The 2023 CBIOMES Annual Meeting in New York saw a re-energizing return to large-scale, in-person programming
Reporting by Helen Hill for CBIOMES News
This year’s meeting was built around eight mini-workshops; three building on the smaller in-person workshops that took place in October 2022, and January, and April 2023 (Diel Processes, Transects and Ecoprovinces, Zooplankton); four exploring new directions as well as existing areas of collaboration (Haptophytes, Macro-ecological Models, Regional and Global Modeling, Macromolecular and Protein Allocations); and an eighth providing hands-on experience with DAR1, the 0-D cloud sandbox Darwin model.
Diel Processes
The Diel Processes mini-workshop provided an opportunity to revisit October 2022’s in-depth discussions on diel processes through the planning of a diel review paper.
During this session, participants narrowed down the major diel processes to be included in the paper, focusing on phytoplankton physiology and diversity, and the modeling and observational challenges.
Project leads look forward to a long list of CBIOMES members engaging on the paper (all are welcome) which it is intended will highlight the need for research on fast environmental fluctuations such as the diel cycle within the marine ecosystem research community, as well as offering guidelines for future research on observational, experimental, and modeling approaches. A timeline towards a January 2024 submission date can be found on the Diel Processes Working Group page.
Contact: Ioannis Tsakalakis (MBL)
Transects and EcoProvinces
The Transects and EcoProvinces mini-workshop returned to questions around characterizing and interpreting the biogeography of plankton in observed and simulated ocean transects.
In this mini-workshop researchers picked up discussions concerning the spatial distribution patterns of different plankton species, with a special focus on Prochlorococcus, examined the similarity of the patterns across different datasets (e.g., SeaFlow, ASVs), and assessed the accuracy of current mathematical models in predicting these biogeographic patterns.
Subgroups also made progress on the various manuscripts identified at the January workshop. These manuscripts are currently in various stages of drafting, review, and revision.
Lastly, attendees engaged in a fruitful discussion about potential synergies with the SCOPE-Gradients project. The mini-workshop also explored how CBIOMES’ research on ecoprovinces can align and complement SCOPE’s focus on understanding microbial processes across the tropical and equatorial Pacific oceans. The discussion presented opportunities for future collaboration and data exchange.
Contact: Francois Ribalet (UW)
Zooplankton
Despite only a few weeks having elapsed between the originating workshop on Zooplankton in Marine Ecosystem Models at the end of April and the Annual Meeting, the Zooplankton mini-workshop allowed participants to reconvene and begin formulating next steps as well as to showcase some of the developments that have already been made.
Niall McGinty began by giving an outline of the April workshop to bring those unable to attend up to speed. The second part of his talk focused on the types of data that have been collected for zooplankton and a discussion followed on the ways they may be integrated or compared.
One of the most promising avenues for inter-comparisons between data is seen to be between ASVs and imaging/net data. Yubin Raut (Fuhrman Lab, USC) discussed the progress made in analyzing the biogeography of several large zooplankton groups along transects in three ocean basins using ASV data, Attendees identified several other zooplankton datasets collected along the same routes that can be used to calibrate and compare the ASV results with other data.
Following coffee, Joe Vallino (MBL) discussed work in his group using experimental and mathematical models to understand microbial food web structure and function. Time scale is important when modeling these microbial food webs with the emergence of specialist or generalist bacteria depending on duration. The results of stable isotope experiments will help constrain time scales for emerging dynamics in microbial communities.
Stephanie Dutkiewicz (MIT) led the final section of this mini-workshop with a discussion on zooplankton modeling in ecosystem models. It was highlighted that the issue of modeling zooplankton is a widespread problem. An inter-comparison between different models shows a large disparity between models with phytoplankton loss rates due to grazing contributing the largest source of uncertainty between models. Initial sensitivity experiments, where different aspects of zooplankton grazing were changed (e.g. wider feeding kernel, new functional groups) resulted in some interesting consequences that will be explored further.
With the intention of forming a working group for zooplankton modeling focusing on 1) Data comparisons and 2) Ways to improve zooplankton estimates in ecosystem models, all interested CBIOMES members are inviteed to bring their own ideas forward. Contact Niall (nmcginty@dal.ca) to participate.
Contact: Niall McGinty (Irwin Group, Dalhousie)
Haptophytes
Haptophytes include functionally significant coccolithophores and phaoecystis. In this session investigators from MIT, the Fuhrman Lab at USC, and the Bertrand Lab at Dalhousie, addressed a range of questions including: What do we understand of the key traits, trade-offs, and ecology of these organisms? What are the key ecological and biogeochemical questions we might address by resolving these species?
Stephanie Dutkiewicz (MIT) began the session by outlining the diversity within the haptophyte group, which includes biogeochemically important calcifiers, significant pico-eukaryotes, as well as phaeocystis which is ubiquitous and sometimes dominant in biomass and productivity. However, this diversity is not currently resolved in ecological and biogeochemical models: how much is necessary for specific applications, and do we know how to resolve these groups?
