Lijing Cheng, relaxing with KangaroosHeating of the oceans is, pardon the pun, a hot subject. There is a broad recognition that the oceans, which absorb approximately 90% of excess greenhouse gas energy, are key not only to how fast the planet will warm, but also how hot it will get in the end. Many recent studies have tried to measure deeper ocean regions or previously uncharted areas in the search for heat. A new study by Lijing Cheng and Jiang Zhu takes a different approach. They ask how large are biases in the estimates of ocean heating from the finite resolution of the devices themselves. Their findings are exciting, but first, let’s talk about the details of the study.
Lijing Cheng, relaxing with Kangaroos
Measuring the oceans is difficult; they are vast and deep. In order to measure the total energy in the ocean, you have to obtain temperatures at many locations and at many depths. Not only that, you need to make measurements over many years if you want to identify long-term trends.
Oceanographers have been making such measurements for decades. But the density of measurements is not spread uniformly over the oceans. There are more in coastal regions or major shipping lanes than in other locations. To further complicate the problem, the measurement methods have changed over the years.
Decades ago, insulated buckets, then, bathythermograph devices, and now ARGO floats have been used. While these devices all go down into the ocean depths, they have different depth resolutions. Over the years, we have a large number of measurements near ocean’s surfaces but as we measure deeper and deeper, fewer and fewer data points are available. As a result, we cannot construct exact temperature-depth curves. Consequently, our discrete data points give us some error, some bias compared to real ocean temperatures.
In their paper, Lijing Cheng and Jiang Zhu quantify our ocean errors. They started with a “real” ocean temperature and then they extracted discrete data and asked themselves how their discrete data matched the original temperatures. By discrete data, I mean that they extracted temperatures every 10 meters, 20 meters, 30 meters, and so forth. Somewhat like the science of calculus where smooth curves are approximated by a series of straight-lined segments.
What they found was very interesting. In the upper regions of the oceans, the discrete data was colder than the real ocean temperatures. However, deeper in the waters, the trend reversed and the discrete data was warmer. But to make things more complicated, the errors differed depending on location in the oceans. Near the equator (tropics), the discrete data exhibited a warm bias but further from the equator, the bias was cold. Furthermore, the extent of the error changed throughout the year.
The authors use their findings to calculate how close together measurements would have to be to obtain accurate ocean temperature measurements. The authors also propose a method to correct past temperature data to account for these biases. It is important for readers to recognize that the biases themselves do not make us think climate change will be worse or milder. What really matter are the changes to biases over time.
Dr. Cheng, who recently graduated with his PhD, told me,
Posted by John Abraham on Monday, 23 June, 2014
 |
The Skeptical Science website by Skeptical Science is licensed under a Creative Commons Attribution 3.0 Unported License. |