A Data Analysis of Ocean Temperature Trends and Anomalies

Open Access
Barkley, Zachary Robert
Area of Honors:
Bachelor of Science
Document Type:
Thesis Supervisors:
  • Chris Eliot Forest, Thesis Supervisor
  • Chris Eliot Forest, Honors Advisor
  • Marcelo Chamecki, Faculty Reader
  • oceans
  • climate
  • temperature
  • trends
  • anomalies
Oceans play an essential role in Earth’s global energy budget, storing greater than 90% of excess energy accumulated by the planet since the industrial revolution. Sparse sampling of data in the ocean, especially at deeper depths and prior to the year 1950, makes analysis of the entire ocean difficult. This study focuses on the density of ocean temperature data as a function of time, space, and depth, and on conclusions that can be drawn from the available data. Temperature trends are also calculated where sufficient data are available. All available data profiles from Expendable Bathythermographs (XBTs), Mechanical Bathythermographs (MBTs), High-resolution Conductivity-Temperature-Depth (CTD), Ocean Station Data (OSD), and Argo Float Data (Argo) are used in this analysis. A bias correction is applied to the XBT and MBT profiles to account for errors in its fall-rate equation. To calculate trends, data are divided into grids of 5°× 5° latitude by longitude boxes, and temperatures are averaged in these boxes over 5 year intervals. A least squares approximation method is then applied, and an overall temperature trend is calculated for the grid. It is found that strong warming can be observed over the last 60 years across much of the surface, with the North Atlantic experiencing the largest heating. Warming trends weaken with depth until 500 meters, where a more gradual decrease takes place down to the ocean bottom. In terms of data coverage, data becomes progressively sparser with depth past 500 meters, with the poorest coverage in the Pacific Ocean. The deployment of Argo floats alleviates many of these gaps in data down to 2000 meters, but results in trend calculations more biased towards recent years. At depths greater than 2000 meters most temperature trends become either insignificant or incalculable. Both the North Atlantic and Southern Ocean are exceptions to this rule. Future data provided by the recent installment of the Argo network should allow for better documentation and analysis of the ocean down to 2000 meters in depth, but gaps in data will continue to affect the uncertainty of analyses at deeper levels.