The talks on the second day were very interesting and were mainly about the status of the science, future challenges and uncertainties. The earth’s climate is just inherently complex . These notes are also partly for my benefit !
1. Clouds : David Randall
Clouds are hard to model. They scatter and emit raditaion. They also transport energy, moisture and momentum large distances. Cloud models go back to the 60s but have remained very much at the macro scale and this is one of the main problems because cloud formation depends more on micro-physics. Cloud seeding is from natural and anthropogenic aerosols and models lack resolution at these scales.
Globally the net cooling effect of clouds is about -20 W/m2. The water content of clouds is 100 times smaller than the water vapour content of the atmosphere, yet their effect on climate is huge. Large convective rain clouds tend to be on average radiation neutral while low clouds are strongly cooling and can be caused by sinking air from convective areas. Cirrus clouds have a net warming effect. CMIP5 models have an average positive cloud feedback of ~ 0.6 w/m2/degC. There are two arguments for a net positive feedback.
- Anvil hypothesis: Tropical anvil convective clouds flatten off at a fixed temperature of 200K so emit the same IR independent of surface temperature. Therefore as surface temperature rises IR radiation doesn’t – a net positive feedback .
- Low clouds diminish with warner temperatures ( low confidence).
It is entirely possible that cloud feedback is in fact negative and this is the largest current uncertainty of GCMs.
2. Aerosols: Olivier Boucher
Aerosols have 3 main effects:
- They scatter incoming solar radiation cooling the earth.
- They (e.g. black carbon) absorb both incoming solar radiation and surface IR radiation
- They help seed clouds formation – net cooling effect.
Energy imbalance where is the aerosol feedback.
Models trade off aerosols against Climate Sensitivity to match observed temperatures. Aerosols are essentially the tuning parameter that match GCMs in hindcasts to previous surface temperatures.
Globally 20-40% of aerosol optical depth is of anthropogenic origin. Somewhere between 1/4 to 2/3 of cloud condensation is on nuclei concentrations of anthropogenic origin. The amount of black carbon emitted in Asia is underestimated but probably overestimated elsewhere. One question that occurred to me was whether we know if the early 19th century temperature measurements in Europe were suppressed due to the wide-scale burning of coal both in households and industry. Is some of the warming after 1956 not actually due to the clean air acts across Europe ? Weather station data were concentrated in Europe and N. America before the early 20th century. All CMIP5 models use the same historic aerosol trends. Just how well are these trends really known ?
There is no correlation of cloud cover with cosmic rays. The statement for the Hiatus was that 50% of the pause can be explained by natural variation ( 1/3 volcanic, 2/3 solar). The other 50% is of unknown origin !
Aerosol-clouds are estimated as a negative feedback ~-0.45 Wm-2deg.C-1. The total effective radiative forcing due to aerosols is ~ -0.9 Wm-2.
3. Carbon + Geochemical Cycles: Corinne le Quere
CO2 is 40% above pre-industrial levels. An extra 180 Gtons of Carbon has been added to atmosphere. Sources of atmospheric CO2 emissions which are currently running at 10 Gtons/year :
- Deforestation = 2.9 Gtons/year regrowth = 1.3 Gtons/year
- Emissions is 2/3 of rise or 6.6 Gtons/year
- Land use is 1/3 of rise or 3.3 Gtons/year
4. Weather extremes: Denns Hartman
The AR5 assessment plays down the risk of climate catastrophes. Increases in extremes are based on
- warming is small ~0.6C since 1950
- Corresponds to a 4% change in saturated vapour pressure (C-C equation)
- making statements about changes in extremes is very difficult but studies show (based on datasets HADEX2 HADGHCND ) that warm nights have increased 4.5 ± 0.9 % and warm days increased by 3± 1.8 %
Weak statements can be made about precipitation. In general wet areas get wetter and dry areas get dryer. There is little evidence that droughts have got worse. There is also low confidence that flooding also has got worse.
- Tropical Storms: No significant change
- Intensity of storms: No significant change.
5. Model confidence : Peter Stott
Very much a party line talk this one. Humans are the “dominant” cause of warming since the mid-20th century. It is extremely likely(95% confidence) that more than HALF of the observed warming is anthropogenic. See table 10.1 for justification for this statement. A summary of forcing causes is shown in Figure 10.5 below.
Presumably then up to HALF of the observed warming could also be natural. He accepts that solar output and volcanic aerosols are forcing agents which can explain half the hiatus, but he dismisses natural internal variability (PDO, ENSP, AMO etc) as being only of order ± 0.1 W/m2.
Kenneth Trenberth then asks the question: Given the observed 15 year hiatus in global warming how can natural forcings be just zero ± 0.1 ? Why is ENSO not included in assessments ?
Peter Stott gave a hand waving non-answer to this point. Has Trenberth become a skeptic ?
6. Circulation: Ted Shepherd
His point is that all GCM models really deal with is energy balance locally. They do not really handle circulation at all and exhibit severe biases. The importance of dynamic circulation is the horizontal transport of moisture, energy and momentum. A severe limitation is understanding the dynamic circulation aspect of climate.
- North Atlantic Oscillation is natural and so is are other internal variability.
- Hadley cell has been widening but it narrows during el Nino.
- WCRP grand challenge on Clouds, Circulation and Climate Sensitivity.
- CMIP5 are not true ensembles and probability distribution funstions do not apply
7. Paleoclimate models: Gavin Schmidt
CMIP5 models have been applied to understanding climates of the past. This means using the known orbital parameters, CO2 , Ice cover , sea-level that were prevalent then.
- Tier 2: Mid-holocene (6000 – LGM 21 K years ago)
- Tier 3 last Millenium
He claims that overall they work well but there are model dependent divergences in Sahel, SW US etc.
However he admits that none of the models are able to reproduce the on-set of an Ice Age or predict the next Ice Age ! They are static representations rather than dynamic.
8. Politics & Propaganda: John Ashton
I won’t say much about this as I already commented on the politics in the previous post.