Ocean, Atmosphere, Ice, and Land - Bumps and Wiggles (8):
Ocean, Atmosphere, Ice, and Land - Bumps and Wiggles (8):
Introduction: This is the eighth (and last!) in a series on understanding climate variability, global warming, and what we might do about it. The series focuses on the past 30 years and the next 30 years.
From the beginnings of the WU Climate Change Blog there have been entries on “natural variability.” A number of known modes of natural variability have been discussed. The most “famous” source of natural variability is El Nino, which is a warming of the sea surface temperature in the eastern Pacific. This warming changes the atmosphere, with the impact of these changes realized throughout most of the globe. El Nino and its related cold phase, La Nina, are part of an oscillating behavior characterized by several attributes. First, this is an example of the ocean and the atmosphere varying together - in a correlated way. Second, the ups and downs of the oscillation are not “regular.” That is, the time period is not the same from one oscillation to the next and the peak and the duration of the warm and cold phases varies a lot. Third, despite significant observations and the development of useful theories, we do not fully understand the mechanisms that cause the El Nino – La Nina cycle. Hence, our ability to represent El Nino in climate models could be substantially improved. For the purposes of this article these attributes can be summarized as - a not fully quantified example of the oceans and the atmosphere irregularly varying in sync with each other.
There are many such features in the Earth’s climate; that is, not fully quantified examples of the oceans and the atmosphere irregularly varying in sync with each other. Other important ones are the North Atlantic Oscillation and the Pacific Decadal Oscillation. (Summary of such variability) When compared with El Nino, these other oscillations are less well quantified, less regular, and the relation between the atmosphere and ocean is more difficult to describe. Another characteristic of these modes is that they slosh back and forth with characteristic times of several years. Hence, these modes contribute to the bumps and wiggles in the observations that are the subject of this series of blogs. Implicitly and intuitively, when people start to think about climate “prediction,” as opposed to “projection,” it is how to account for these sources of natural internal variability. This is and will be a hard problem. There are a few Big Points I would like to make.
Big Point 1: One of the spurious statements about climate change, climate models, and climate projections is that it is impossible to forecast the climate because we know that “weather” is “chaotic,” and cannot be predicted beyond, approximately, 2 weeks. (Chaos and Weather Prediction from National Geographic) This misconception is, perhaps, based on the idea that the climate is “average weather,” and therefore, we have to predict weather before we can predict climate. Several points – 1) We might not be able to predict weather on an event-by-event basis beyond a few days, but we are likely to be able to (and can pretty well) represent the “average” weather and the correct amount of variability. 2) The definition of climate as “average weather” is, in fact, an inadequate definition of climate based on our experience of focusing on the weather and weather prediction and knowing the most about the atmosphere. The climate is all about the ocean and ice sheets and land and how they interact with the atmosphere and impact people. 3) So the details of Lorenz’s famous and powerful theory about chaos and the atmospheric weather is not exactly relevant to the problem of climate prediction – or more concretely ocean prediction. 4) We know from our El Nino experiences that the atmosphere responds in a, more or less, predictable way to changes in the ocean. (El Nino responses La Nina responses) Therefore, if we can predict the ocean better, we can predict, more accurately, what will happen in the atmosphere and how climate change will impact man.
Big Point 2: Sometimes I hear in the criticism of climate change science that climate models and modelers and scientists do not account for El Nino, the North Atlantic Oscillation and all of these sources of natural variability. That models and scientists do not account for these processes is simply an untrue statement. Climate models that include the interaction of the ocean and the atmosphere do contain modes of variability that are “like” El Nino, the North Atlantic Oscillation, etc. They are like the phenomena that are observed on Earth, but they are not an event-by-event representation of the Earth. And while they are like these events, our ability to simulate these events is far from perfect. Remember, I said earlier, these sources of variability are not fully quantified examples of the oceans and the atmosphere irregularly varying in sync with each other. When we look for signals in the models of warming by greenhouse gases, that warming is deemed significant above natural variability in the models that is comparable in magnitude of what is observed. see wiggles on figure in this blog. When all of the models are added together and averaged, as is often done in the IPCC figures, all of this variability averages out; the projection appears smooth. For the purpose of this series of blogs, the real observations of temperature have bumps and wiggles that are not in the projections that are the source of the discrepancies I have been writing about.
