Sunday, November 18, 2012

The Ice Melts at Midlight

As part of the Climate Dialogue, there is the usual but but but, and one of the buts has been pointing to the role of clouds (Sgt. Schultz knows nothing,  nothing)

The past 25 years of carbon dioxide increase has contributed to less than 1 W m-2 in radiative forcing. There is no simple calculation to be made of the effect of the sea ice albedo change on the global energy balance. The better comparison is with a change in fractional cloud cover (which is highly variable), which dominates the planetary albedo far more than the arctic sea ice.
but to be fair, and you know that Eli is always fair, others, such as Rob Dekker point out that
Here, I noticed that GCM projections are lagging some 20 years (CMIP5) to 40 years (CMIP3) behind actual Arctic sea ice decline. Apparently, we are currently experiencing Arctic “climate” conditions that were not expected until another 40 – 80 ppm increase in CO2 concentration (which implies some 0.5 – 1.0 W/m^2 forcing).
And in the middle of this, the Rabett remembered something, the further north you go, the larger the annual CO2 cycle

Using flask sample data from Alert (in the blue, purple is Mauna Loa) , it's about a 15 ppm swing from early spring (April, May) to late summer (August, September) and about 8-10 ppm more of a swing than ML.

And the latest measurements from Barrow show a difference of 20 ppm

 Applying the normal equation for CO2 forcing, 5.35 ln (C/Co), where Co is 280 ppm, that is a difference of a bit more than 0.25 W/m^2, in other words, significant when compared to the total average increase of 0.5 to 1 W/m^2 in the past 25 years.  Moreover, the net effect is to limit cooling in the winter as compared to using the average carbon dioxide mixing ratio, increase melting in the spring and force less melting in the late summer.  Over the year, the average concentrations are about the same at ML and Alert.

It takes about six to seven years before the low mixing ratio catches up with the high at Alert (e.g. the minimum in 2007 matched the maximum in 2000.

To tell the truth, Eli has not a clue as to whether this is included in models or evaluations of the Arctic ice, but it looks interesting.  Just thought the bunnies might like to know


David B. Benson said...

Thanks Eli.

Limit cooling in the winter: maybe delays some winter ice formation?
Increase melting in the spring: start the death spiral earlier?
Less melting in late summer? Huh?

You ought to ask over on Real Climate whether the models include this interesting observation.

Anonymous said...

That's been known for almost like ummm.. since Anthony Watts was born. I don't know if a plea for incompetence is enough if that hasn't been factored in. Reason: more active greenery at low latitudes.

Anonymous said...

well not quite:

Anonymous said...

William Connolley said...

Its an interesting point. The models certainly don't bother with location-dependent CO2.

I doubt it matters. If the atmosphere was a collection of single columns, it might. Since it isn't, and the circulation is quick, my guess would be the effect is too small to bother with.

Russell said...

Yes, but though the circulation is quick, the midnight sun is slow to set , amplifying the seasonal impact on regional temperature

EliRabett said...

Wm. the mixing is slow enough that regional difference appear in the column even on an annual scale which is where this guess came from. Certainly the ML annual cycle as opposed to the lack of a cycle shows this as do the differences btw NH and SH. True, your column has to have a pretty wide base. .

Aaron said...

Likewise, there is now some seasonal CH4 (not included in models). see for example

And, for the last few summers, there was some H2O vapor (that is not in the models).

EliRabett said...

Water vapor is intrinsically in the models as relative humidity

jyyh said...

using unconventional maths again.

Is this a case of 'as we can't model plants/animals perfectly, let's leave the effects they seasonally produce out of the model.'? That 20ppm annual cycle is about what, 8 years of anthropogenic emissions? Well ok, statistics says one has to take (missing the link to Tamino) c.16 years of data in order to find out if the Globe has warmed, so the equivalent of 8 years of local anthro emissions has no effect on the conlusion. So, do we always have 8 years to pray for a miracle? That's conveniently 2 election cycles so everything fits. :-(.

Paul S said...

I've had a peek at some files at PCMDI and William appears to be correct that all the standard runs use one global value for CO2 concentration in all grid squares.

However, CMIP5 includes some 'ESM' experiments for models with interactive carbon cycles. In these realisations CO2 concentrations are regulated by the climate live in the model, so you do get geographically-heterogeneous values and seasonal cycles. A few things from this, just looking at a single annual cycle:

- In the model the difference from peak to trough in the seasonal cycle (monthly resolution) at Barrow's latitude is ~17ppm. However, the amplitude actually decreases from there to the Pole, where it is ~12ppm.

- In the model the higher amplitude does not carry vertically. In the upper troposphere it is closer to ~7ppm.

These two factors mean the radiative forcing change is not quite as simple as stated in the post, and the ~0.2 figure is likely an overestimate.

As a general point, the seasonal change in downwelling shortwave radiation around the Arctic is something like 400 W/m^2 from peak to trough. Seems to me the CO2-cycle effect is a few orders of magnitude too small to make a noticable difference.

Anonymous said...

You'd have to look at the annual cycle in temperature in the arctic to know if there is any appreciable effect of the annual cycle in CO2.

The response of air temperature to change in forcing is not instantaneous, so the real question becomes "is the response fast enough to have an effect on such a short time scale?"

The upper troposphere might respond quickly, but the air immediately above the ice (and open water) would take longer to heat up, so a brief change in forcing due to annual CO2 cycle may not even show up in the surface air temperature.

Or perhaps the effect is actually indirect, with cloud cover being affected (reduced due to warming of the upper troposphere) which in turn affects the amount of sun reaching the surface.