Climate Prediction Coprocessor

of 26
17 views
PDF
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Climate Prediction Coprocessor
Document Share
Document Tags
Document Transcript
  “Grand Challenge” Problems : Computational Climate PredictionA possible solution :A Low Complexity Dynamic Grid Coprocessor for Distributed ComputingDo you see any design logic or any other !ind of error Do you see any design logic or any other !ind of error ContactContact“Climate Prediction CP#”“Climate Prediction CP#”dist$% at &uno dot comdist$% at &uno dot com http:''hireme(gee!(n)' “Climate Prediction Coprocessor”   *umerical weather and climate prediction uses !nown weather conditions as  baseline inputs to extrapolate “short term” weather forecasts and “long term” climate outcomes(+nitial weather and climate conditions are fed into an array of simplified “ensemble models” of the atmosphere typically with only a few minor ,alue changes between hundreds of models( Due to chaos theory the slightest change in a model-s input parameters typically yields hundreds if not thousands of ,astly different outcomes(Large numbers .do)ens if not hundreds/ of climate and weather simulations must  be run .and a,eraged/ to filter or “sort out” unrealistic ,ariance( #se of model ensemble forecasts helps to define the forecast uncertainty and extend weather and climate predictions farther into the future than would otherwise be possible(0inor changes in model inputs and outputs can be used to extrapolate historical climate conditions .called hindcasting/(0anipulating these huge datasets 1as well as performing the complex prediction calculations necessary at a resolution high enough to ma!e the results useful2 re3uires ,ery powerful computers( 4upercomputers are ideal for this tas! but ,ery few supercomputers are a,ailable to use for this tas!( 5istory6ritish mathematician Lewis 7ry 8ichardson first proposed numerical weather  prediction in 9$$ although his initially unsuccessful attempts at the method date bac! to ;;+( 8ichardson attempted to perform many !inds of low complexity numerical forecasts before ;orld ;ar ++ but success was ne,er achie,ed(<he first successful numerical prediction was performed in 9=> by a team composed of the American meteorologists ?ule Charney Philip <hompson Larry Gates and 8agnar 7&@rtoft and applied mathematician ?ohn ,on *eumann using the *+AC digital computer(<hey used a simplified form of atmospheric dynamics based on the barotropic ,orticity e3uation( <his simplification greatly reduced demands on computer time and memory so that the computations could be performed on the relati,ely primiti,e computers a,ailable at the time( Later models used more complete e3uations for atmospheric dynamics and thermodynamics(Bperational numerical weather prediction began in 9== under a &oint pro&ect by the #4 Air 7orce *a,y and the #4 ;eather 6ureau .later *BAA/( http:''hireme(gee!(n)' “Climate Prediction Coprocessor”  5ow is modern numerical forecasting done  *umerical atmosphere models are initiali)ed using data collected from radiosondes weather satellites and surface weather obser,ations( <hese irregularlyspaced obser,ations are processed by data assimilation and ob&ecti,e analysis tools which  perform 3uality control chec!s that filter out misbeha,ing sensors( Guesstimate data ,alues are extrapolated for locations that are physically unreachable by the sensor net( xtrapolated data points .due to corrupt sensors or lac! of data/ are imputed using !nown and expected seasonal ,ariances(<he so called atmospheric “primiti,e e3uations” are then applied to the atmospheric state to find existing and new rates of change( <he !nown rates of change predict the atmosphere in a future state( <ime stepping procedures are continually repeated until the solution reaches the desired forecast time(<he length of the time step is related to the distance between the points on the computational grid( <ime steps for global climate models may be on the order of tens of minutes while time steps for regional models may be %> seconds to %> minutes(Long term problems see!ing solutions0ost climate models simulate a region of the arth-s atmosphere from the surface to the 4tratopause( <here also exist numerical models which simulate the wind temperature and composition of the arth-s tenuous upper atmosphere from the mesosphere to the exosphere including the ionosphere(<his region is affected strongly by the 99 year 4olar Cycle through ,ariations in solar # ' # ' Eray radiation and solar wind leading to high latitude particle precipitation and aurora( +t has been proposed that these phenomena may ha,e an effect on the lower atmosphere and should therefore be included in simulations of climate change(7or this reason there has been a dri,e in recent years to create Fwhole atmosphereF models to in,estigate whether or not this is the case(<he atmosphere as well as the world-s oceans .that influence and buffer atmospheric trends/ are dynamic fluids( <he basic idea of numerical weather prediction is to sample the state of the fluid at a gi,en time and use the e3uations of fluid dynamics and thermodynamics to estimate the state of the fluid at some time in the future( http:''hireme(gee!(n)' “Climate Prediction Coprocessor”  Contrasts (goals)  *;Ps .predict weather/GC0s .predict climate/4patial co,erage8egional or GlobalGlobal<emporal range% Days to $= ;ee!s$ ears to 9= Decades4patial resolutionariable .$>9>> !m $ /#sually H $>> !m $ 8ele,ance of initial conditions5igh0edium to Low8ele,ance of clouds radiation5igh0edium8ele,ance of surface .land ice ocean(((/0edium to Low5igh8ele,ance of ocean dynamicsLow5igh8ele,ance of model stability0edium to Low5igh<ime dimensionssential7or comparison with other model runs<ime 4tep$ minutes to $> minutes%> minutes to I> minutesPhysics3uations of motion .plus radiati,e transfer e3uations water conser,ation e3uations ((/0ethods7inite difference expression of continuous e3uations or spectral representationJ run prognostically0aximum time stepControlled by spatial resolution .C7L condition/http:''hireme(gee!(n)' “Climate Prediction Coprocessor”
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x