U.S. Navy sailors repair NOAA weather instrument.
U.S. Navy sailors repair a National Oceanic and Atmospheric Administration weather instrument buoy in the Atlantic Ocean. The fourth in a series of marine climate data workshops addressed issues in developing useful archives of marine weather and surface ocean measurements from diverse data sources that span several centuries. Credit: U.S. Navy

Since the 1600s, mariners have measured the weather and surface ocean conditions as part of daily operations of merchant and naval vessels. Early observations were primarily visual estimates of weather conditions and later included measurements from early versions of weather instrumentation, including thermometers, wind vanes, and barometers. In the latter half of the 20th century, scientists developed new technology, including moored and drifting buoys, gliders, and autonomous profiling floats, to further measure environmental conditions near the ocean surface.

A diverse mix of historical and modern marine measurements provides the basis of our understanding of the climate over the world’s oceans.

This diverse mix of historical and modern marine measurements provides the basis of our understanding of the climate over the world’s oceans and is the foundational data used to model past, present, and future climate. Developing homogeneous collections of weather and surface ocean measurements is critical to support ongoing global climate research.

Last July, 51 scientists met at the National Oceanography Center in Southampton, UK, with an overall objective to recommend actions to update and improve available marine climate data sets and products. Such improvements will support future marine climate research and operational activities. Topics at this workshop, dubbed MARCDAT-IV, included data homogenization, quantification and estimation of data biases and uncertainty, applications of marine climate data, comparison of marine data from various observing platforms, and the future of ocean observing technology.

The key data set that workshop participants discussed was the International Comprehensive Ocean-Atmosphere Data Set (ICOADS). This resource is recognized as the premier marine climate database; the latest version contains data from 1664 to 2014, with near-real-time updates to the present time.

A word cloud that synthesizes more than 1400 published manuscripts, covering a wide variety of marine climate topics. These manuscripts cite previous versions of ICOADS, a marine climate database that spans more than 300 years.

MARCDAT-IV focused on the diverse mix of platforms, instruments, and human observing practices that can be found in ICOADS. Participants emphasized the need to develop new approaches to data homogenization, that is, integrating and standardizing data from disparate sources. Approaches include those based on multivariable techniques, statistical methods, and processes that take advantage of techniques developed by the land surface climate community.

Homogenization techniques rely on access to quality metadata (data about the data). Attendees reiterated the continuing need for better contemporary and historic metadata to meet the needs of the marine climate community. In particular, they noted an appalling lack of metadata for moored and drifting buoys available from some national meteorological services (NMS). Participants also recommended that countries continue to support long-term climate reference stations to measure essential climate and ocean parameters.

Participants recommended that countries continue to support long-term climate reference stations to measure essential climate and ocean parameters.

Looking forward, participants expressed concern about the expanded use of autonomous observing platforms and the desire by NMS to deploy automated weather systems (AWS), noting that such new technologies could add new discontinuities into the marine climate record. Attendees recommended that AWS and new autonomous systems be tested and compared to existing observing systems prior to deployment. Also, AWS should measure multiple climate parameters (e.g., winds, pressure, air and sea temperature, and humidity), not just sea level atmospheric pressure, as is being recommended by some NMS.

Finally, attendees provided feedback on ICOADS to the international steering committee that guides ICOADS development, focusing on addressing the needs of a wide user community. Recommendations include

  • providing better documentation that is searchable, easily updated, and comprehensive
  • enhancing data distribution services to create data subsets and provide observations in additional formats used within the research community
  • improving data quality control
  • issuing more frequent updates

The recommendation for more frequent updates arose because participants noted the nearly 10-year interval between ICOADS releases 2.5 and 3.0; ideally, annual updates would be desirable. These recommendations all focus on improving the user experience and modernizing ICOADS to maintain its role as the foundational data set for marine climate research. The meeting report, program, and presentations can be found on the meeting website.

—Shawn R. Smith (email: smith@coaps.fsu.edu), Center for Ocean-Atmospheric Prediction Studies, The Florida State University, Tallahassee; and David I. Berry, National Oceanography Centre, University of Southampton Waterfront Campus, UK


Smith, S. R.,Berry, D. I. (2017), Expanding a 300-year record of marine climate, Eos, 98, https://doi.org/10.1029/2017EO068241. Published on 13 March 2017.

Text © 2017. The authors. CC BY 3.0
Except where otherwise noted, images are subject to copyright. Any reuse without express permission from the copyright owner is prohibited.