Skip to content

Latest commit

 

History

History
35 lines (24 loc) · 3.42 KB

README.md

File metadata and controls

35 lines (24 loc) · 3.42 KB

This is under development.

objFeatures

pytest workflow

We built objFeatures to enable users to calculate multiple object tracking measures. objFeatures first developed as an extension of the morphological image processing Python package Ocetrac, where we worked with tracked unique geospatial features (i.e., sea surface temperature extremes, particle tracer tracks, and chlorophyll-a concentrations among others) in gridded datasets, such as climate model simulations, reanalysis datasets, and observations. The motivation is to compare characteristics between multi-dimensional objects and to compute distances between the objects to understand similarities and streamline the data analysis.

Use case: marine heatwaves

We apply objFeatures on sea surface temperature anomalies (SSTAs) in CESM Large Ensemble climate model projections, which consists of 100 ensemble members at 1° spatial resolution covering 1850 to 2015. Our multi-dimensional SSTA objects are called marine heatwaves (MHWs). Classifying and finding patterns in the spatiotemporal evolution of these moving SST ‘groupings’ can fill multiple knowledge gaps, such as our understanding of key MHW characteristics like distribution, variability, and trends, and the physical mechanisms that cause MHWs in different parts of the ocean. We can then generate statistics from our MHW groups, which will allow us to examine the global and regional scale drivers of MHWs. This pipeline allows this analysis to be replicated for (1) different regions and (2) using different datasets.

Scientifically, this project will add merit to the field by allowing us to 1) assess the possible physical processes causing MHWs, 2) to define the the statistical properties of MHWs, including their varying intensities, duration, and spatial extents and 3) to characterize the spatio-evolution of connected MHWs. More generally, this project will advance our mechanistic understanding of the MHWs, relate one MHW to another, and our understanding of MHW evolution.

Screen Shot 2023-03-09 at 3 16 35 PM
Figure 1. Three-dimensional visualization of marine heatwaves over time

Installation

How you can contribute

  • You can get involved by trying objFeatures, filing issues if you find problems, and making pull requests if you make improvements.

Wishlist

Acknowledgements