/upper trophic level and coupled deep-pelagic/epipelagic ecosystem model using ECOPATH and end-to-end ECOTRAN methods; and 4) apply data-driven ecosystem models to quantify rates of energy and biomass transfer between the epipelagic and deep-pelagic ecosystems that occur via trophic interactions and to estimate the consequences to ecosystem dynamics from perturbations in these linkages.

     At the core of this study are a wealth of unpublished data and plankton samples collected in 2010/2011 during the National Resource Damage Assessment (NRDA) response to the DWHOS, which includes many collections from the deep-pelagic region (a first for the GOM). Many of the NRDA deep-pelagic plankton collections overlapped spatially and temporally with data collected during concurrent deep-pelagic micronekton and nekton cruises. Combined, the plankton and micronekton/nekton data provide an unprecedented opportunity to examine the deep-pelagic ecosystem.

     The work being conducted for this project is highly responsive to GoMRI Theme 3 in that we are assessing the dynamics of the deep-pelagic planktonic community, which will contribute to evaluation of the “environmental effects of the petroleum/dispersant system” in the deep-pelagic GOM.  Oil and gas exploration/extraction in the GOM continues to push further offshore, making the probability of another deep water incident more likely, and the need for basic ecosystem data more critical.  Deep-pelagic planktonic communities were greatly understudied in the GOM prior to the DWHOS, with limited understanding of natural ecosystem conditions for post-spill assessments.  The work being conducted will fill a major data gap.  In doing so, the proposed work is highly responsive to GoMRI's specific requests in RFP-VI for "data integration from various sources" and "scientific synthesis across themes and consortia". 

Research Highlights

     Dr. Hernandez’s research, which included 15 conference presentations, 5 outreach products and activities, and 31 datasets to the GoMRI Information and Data Cooperative (GRIIDC), which are available to the public.  Significant outcomes of their research (all related to GoMRI Research Theme 3) are highlighted below.  All analyses and results described are in the process of being finalized for manuscript preparation and are subject to change.

      Deep-pelagic larval fishes are a large component of open ocean food webs and carbon cycling, yet their trophic ecology remains understudied.  Knowledge of trophic linkages in the deep-pelagic was improved by examining stable carbon and nitrogen isotopes in dominant deep-pelagic larval fishes (Objective 1).  These analyses will also contribute to quantification of trophic transfer between epipelagic and deep-pelagic ecosystems (Objectives 3 & 4).

• Stable isotope analysis of δ¹³C and δ ¹⁵N was conducted on ichthyoplankton from four dominant families of mesopelagic fishes (Myctophidae, Gonostomatidae, Sternoptychidae, and Phosichthyidae – all zooplanktivores) and one family of fish with a less common early life history feeding strategy (Paralepididae – micronektonivore).  Trophic niche position and niche size were estimated, as well as the probability of niche overlap among taxa.  Values of δ ¹⁵N, which represent relative trophic position, were similar among all larval fish taxa, and many displayed overlapping trophic niches regardless of feeding strategy.  However, within the family Myctophidae there were several genera that displayed small trophic niche areas and extremely low probability of trophic niche overlap, suggesting that trophic niche partitioning may occur even within families of fishes occupying similar depths and feeding guilds.         

      Zooplankton are also major contributors to the oceanic carbon cycle, and remain understudied in mesopelagic and bathypelagic zones.  Knowledge of deep-pelagic zooplankton distributions in the GOM was improved by describing zooplankton community composition and vertical distribution patterns (Objective 2), which will also help to inform diet availability for mesopelagic larval fishes (Objective 1).  

