// lecture 23 Flashcards
why is CO2 highly correlated to ice volume?
incomplete answer: colder oceans can dissolve more atmospheric CO2.
- but possibly more plankton active taking CO2 out of atmosphere and/or seawater exchange between surface and deep was greater
Younger Dryas (YD)
- example of rapid climate change
- 14700 kbp, the warming trend reversed
- relatively cold period lasted about 2,000 years
- warmed very abruptly about 12,000 years ago, and has been relatively stable since
cause of YD
- probably by ice sheet breakup and flooding in the northern N. Atlantic
- meltwater pulse could cause the thermohaline circulation to shutdown
- reducing heat transport into northern n. atlantic
numerical weather prediction (NWP)
- improvement in weather prediction data over the last 60 years is among the most impressive accomplishments of society; 3, 5, and 7 day forecasts have improved in both NH and SH
Lewis Fry Richardson
- made the first numerical weather prediction in 1922
Richardson’s Dream: The Forecast Factory
- filled with employees (“Computers”) doing calculations; estimated 64,000 people would be necessary to forecast over the global
Richardson’s Experiment
- used data from - May 20, 1910 and made Leipzig charts for surface pressure and temp.
- data was taken when haley’s comet was passing through; all values were tabulated by hand.
Richardson’s calculations
- took 10,000 hours of work to perform calculations
- book lost during a battle, but eventually recovered and published in 1922
Richardson… failure or success?
- first prediction was for pressure to change by 145 mbar in 6 hrs… that would be record setting
- he realized that noisy wind data was likely the problem and suggested 5 diff. filtering methods to fix this
- but he couldn’t try this experiment again, need computer’s of today
computer forecast w/ richardson’s proposed fix
- becomes a good forecast with his filtering methods used
first weather prediction on computers
- 1950: Charney, Fjortoft and von Neumann paper
- May 1955: Joint numerical weather prediction unit, maryland: first operational computer forecasts in US
- global coverage since 1973
- computers surpassed human forecasts: 1980’s?
ENIAC computer (1943-56)
- 17,468 vacuum tubes, was 1000 times faster than that of previous computing machines
- original computer weather forecast done in 1950 can now be done on a cell phone
weather vs climate forecasting
- similar bc use similar mathematical equations
- but weather is short term; climate is long term
Chaos Theory
- Ed Lorenz was running a computer model and put in slightly different inputs; found the predictions were similar for a while but then wildly diverged to diff. solutions
Butterfly effect
- Ed Lorenz; weather forecasts depend on initial observations but climate models don’t
Climate forecasts
examples
- summer is hotter than winter, after a strong volcano erupts, the earth will cool, the earth will be hotter with more GHG’s, shifts in weather patterns when El nino is present
Suki Manabe: father of climate modeling
- gradually builds up more sophisticated climate models:
- radiation only model (LW and SW): M. and Moller (1961)
- above plus convection: M. and Strickler (1964)
- model with atmospheric motions (but no ocean yet): Smagorinsky, M. and Holloway (1965)
first coupled climate model
Manabe and Bryan (1969)
first global warming forecast
Manabe and Wetherald (1975)
other early manabe studies
- effect of ocean circulation on climate: turn off ocean model
- effect of moisture: don’t allow condensation to occur
- effect of mountains: bulldoze all topography
- effect of changing soalr radiation, doubling CO2, ice sheets, clouds, soil moisture, etc.
GCM (global climate model) Components
- equations of fluid motion on a rotating sphere
- both the atmosphere and ocean are just fluids
- equations put simple physics principles in mathematical form
parts of climate model - 1968
- uses laws of physics
- momentum equation
- heat equation
- water equation
- for both atmosphere and surface and later ocean too
change in resolution overtime
FAR -> TAR -> SAR -> AR4
within each grid cell
there are things that are not explicitly modeled (clouds) that must be approx. or parameterized
cloud schemes
cloud interactions are the most uncertain process in GCMs, lead to the largest differences between models
