As climate change intensifies and starts to cause changes in the water column (e.g., increased temperatures and stratification) mesophotic communities (ca. 30 m to 150 m depth) will become exposed to climate change impacts that regularly influence shallow-water coral communities (i.e., < 30 m). Mesophotic coral reefs, ecosystems that have received increased attention in recent years due to new, cost-efficient ways of quantitatively studying them, contribute to the ecosystem services of U.S. coral reefs valued at US$3.4 billion per year (Brander and van Beukering 2013). Habitat suitability models help us statistically predict where important species occur in difficult to survey regions, like the ocean, and contribute to spatial planning. They have been applied to cold-water coral (Davies and Guinotte 2011) and used for mesophotic coral habitats in Hawaii (Costa et al. 2012). These models require a large amount of information and computing power to run, especially if there are ranges of environmental parameters that are tested (e.g., temperature ranges that reflect climate change scenarios). Being able to transition this work to high performance computer clusters like UNITY would transform the habitat suitability models I can run and increase the impact of my work (e.g., test how varying climate change factors impact model outcomes in a matrix) on how climate change will impact mesophotic coral community distribution in the northern Gulf of Mexico. Part of my dissertation will help improve the Flower Garden Banks National Marine Sanctuary’s marine spatial management of ecosystems that they are mandated to protect and preserve.
Project Information Subsection
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
University of Rhode Island -- Bay Campus
{Empty}
CR-University of Rhode Island
{Empty}
Yes
Already behind3Start date is flexible
6
{Empty}
{Empty}
{Empty}
{Empty}
Milestone Title: Milestone #1 Milestone Description: Determine project scope, HPC access, presentation of goals, and set up the project on GitHub/UNITY Completion Date Goal: 2024-01-01
Milestone Title: Milestone #2 Milestone Description: Gather and organize the data and create a logical file structure. Completion Date Goal: 2024-02-01
Milestone Title: Milestone #3 Milestone Description: Complete a preliminary species distribution model. Completion Date Goal: 2024-03-01
Milestone Title: Milestone #4 Milestone Description: Test and debug, utilize output file structure for model runs. Completion Date Goal: 2024-05-01
Milestone Title: Milestone #5 Milestone Description: Wrap up development, final model run, update project and documentation, and final presentation. Completion Date Goal: 2024-06-30
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
Final Report
The CAREERS Cyberteam grant helped establish the computational aspects of the project (i.e., using computer clusters and running models using R code) for me and introduced me to a wider range of computationally based projects that helped me as I worked through issues. The principal discipline, ecological oceanography, requires computational methods to make sense of data and the world -- and developing distribution models for important and vulnerable mesophotic coral ecosystems is a great contribution.
The impacts likely extend to various earth science disciplines that think about the natural world in computational ways (e.g., grid formats, environmental variables, resolution) because increasing resolution size can help us understand greater intricacies and processes in the world -- but also lead to higher uncertainty in our predictions.
{Empty}
{Empty}
{Empty}
{Empty}
{Empty}
Modeling the distribution of coral reefs and how they might shift in climate change is important to society because of the ecosystem services that reefs provide - from fish habitat, to nutrient cycling.
We learned how to (1) gather and build the data layers required for habitat suitability modeling, (2) build the code base to process and run the data through distribution models, (3) use UNITY computer cluster to troubleshoot and run the models in an efficient manner on powerful machines, and (4) interpret preliminary results. We made progress in Python and arcpy, R, and UNITY interactive sessions.
We created a habitat suitability model for mesophotic coralline algae beds in the Flower Garden Banks National Marine Sanctuary at high-resolution (5 m) for the first time.