• Zooplankton images from depth-discrete net sampling conducted on two DWH NRDA deep-pelagic plankton cruises conducted in Fall 2010 and Spring 2011 were classified into 11 distinct functional groups: amphipods, bryozoans, cnidarians, copepods, crustaceans, echinoderms, ichthyoplankton (minimal residual numbers after samples were sorted and larval fishes removed for identification and archiving), molluscs, pelagic tunicates, protozoans, and worms.
• Epipelagic depths (< 200 m) contained the most zooplankton overall.  When comparing functional groups, relative proportions differed between depth zones and seasons, but were similar during day and night. 
• Copepods were the most abundant functional group overall by two orders of magnitude, and represented five taxonomic orders.  Copepod composition changed between depth zones and seasons.  Diel migratory signals were observed in copepods during both Spring and Fall.  Copepods are a targeted food source for many forage fish, and understanding their vertical migration behavior will aid in construction of the ecosystem model and describing the movement of energy through the open ocean system.   

      Little is known about the taxonomic composition of ichthyoplankton in the deep-pelagic GOM, as major plankton surveys (i.e., SEAMAP) do not sample beyond the epipelagic.  In order to improve our understanding of planktonic distributions in the GOM, the aims of Objective 2 were to describe deep-pelagic larval fish assemblages and vertical migration patterns, and identify the biophysical drivers that structure these distributions.

• Vertical patterns of distribution were analyzed for four dominant families of mesopelagic larval fishes (Myctophidae, Gonostomatidae, Sternoptychidae, and Phosichthyidae) sampled in the northern GOM from 0 – 1000 m using depth-discrete nets.  Results suggest that vertical distribution patterns of larval fishes are strongly size-dependent, and that vertical migration patterns were variable among taxa.  Small (< 10 mm) myctophid, gonostomatid, and phosichthyid larvae were found almost exclusively in epipelagic waters (< 200 m).  At 10 – 20 mm, myctophid and phosichthyid larvae were found much deeper, mostly beyond 400 m during the night, with a small proportion of larvae moving to the epipelagic during the day.  Patterns of diel vertical migration were strongest for myctophids and phosichthyids at > 20 mm length.  While sternoptychid larvae < 10 mm were most abundant in the epipelagic, a proportion were present in the upper mesopelagic even at this early life stage.  Both sternoptychid and gonostomatid larvae > 10 mm were found almost exclusively in mesopelagic waters regardless of diel period. 
• Community assemblage structure was evaluated for ichthyoplankton sampled from six DWH NRDA deep-pelagic plankton cruises.  Larval fish abundance, richness, and diversity were highest in the epipelagic and decreased with increasing depth.  The vast majority (97%) of ichthyoplankton sampled beyond 200 m depth were from the four dominant mesopelagic families described above.  Multivariate regression trees were applied to evaluate the effect of environmental and spatiotemporal features on ichthyoplankton assemblage structure.  Within all pelagic regions, larval assemblages were largely structured by depth and season.  Epipelagic assemblages were additionally structured by upper water column features (e.g., surface salinity and mixed layer depth), driving patterns of abundance in mesopelagic taxa in lower epipelagic waters (25 -200 m).  While broad-scale temporal variability (season) was a significant driver of assemblage composition, fine-scale temporal variability (diel period) and spatial variability in larval fish assemblage structure appeared limited in the northern GOM.   

      Previous ecosystem models for the Gulf of Mexico have focused on continental shelf ecosystems.  Development of a vertically resolved open ocean ECOTRAN model (Objectives 3 & 4) helps us to quantify the rates of energy and biomass transfer via trophic interactions between the deep-pelagic and epipelagic regions and estimate how perturbations within any one depth regime propagate throughout the water column.

