Data assimilative microbial modeling and other model development and analysis for the Northeast Pacific

Chris Edwards
University of California, Santa Cruz

Marine ecosystem models generally combine numerical representations of basic processes like biological production and consumption with ocean circulation to simulate complex dynamics, usually governing photosynthesizers, their grazers, and the inorganic nutrients required for their growth. These models are necessary simplifications of nature, and accurate marine ecosystem simulation is subject to many unavoidable errors, including in processes, organisms, and in conditions used to initialize simulations.

The UCSC and UH groups focus on coupled physical and biogeochemical modeling and data assimilative studies in regional environments. We seek to apply these methods to the question of phytoplankton community structure in the north Pacific as well as pursuing general model development and analysis for understanding regional Darwin estimated microbial distributions in the northeast Pacific Ocean.

Data assimilative models use marine biogeochemical data to constrain coupled physical and NPZD-type ecosystem models to yield model output that more closely resembles ocean data than unconstrained model output. Both parameter estimation (Mattern et al. 2016) and ocean state estimation (Song et al. 2016a, 2016b, 2016c, and Mattern et al., 2017) are viable approaches. Our previous CBIOMES research has worked to make these methods more broadly applicable to any biogeochemical model (Mattern and Edwards, 2019, and Mattern et al., 2019). Most recently, we have investigated efficient methods that scale practically for the Darwin model (Mattern and Edwards., subm., and Izett et al., in prep).

These methods can now be applied to a specific problem in the northeast Pacific Ocean, namely the transition in Prochlorococcus concentrations observed in the Simons-funded Gradients cruises. High concentrations are found in warm, subtropical waters and low concentrations are observed in more northern waters. Debate presently exists regarding the cause of the transition. Flombaum (2013) argues that concentrations are statistically predictable by physical properties, largely temperature, whereas Follett et al. (2021) argue that temperature is not a strong predictor and that biological interactions, specifically co-predation of heterotrophic bacteria, influence Prochlorococcus population success. Parameter and state estimation methods previously developed in the CBIOMES project can be applied to bring the Darwin model closer to available observations allowing tests of hypotheses toward mechanistic controls on Prochlorococcus and other distributions. In addition, successful demonstration of the state estimation capability would represent a major step forward in ocean data assimilation, moving the field toward data-constrained predictions of presently unobserved quantities.

 

 

 

 

CBIOMES Collaborators working with Chris Edwards

News from the UCSC Group

Past Annual Reporting

2020 | 2019 | 2018  [login required]