Following her was Yubin Raut (Fuhrman Lab, USC) with a review of haptophytes in the Fuhrman lab’s GRUMP data set: haptophytes account for as much as 50% of total phytoplankton ASVs and pigments at some places in the transects discussed.
Erin Bertrand (Dalhousie) reviewed and then led a discussion of Phaeocystis ecology and physiology, discussing large blooms on the Labrador Sea (also seen in Antarctic waters). As much as 50% of carbon fixation goes into mucilaginous material associated with colonies.
These presentations led to discussions framed around key traits and trade-offs for Phaeocystis and coccolithophores. For the former, these highlighted the potential significance of flagella and haptonema. For Phaeocystis colonialism is a defining trait, but what impact does it have on fitness, and how? For coccolithophores, the importance of linking calcification and in-water carbonate chemistry was emphasized if we are to have dynamic, relevant models of calcification under global change. Some interest was expressed in possible further discussions or working groups around parameterizing Phaeocystis.
Contact: Mick Follows (MIT)
Macroecological Models
MITgcm-Darwin simulations and taxonomically-resolved observations reveal local and regional ecosystem structures. These can be characterized using “macro-ecological” metrics including Species Abundance Distributions, Taylor’s Law, and the slope of the size spectrum which can provide compact, quantitative metrics of simulation skill. In turn, models and simulations can provide interpretations of these characteristics. Investigators from MIT, Dalhousie, and USC, used this session to review potentially useful metrics and to discuss applications.
Mick Follows (MIT) gave a plenary overview, outlining the topic and describing a few potentially valuable macro-ecological metrics (e.g. scaling relationships between population sizes, demographics of cell size, and Species Abundance Distributions.)
Andrew Irwin (Dalhousie) followed with a stimulating tour through some recent research and macro-ecological analysis of data sets including the Continuous Plankton Recorder and Darwin-MITgcm simulations. As an example (linking also to the recent Zooplankton workshop), he noted that emergent Z:P biomass ratios in Darwin compare well with observed relationships.
Jesse McNichol (St. Francis Xavier Univ., Canada – formerly of the Fuhrman Lab, USC) followed with a presentation and discussion on relevant observational data sets, with some particular emphasis on GRUMP; the global-scale ASV biogeography developed by the Fuhrman Lab. Jesse showed initial analyses of Species Abundance Distributions from the data set with much discussion. Amongst other topics, it was discussed how to interpret genetic signatures on filters of organisms too large/rare to have been physically present as a live organism in a sample.
In summary, the session provided an excellent seed for yet further discussion and continued applications of macro-ecological measures of ecosystem structure in a data-model context: both as a “test” of simulations, but also as a way to exploit models as interpretive tools.
Contact: Mick Follows (MIT)
Molecular and Protein Allocation
Talks from investigators at MIT and Dalhousie reviewed recent progress in understanding and modeling macro-molecular and protein allocation at the cellular scale – including field and laboratory data – leading to a discussion of insights from the initial implementation of allocation models in ocean simulations and next steps.
Mick Follows (MIT) kicked things off with a revie of the motivation for macro-molecule resolving models, including biologically meaningful mechanisms and the controls on elemental stoichiometry.
Zoe Finkel (Dalhousie) then briefly reviewed findings from laboratory studies of the macro-molecular and elemental composition of 17 species of phytoplankton under a variety of growth conditions. This remarkable data set provides the foundations for continued development and constraint of physiological models for biogeochemical and ecological modeling. Zoe demonstrated that the data set provided good support for several key assumptions in our current macromolecular modeling approaches. She also introduced a North Atlantic scale database of field data on the macro-molecular composition of surface particulate, collected by the Bedford Institute of Oceanography over some decades (compiled by Mohammad Matlob, Iriwin/Finkel Labs). This remarkable and little-known data set provides a unique window on macromolecular allocation in the ocean that is expected to prove very valuable.
The Bertrand lab (Dalhousie) including Loay Jabre and alumnus Scott McCain provided a review and discussion of microbial protein allocation, with some particular discussion of allocation to ribosomes and the connection to macromolecular allocation models.
Contact: Mick Follows (MIT)
Regional & Global Modeling
Talks from investigators at U Hawaii, UCSC, MIT, and WHOI about the collection of circulation modeling activities under the CBIOMES umbrella drove discussion about current activities and future directions.
Brian Powell and Tobias Friedrich (U Hawaii) led presentations on the ROMS ocean model with a Darwin model ecosystem applied to the Hawai‘i region. Their talk showed how including fine-scale ocean circulation detail in a regional model produces solutions that are more plausible and qualitatively distinct from solutions derived from state-of-the-art climate prediction models. The fine-scale results (embedded in a 100-year climate forecast) are being used to ask questions about possible changes to the ecosystem mix in the coming century. Notably, the regional solutions show shifts in phosphate and nitrate concentrations in the upper 75m of up to +/-50%.