• The end-to-end ecosystem model platform (ECOTRAN) was adapted to incorporate vertical structure and physical processes in the open ocean.  Discrete food webs were included for epipelagic, mesopelagic, bathypelagic, and benthic zones.  Numerous processes were modeled, including vertical migration, particle sinking rates, bacterial metabolism of detritus, and vertical mixing of nutrients and detritus.  Model sensitivity tests were conducted using an idealized oceanic food web.  Sensitivity to detritus sinking speeds was tested, and increased sinking speeds resulted in increased production in the mesopelagic and decreased production in the epipelagic.  Sensitivity to fish diel vertical migration rate was also evaluated, and increased rates resulted in increased higher level trophic production in both the epipelagic and mesopelagic, and increased total production.
• A perturbation event was tested using an idealized oceanic food web within ECOTRAN.  Recovery time of a short-duration but high mortality (50%) event within the planktivorous mesopelagic fish population was on the order of 4 years.  Recovery time for their piscivorous fish predators was longer, approx. 5 years. The recovery of the non-migratory piscivorous fish population was faster in the epipelagic depth zone than in the mesopelagic.  Recovery times within both the epipelagic and mesopelagic depth zones were reduced as the diel vertical migration rates within the planktivorous fish community increased.
• A higher resolution, data-driven food web model consisting of 46 distinct functional groups was developed for a model domain encompassing a portion of the northern Gulf of Mexico.  The functional groups include categories from detritus to marine mammals, with numerous distinct groups of zooplankton and fishes.  Using this high-resolution food web model, the ECOTRAN model platform is being used to evaluate trophic transfer among pelagic regions in the northern GOM and to simulate the propagation of trophic impacts in perturbation scenarios and estimate ecosystem recovery times.   

Much of what is known about the Gulf of Mexico (GOM) ecosystem is limited to coastal and upper ocean regions, even though > 90% of the GOM's volume occurs at depths > 200 m. The Deepwater Horizon oil spill (DWHOS) occurred in the deep GOM, and the lack of baseline data for this region was a major impediment to the damage assessment efforts. Although remote and understudied, the deep-pelagic environment plays a vital role, as many deep-pelagic organisms (including fish larvae) undergo wide-ranging and daily vertical migrations that drive a "biological pump" by actively transporting nutrients from the epipelagic zone to the deep GOM. To date, the Deep-Pelagic Nekton Dynamics (DEEPEND) Consortium is the only science team researching the deep-pelagic ecosystem. However, the focus of DEEPEND is on micronekton and nekton, and not planktonic communities, which include the early life stages of deep-pelagic fishes and invertebrates, as well as prey resources for micronekton and nekton predators.

Our goal is to address major knowledge gaps for the GOM by describing the community structure and trophic ecology of deep-pelagic plankton assemblages, and their connectivity with the upper ocean. Our specific objectives are to: 1) describe diets and trophic linkages for dominant, deep-pelagic larval and juvenile fishes using gut content analysis and stable carbon and nitrogen isotope analysis; 2) identify environmental drivers that structure deep-pelagic planktonic assemblages and vertical migration patterns; 3) develop a coupled plankton/upper trophic level and coupled deep-pelagic/epipelagic ecosystem model using ECOPATH and end-to-end ECOTRAN methods; and 4) apply data-driven ecosystem models to quantify rates of energy and biomass transfer between the epipelagic and deep-pelagic ecosystems that occur via trophic interactions and to estimate the consequences to ecosystem dynamics from perturbations in these linkages.

At the core of this study are a wealth of unpublished data and plankton samples collected in 2010/2011 during the National Resource Damage Assessment (NRDA) response to the DWHOS, which includes many collections from the deep-pelagic region (a first for the GOM). Many of the NRDA deep-pelagic plankton collections overlapped spatially and temporally with data collected during concurrent deep-pelagic micronekton and nekton cruises. Combined, the plankton and micronekton/nekton data provide an unprecedented opportunity to examine the deep-pelagic ecosystem.

The proposed work is highly responsive to GoMRI Theme 3 in that we will establish a baseline needed to examine "environmental effects of the petroleum/dispersant system" in the deep-pelagic GOM. Oil and gas exploration/extraction in the GOM continues to push farther offshore, making the probability of another deep water incident more likely, and the need for baseline data more critical. No baseline data for deep-pelagic planktonic communities existed prior to the DWHOS for post-spill assessments, therefore the proposed work fills a major data gap. Further, our project complements the work of the DEEPEND consortium. Our proposed ecosystem models will integrate our findings with those of DEEPEND and will be used to investigate the propagation of nutrient and plankton dynamics to higher trophic levels. In doing so, the proposed work is highly responsive to GoMRI's specific requests in RFP VI for "data integration from various sources" and "scientific synthesis across themes and consortia".

 Proposal Abstract - RFP-VI Frank Hernandez