Mixed Signals from the NOAA ENSO Blog about Climate Models

There’s a new post at the NOAA ENSO blog titled Climate Change and ENSO: Take 2.  It’s a guest post by Matt Collins of the University of Exeter.  That post probably left most people scratching their heads…not for its content, but for its purpose.

PREFACE

We’ve posted and discussed two studies over the past few years that indicate climate models are incapable of modeling ENSO. And it’s not a question of the model abilities to replicate the instrument temperature record, which they can’t.  The problems run very deep. The climate models used by the IPCC for attribution studies and projections of future climate cannot simulate the basic coupled ocean-atmosphere feedback in the tropical Pacific that underlies ENSO.  It’s called Bjerknes feedback.  It’s the positive feedback relationship between the strength of the trade winds and the surface temperature gradient (cooler in the east, warmer in the west) of the tropical Pacific.  Stronger trade winds yield a larger temperature gradient. And a larger temperature gradient yields stronger trade winds.  The two are interdependent, providing positive feedback to one another.

There are a multitude of other problems, which we’ll now briefly mention.

The two papers that present model failings at simulating ENSO are:

• Guilyardi et al. (2009) Understanding El Niño in Ocean-Atmosphere General Circulation Models, and
• Bellenger et al (2012) ENSO representation in climate models: from CMIP3 to CMIP5.

You’d be hard pressed to find any component of ENSO that climate models simulate properly. One of the findings of Guilyardi et al. (2009) was that climate models do not use enough sunlight, and that’s a key finding because sunlight provides the fuel for ENSO.  That is, ENSO acts as a sunlight-fueled recharge-discharge oscillator. Introducing more sunlight to the models would create an obvious problem for climate modelers: more sunlight undercuts the modelers’ reliance on infrared radiation from manmade greenhouse gases and allows sunlight through ENSO to explain more of the warming from the mid-1970s to the turn of the century, a period when ENSO was skewed to El Niño dominance.

Because of all those model failings, Guilyardi et al (2009) include the following note:

Because ENSO is the dominant mode of climate variability at interannual time scales, the lack of consistency in the model predictions of the response of ENSO to global warming currently limits our confidence in using these predictions to address adaptive societal concerns, such as regional impacts or extremes (Joseph and Nigam 2006; Power et al. 2006).

The section titled “Discussion and Perspectives” in Bellinger et al. (2012) begins:

Much development work for modeling group is still needed in order to correctly represent ENSO, its basic characteristics (amplitude, evolution, timescale, seasonal phaselock…) and fundamental processes such as the Bjerknes and surface fluxes feedbacks.

“Amplitude” refers to the strengths of ENSO events.

“Evolution” refers to the formation of El Niños and La Niñas and the processes that take place as the events are forming.

“Timescale” can refer to both the how long ENSO events last and how often they occur.

“Phaselock” refers to the fact that El Niño and La Niña events are tied to the seasonal cycle. They peak in the boreal winter. (See the post Why Do El Niño and La Niña Events Peak in Boreal Winter?)

“Bjerknes feedback,” which we discussed above, very basically, means how the tropical Pacific and the atmosphere above it are coupled; i.e., they are interdependent, a change in one causes a change in the other and they provide positive feedback to one another.  The existence of this positive “Bjerknes feedback” suggests that El Niño and La Niña events will remain locked in one mode until something interrupts the positive feedback.

“Surface fluxes” refers to the variations in heat exchange between ocean and atmosphere in the tropical Pacific in response to ENSO. (See the National Oceanography Centre webpage here, for a quick overview.)

In short, according to Bellenger, et al. (2013), the current generation of climate models (CMIP5: used by the IPCC for their 5^th Assessment Report) still cannot simulate basic coupled ocean-atmosphere processes associated with El Niño and La Niña events–basic processes, really basic processes.

REGARDLESS OF ALL THOSE FAILINGS

Some scientists found that some models simulate some part of precipitation in the tropical Pacific reasonably well in response to some parts of ENSO.  They then go on to write papers about the future of ENSO. That means they’re overlooking the fact that the models don’t simulate other ENSO functions properly, and the precipitation portion they’re focusing on was achieved under conditions that do not relate to ENSO processes as they exist in nature.  So the precipitation-based ENSO studies are basically meaningless.  We discussed one of those studies in the post Will Global Warming Increase the Intensity of El Niño?

BACK TO THE NOAA ENSO BLOG POST

The post begins:

Tom previously touched on how climate change might affect ENSO, emphasizing the 2013 AR5 Intergovernmental Panel on Climate Change (IPCC) statement (footnote 1), which basically said that ENSO will continue, but we don’t know if or how its frequency or intensity might change.

That’s the answer we get when we look at the question head-on. But what about when we look at it more indirectly? Looking at other elements of the climate system, for example, we can focus on one reasonably robust finding: an intensification of mean rainfall in the central and eastern equatorial Pacific is ‘likely,’ according to the IPCC criteria. Recently, my colleagues and I have been focusing on the possibility that these overall rainfall changes may impact ENSO.

Answers at last?

There have been a number of modeling studies…

That’s as far as I originally read for reasons discussed above.  I came back to the NOAA ENSO blog post later because I’d been asked to comment about it. The body of the post was about how ENSO might change in the future if the climate models were correct…which they can’t be.

THEN THERE’S THE CLOSING

The first paragraph of the closing begins:

But is that the final answer?

Just because the change is seen in models, or even a subset of models, doesn’t mean that we should believe it without question. Not only do we have to factor in errors or biases in the models, but we also have to have a convincing physical argument for the changes.

That quote’s a keeper.  File that one away.  It’s applicable to climate models in general.

Then there’s the final paragraph:

So, the picture of changes in ENSO, when viewed in terms rainfall response patterns, may be limited by errors and biases that have been long-term features in climate models. Research is required to test the potential impact of SST biases on the change in average precipitation in the tropics. We must improve models, but we must also to better understand the processes whereby biases in present-day simulations link to future projections. Until we get a better handle on these issues, the prediction of an overall increase in rainfall in the eastern tropical Pacific, and its year-to-year variability, remains uncertain.

One wonders why he presented possible changes to ENSO, when he knew the modeled answers were wrong.

I believe it was Richard Lindzen who once wrote that modelers should have first started with the oceans, if they wanted to understand climate on this planet…or something to that effect.  That’s becoming more and more obvious to more and more people.