Paul Mattern (UCSC) presented an overview of the ROMS ocean model with a Darwin model ecosystem applied to the California Current region. Paul described results from a series of Darwin model configurations of different complexities and looked at how different sets of virtual species mix correlated with observations. The progression of complexity ranged from a six phytoplankton, four zooplankton functional group scenario to a scenario with 37 phytoplankton, 27 zooplankton functional groups.
Stephanie Dutkiewicz (MIT) presented a review of a range of MITgcm and Darwin model configurations with both global and regional applications. The setups have a range of different ecosystem model complexities, with different characteristics with respect to observations. They span experiments with a few phytoplankton classes and quota-based stoichiometry schemes to scenarios with 350 phytoplankton categories, 16 zooplankton, and spectral light. A range of virtual experiments have looked at the role of bacteria, the impact of diel cycling, the role of submesoscale and vertical motions in explaining the onset of growth, and multiple other questions. Stephanie’s talk also looked at a number of new capabilities being explored with the Darwin system, including coastal sediment schemes, diel cycling, heterotropic functional groups, enhanced representation of radiative scattering, and carbon storage reserve pools.
Greg Britten (WHOI) presented new work that he is just getting started in collaboration with others at Woods Hole, Boston, Santa Cruz, Halifax, and elsewhere. It is centered on the shelf-break region of the North-West Atlantic. This region includes the confluence of the Labrador Current and the Gulf Stream and contains some of the most highly productive and diverse marine ecosystems on the Earth. Observational records extend back to at least the early twentieth century with active in-situ and remote sensing programs ongoing and future campaigns planned. Greg highlighted how the ecosystem dynamics of the Georges Bank region might interact with variable dynamics of warm southern waters and cold northern waters and with persistent stable eddy events. These dynamics interact with large-scale marine food webs that support zooplankton and fish populations that have widespread impacts on Atlantic basin fisheries and marine life. Greg’s talk showed some preliminary results from modeling of the region that is currently under development.
Gael Forget (MIT) gave an insightful presentation on open-science approaches to collaborative research using modeling. He shared multiple examples of tools that can help grow effective and powerful open science projects. His talk reviewed a wealth of MITgcm and Darwin project open-science impacts that have accrued over many years. The slides looked at a specific example of reasoning about the Earth’s energy imbalance in the coming century and the ways in which modeling is being used to examine both natural system responses and geoengineering activities, including marine carbon dioxide removal. His talk presented a stepped plan for ensuring that work is reproducible in practice with minimal friction. Gael went over many of the most useful recent collaboration tools that are making open science practical and showed how he is applying these tools in his work and highlighted how the ClimateModels.jl initiative, that Gael is leading, is using these approaches to bring together many modeling tools, including MITgcm and Darwin model configurations, into a lightweight unified environment that helps remove challenges to sharing and to working across tools and discipline perspectives.
The closing discussion was wide-ranging. It spanned multi-petabyte physical and biogeochemical solution integration into CMAP, automating global and regional analyses following the statistical approaches that have been adopted in various CBIOMES papers, flexible deployment of Darwin system into different physical circulation models, coastal and shelf region investigations, connections to higher trophic level food webs. A few interesting themes for future activities were identified from these. These included (i) technical work on making the master Darwin code a flexible stand-alone tool, (ii) data work on incorporating model ecosystems and physical solutions into CMAP at full temporal and spatial resolution, (iii) ecosystem and food web modeling directions that include active shelf break to coastal margin ecological and physical dynamics, (iv) enabling flexible use of general Bayesian approaches with a range of Darwin system configurations. There was some subsequent discussion that a more extended multi-day workshop organized around some or all of these themes could be explored and might be of broad interest.
Contact: Chris Hill (MIT)
DAR 1
During the hands-on DAR1 workshop facilitated by DAR1 co-author Barbara Duckworth (MIT), attendees were able to download and run their own 0-D instance of the Darwin model. After running a default bottle experiment, participants set their own nutrient concentrations and saw how it affected plankton abundance. Participants then discussed other features and capabilities people would like to see in DAR1, such as default light and temperature, diel, and seasonal cycles, a 1-D water column, and adding tracers.
Contact: Barbara Duckworth (MIT)
The schedule, which lasted two and a half days, included pre-dinner poster sessions on each of the full days and an extremely engaging guest lecture from Ginger Armbrust “Deciphering the hidden worlds of microbes encountered during Gradients cruises.”
You can access posters, a recording of Ginger’s guest lecture, and group overviews here [login required]
Giant thanks to all the workshop leads for their help in preparing this write-up.
Note: The 2024 meeting is scheduled for June 25 – 27 in the Simons Foundation in NYC.