// lecture 24

Question 1

K Computer - fastest known

Answer 1

housed at the riken advanced institute for computational science in Kobe, has 672 computer racks and 68,544 CPUs

Question 2

2012 update

Answer 2

says there’s more powerful computer called Sequioa at the Livermore National Laboratory and can do 16.3 petaflops compared to K Computer’s 10.3 petaflops

Question 3

2014 update

Answer 3

• Tianhe2 is even faster.

Question 4

2012 update: increase with time

Answer 4

• factor or ten increase in speed about every ~5 years.
• doubling every 18 months, Moore’s Law
• log scale

Question 5

validating climate models

Answer 5

• we validate forecasts incrementally as we go along, we are experimenting in real time
• we have a challenge to estimate how accurate 2050 or 2100 climate forecast will be

Question 6

other ways to validate climate models

Answer 6

• how much cooling after a volcano?
• can we reproduce the last ice age conditions given CO2, solar, etc.
• can the climate of the 20th century be repdocued given GHG, solar, volcanoes, and aerosols

Question 7

cloud feedback

Answer 7

• Clouds have a strong impact on the radiation balance
  of Earth
• Reduce OLR by about 30 Wm-2
• Reduce Absorbed Solar Radiation ~ 50Wm-2
• Net effect about -20 Wm-2
  -The effect of doubling CO2 is only about 4Wm-2
• If their radiative effects change with global warming,
  the effect could be a very significant cloud feedback.

Question 8

cloud feedback continued

Answer 8

• If clouds get brighter and reflect more solar radiation
  as the climate warms, that would be a NEGATIVE FEEDBACK
• If clouds get higher and colder and trap more
  outgoing longwave radiation as the climate warms, that would be a POSITIVE FEEDBAC

Question 9

cloud feedback in models

Answer 9

• varies because of SW cloud feedback, the reflectivity of clouds changes as the climate warms
• mostly models say clouds will reflect less radiation as the climate warms, a pos. feedback
• but this is uncertain

Question 10

other successful predictions of climate models

Answer 10

• more warming at night than day
• most warming in arctic than anywhere else (Esp during winter)
• least warming in/around Antarctica
• wet gets wetter and dry gets drier
• tropopause (at the top of the weather layer of the atmospehre) moves upward
• large scale tropical circulations weaken

Question 11

CMIP5 models and ENSO

Answer 11

• many models do a good job simulating ENSO variability

Question 12

different CO2 scenarios give different warmings

Answer 12

business as usual gets about 4C warming by end of century of 8C by 2300.

• a couple possible scenarios keep warming under 2C
• model spread gives an estimate of uncertainty of these projections