CBIOMES Provinces Working Group

Contact: Chris Follett

The Provinces Working Group is a cross-institutional collaboration that grew out of discussions at the 2019 Annual Meeting. Drawing together CBIOMES investigators with expertise in statistical methods, physical and biogeochemical modeling, and remote sensing, the group is exploring questions concerned with spatial and temporal controls on marine phytoplankton distributions observed in nature and in models.

Current work has focused around defining a simplified, gridded, monthly climatology to explore methods for comparing models and observations utilizing provinces. Group members have used this data-set to test methods for objectively defining provinces (Christian Müller, Flatiron; Thomas Jackson, PML); used previously defined ‘Longhurst Provinces’ to compare model-data distributionally (Bror Jönsson, PML); and are developing transport methods to optimally partition deviations into quantity, space, and time components (Sangwon Hyun, Jacob Bien; USC). Additional comparisons have been done focused on temporal shifts (Marie-Fanny Racault, PML).

Moving forward, group members are keen to utilize objectively defined provinces and their boundaries to test the mechanisms which control the location and time-evolution of ecosystems and the biogeochemical cycles they control.

To find out more or to get involved, email follett@mit.edu

Papers in Preparation/ Under Review:

Sangwon Hyun, Aditya Mishra, Christopher L. Follett, Bror Jonsson, Gemma Kulk, Gael Forget, Marie-Fanny Racault, Thomas Jackson. Stephanie Dutkiewicz, Christion L. Mueller, Jacob Bien (2021), Ocean Mover’s Distance: Using Optimal Transport for Analyzing Oceanographic Data, arXiv: 2111.08736 [stat.AP] (accepted for publication in Proceedings of the Royal Society A.)

Abstract: Modern ocean datasets are large, multi-dimensional, and inherently spatiotemporal. A common oceanographic analysis task is the comparison of such datasets along one or several dimensions of latitude, longitude, depth, time as well as across different data modalities. Here, we show that the Wasserstein distance, also known as earth mover’s distance, provides a promising optimal transport metric for quantifying differences in ocean spatiotemporal data. The Wasserstein distance complements commonly used point-wise difference methods such as, e.g., the root mean squared error, by quantifying deviations in terms of apparent displacements (in distance units of space or time) rather than magnitudes of a measured quantity. Using largescale gridded remote sensing and ocean simulation data of Chlorophyll concentration, a proxy for phytoplankton biomass, in the North Pacific, we show that the Wasserstein distance enables meaningful low-dimensional embeddings of marine seasonal cycles, provides oceanographically relevant summaries of Chlorophyll depth profiles and captures hitherto overlooked trends in the temporal variability of Chlorophyll in a warming climate. We also illustrate how the optimal transport vectors underlying the Wasserstein distance calculation can serve as a novel interpretable visual aid in other exploratory ocean data analysis tasks, e.g., in tracking ocean province boundaries across space and time.

“A Distributional Comparison of Model and Satelite Chlorophyll Observations”, Bror Jönsson, C.L. Follett, J. Bien, S. Sathyendranath, S. Dutkiewicz, S. Hyun, G. Kulk, G. Forget, C. Müller, M. F. Racault, C. N. Hill, T. Jackson

“Quantitatively comparing the time-space evolution of the surface ocean across scale”, S. Hyun and other members of the CBIOMES Provinces Working Group.

Abstract: Many ocean datasets are large, multi-dimensional, and inherently spatio-temporal. Oceanographers are often interested in comparing datasets along one or several dimensions of latitude, longitude, depth, or time. In this paper, we introduce and extend optimal transport and Wasserstein’s distance as rich and useful tools for analyzing ocean data. Wasserstein’s distance improves upon existing common distance measures that conduct a pixel-by-pixel comparison. Such pixel-wise comparisons are inherently limited in detecting meaningful differences in the spatio-temporal regularity and multi-scale patterns that are common in ocean data. Furthermore, the optimal mass transports can provide a valuable visual aid for oceanographers while making ocean data comparisons. We demonstrate the usefulness of optimal transport and Wasserstein’s distance using the key application of comparing synthetic and remote-sensing ocean data.

Related Work:

Follett C. L., Dutkiewicz, S., Forget, G., B.B. Cael, and Follows, M. J., Moving Ecological and Biogeochemical Transitions Across the North Pacific, Limnology and Oceanography (Under Review)

Moving Ecological and Biogeochemical Transitions Across the North Pacific – SCOPE 2020 Annual Meeting Microtalk (login required)

(CBIOMES investigators highlighted in bold)

Explore Related Media

GitHub Page for matched Model-Data Climatology

CMAP Link for Full DARWIN Climatology (CBIOMES-Global)

2019 Annual Meeting Plenary: Defining Provinces (login required)