Big Point 3: There are certainly other modes of atmosphere-ocean-ice interactions that we have yet to observe. Back in the 1990s while I was at NASA, Arthur Hou and Andrea Molod published a paper looking at the interaction between the tropics and high latitudes. This has been a subject of many papers over the years. There was a result that they found that has stuck with me; namely, they found a high sensitivity of the transport of heat to the Arctic to what was going on with the deep convection in the western Indian Ocean. (Hou and Molod) The “deep convection” is responsible for those incredible and sometimes dangerous storms near the equator that drive the Hadley Cell. (Hadley Cell 1, Hadley Cell 2) The two aspects of Hou and Molod’s study that struck me were first, the high sensitivity mentioned above, and second, how difficult it is to model the western Indian Ocean. In recent years I have been impressed by the body of work that is provided by P. D. Sardeshmukh and his colleagues, in this case, especially Joseph Barsugli. In a two studies focused on tropical sea surface temperature patterns, they find great sensitivity of global weather patterns to, yes, the Indian Ocean. (2002, 2006) (These go along with much more studied sensitivity to patterns in the Pacific.) I, rightly or wrongly, even extract from this work a potential clue about that warm period in the 1930s that gets a mention every now and then. The point, there may be, almost certainly are, ocean-atmosphere patterns that we have yet to observe adequately, much less model. We will need to do this better as we understand the bumps and wiggles in the temperature observations. In my opinion, the most important places to focus to improve our modeling ability are on the West Indian Ocean and the Northwest Atlantic Ocean. These studies need to be process-focused and to explicitly focus of how the ocean and atmosphere (and in the Northwest Atlantic the outflow glaciers of Greenland) couple together. It is a fact of history that we focus most on the atmosphere, next the ocean, and only recently ice sheets and glaciers. We leave the “coupling” to just happen. We must focus on the science of coupling to explain the bumps and wiggles and to develop a predictive capability.
Big Point 4: There are modes of variability that we have not even thought about. We are just beginning to introduce ice sheets into climate models. We already know that there are strong connections between glacial flow, glacial melt, and the presence of warm sea water (Learning Abount Ice, Fast Ice – Redux, Sea Ice in Hot Water). But we have really not had an observing system that would measure all of the components, and those modes of variability would be expected to have long time scales. We are just beginning to get real ice sheet models linked into climate models, and that gives us the opportunity to do some numerical experiments. But even weak-minded modelers, ultimately, rely on data.
Short Summary: What I have posed in this series of blogs on short term variability is that the community of climate scientists need to have a research focus on short-term variability, where short term is a “few” years. By a “few” I mean a couple of years to maybe a few decades. The reason for this focus is not only to develop the foundation for decadal climate prediction, but to also provide the intellectual depth to better use climate model projections, as well as to strengthen the credibility of using climate model projections in applications. It is of critical importance to increase the focus on coupling of components in climate models, with particular attention to the processes important for variability on time scales longer than “weather.”
Bumps and Wiggles (1): Predictions and Projections
Bumps and Wiggles (2): Some Jobs for Models and Modelers (Sun and Ocean)
Bumps and Wiggles (3): Simple Earth
Bumps and Wiggles (4): Volcanoes and Long Cycles
Bumps and Wiggles (5): Still Following the Heat
Bumps and Wiggles (6): Water, Water, Everywhere
Bumps and Wiggles (7): Blackness in the Air
And here is
Faceted Search of Blogs at climateknowledge.org