Click images for more details



Recent posts
Recent comments

A few sites I've stumbled across recently....

Powered by Squarespace

Discussion > Models uncertainty

In recent posts about model uncertainties Doug McNeall of the met office pointed us to a new paper on The HadGEM2 Family of MetUM climate configurations, or the met office climate model for the rest of us.

There are a lot of improvements on this model, but I would like to concentrate on a couple of them just now. I want to stress that these are new changes, the AR4 scenarios were based on past versions of the model, and so were a very large body of literature and policy recommendations.

If we look at summary table A2, some of the changes are:

* Land surface emissivity reduced from 1.0 to 0.97

* Land ice, snow on vegetation, and ocean albedo all reduced by 5%

and mainly "achieve closer top-ofatmosphere radiative balance."

Surface emissivity controls the amount of long-wave radiation emitted (and absorbed) by the surface, which can be calculated according to the Stephan-Boltzman Law. This prescribed change in emissivity would result in a change of more than 11 Watts in the radiative balance, assuming a global land surface temperature of 15C (288.15K). Let's ignore for the moment that there is not such a thing as global emissivity, deserts and forests have very different values, or tundra and snow, granite or limestone, etc.

Then a 5% change in ice and snow albedo is a big change, that will probably make the skiing season in the Cairngorms vanish, or will melt half the glaciers in the Alps. A 5% change in ocean albedo means a variation of 12.5 Watts in the earth radiative balance.

Let's recall that the Net anthopogenic radiative forcing ranges between 0.6 and 2.4 Watts, and it turns that the models designed to evaluate the impact of such a forcing were up to 22.5 Watts off the mark (or 15.8 if we consider only 30% of land surface).

Given the very wide range of adjustments needed, it looks to me that we are very badly equipped with these models to detect the influence of any anthropogenic radiative forcing.

Aug 8, 2012 at 10:28 AM | Registered CommenterPatagon

I found nothing about the handling of radiation in the writeup. Are radiative effects modelled or do they come in a black bx. Maybe I skipped that bit or it was described for previous versions and this was only changes. Anyhow, I asked on a thread where Doug was present but he must have missed it, so again, is the radiation bit modelled or parameterised?

Oh, good points, Patagon. Can we get a comparison with the equivalent numbers in other GCMs? They can't all be right, can they?

Aug 8, 2012 at 10:35 AM | Registered Commenterrhoda

My concern is that the numbers don't even agree with the previous version of the model.

The 3-hourly steps for for radiative calculations are based on Edward and Slingo (1996)

according to

The Physical Properties of the Atmosphere in the New Hadley Centre Global
Environmental Model (HadGEM1). Part I: Model Description and Global Climatology. Journal of CLimate, 19,1274

The radiation code is that of Edwards and Slingo (1996) used in HadCM3 with some developments.

J. M. Edwards, A. Slingo
Studies with a flexible new radiation code. I: Choosing a configuration for a large-scale model
Quarterly Journal of the Royal Meteorological Society Volume 122, Issue 531, pages 689–719, April 1996 Part A

It also mentions:

Cusack S, Slingo A, Edwards JM, Wild M (1998) The radiative
impact of a simple aerosol climatology on the Hadley Centre
climate model. Q J R Meteorol Soc 124: 2517±2526

Aug 8, 2012 at 11:10 AM | Registered CommenterPatagon

Let me put it in a different way.

We have twisted several knobs in the new models to make them perform better according to observations. This change means that at least we should add error bars to the older model. The error bars are about 22.5 Watts wide (or at least a minimum of 15W, just for two of these knobs) in the radiative balance.

How can we say anything meaningful about a human related forcing of 0.6 to 2.4 W when the error bars are bigger than 15 W?

Aug 10, 2012 at 11:57 AM | Registered CommenterPatagon

I note Edwards and Slingo has been declared not quite good enough by the BOM in Oz. So now what? I have no problem at all with modellers modelling and trying to get ever closer to observed reality, but they are not ready, nowhere near ready, to have those models used to predict anything or to influence policy. And yet here we are using a bunch of models, not the best but an assemblage designed not to offend any national team, to do just that.

Aug 14, 2012 at 3:16 PM | Registered Commenterrhoda


You are by a large margin far better than me in terms of being able to feed data into computer code and produce models and I have only admiration for those who are trying to model our climate. However I do not understand why you can not see the point Rhoda makes; you are simply no where near understanding enough about our climate to model it in a way which can predict the future.
May I ask you a question?
Of all the factors which we will eventually discover to be affecting climate, what percentage of that ultimate understanding do you think you currently have?

Aug 14, 2012 at 4:08 PM | Registered CommenterDung

I don't think Patagon is claiming anything except the size of the error bars which ought to be there, but are not.

Aug 14, 2012 at 6:15 PM | Registered Commenterrhoda

Patagon, no one seems to have pointed out any errors in the values you pointed out. As you say how can variations that large in parameters critical to the energy balance be used to investigate future temperature changes which depend only on the energy balance in the model. Things like the dynamics of the model will obviously effect energy balance terms, but the changes you've pointed out will completely swamp these kinds of effects.

Aug 27, 2012 at 11:35 AM | Unregistered CommenterRob Burton

Rob, you are completely right. The changes are quite clear in the paper I referred to, and the implications huge, as you point out. I am still waiting for an explanation from someone with better knowledge of the actual parameterisations of climate models. Though from my knowledge of other similar models, I suspect the changes are real, and the implications simply forgotten.

Aug 29, 2012 at 3:57 PM | Registered CommenterPatagon