Arctic | Lower Trophic Levels
Comparison between models to estimate phytoplankton size fractions in the Chukchi and Bering Seas Presenter: Lisa Eisner, lisa.eisner@noaa.gov, Alaska Fisheries Science Center, NOAA Fisheries Presenter: Calvin Mordy, calvin.w.mordy@noaa.gov, CICOES, University of Washington, Seattle, WA, United States Jens Nielsen , jens.nielsen@noaa.gov, Cooperative Institute for Climate, Ocean, and Ecosystem Studies, University of Washington Priscila Lange , prilange@gmail.com, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil Michael W Lomas , mlomas@bigelow.org, Bigelow Laboratory for Ocean Sciences, Maine, United States J eanette Gann , jeanette.gann@noaa.gov, NOAA Alaska Fisheries Science Center Sage Osborne , sage.osborne@noaa.gov, Pacific Marine Environmental Laboratory, Seattle, United States Dale Robinson , dale.robinson@noaa.gov, NOAA Southwest Fisheries Science Center, La Jolla, United States Phyllis Stabeno , phyllis.stabeno@noaa.gov, NOAA Pacific Marine Environmental Laboratory Ecosystem shifts in response to climate change in the Arctic are mediated by changes in phytoplankton stocks and types. As the main primary producer in this ecosystem, phytoplankton size governs the fate of the carbon fixed by these photosynthesizers, with large cells and cell agglomerates sinking faster and supporting benthic life; whereas smaller cells fuel planktonic and neritic production. Therefore, understanding the temporal and spatial variability of phytoplankton size is key to diagnosing potential changes in Arctic ecosystems. Here, we compare 4 empirical models which can be used with satellite ocean color data to estimate phytoplankton size-fractionated chlorophyll across the Bering and Chukchi seas. Two models (Brewin et al. 2011, Hirata et al. 2011) originally used total chlorophyll as the predictor for three size-fractions (< 5, 5-20 and > 20 μm). These models, with their original parameters, were compared to models which were regionally-tuned using 266 in-situ surface chlorophyll samples for each size, collected in 2017 and 2019. In addition, sea surface temperature (SST) was tested as an additional predictor of these size classes. Models based on the same equations were also regionally-tuned (1425 samples from 2003 to 2019) to estimate the contributions of two chlorophyll size fractions (< 10 and > 10 μm). Two other models (GAM and Random Forest) were also tested using the same predictors and with additional environmental forcing (e.g. temperature, season, latitude). Comparison among models show the regionally tuned models perform best regardless of the established size fractions, with total chlorophyll being the most important predictor while SST does not appear to improve the models. The implementation of these models in a time-series of chlorophyll images can help us understand the effects of long-term environmental change in the Bering and Chukchi ecosystems, and aid the prediction of future trophic scenarios, which is necessary for ecosystem management.
Alaska Marine Science Symposium 2023 146
Made with FlippingBook flipbook maker