Flashcards in Climate modelling Deck (28)
Why do we use climate models?
-prediction or projections
What is a climate model?
A computer programme that simulates the 3D behaviour of the atmosphere, oceans and land surface
What do computer programs integrate?
Various equations to determine the time and space-deponent evolution of the climate system
When is numerical integration used?
When an equation is too complicated to solve analytically
What elements are calculated in models?
*double check this future emily*
-Atmoshpere and ocean circulation
-Atmoshperic radiation budget
-Hydrology and water phase change
-Energy flow in rocks and soils
What is the AGCM?
Atmospheric General Circulation Model
What it the OGCM?
Ocean General Circulation Model
What is the AOGCM?
Coupled Atmosphere-Ocean General Circulation Model
What is the ESM?
Earth System Model
How have models progressed?
-AGCM ... prescribed the oceans and land surface
-OGCM ... simple atmospheric calculations (wind stress and heat supply)
-AOGCM ... used extensively in climate change experiments
What must you do before running a climate model?
Define the initial conditions
What is discretisation?
The splitting of continuous quantities into discrete units e.g. spilling the earth up into a grid cells for models
What is the typical resolution of models?
1º x 1º x 60 altitudes levels .... equivalent to 125 km x 125km x 100's m
What is the problem with discretisation?
-Numerical diffusion (flattening of graphs)
-Don't have information everywhere on the grid
-causes a reduction in resolution (detail)
What is parameterisation?
The simplification of processes in terms of simpler equations with physically or empirically derived parameters
How is rainfall parameterised in models?
Models calculate where humidity is above 100% and takes excess as clouds ... the reality of the process is more complicated than this
What is The Unified Model?
A model used to predict weather and climate
What is the difference between climate models and weather forecast models?
Climate models are 'free running' (with every time step there are new observations to work off) whereas Weather forecast models need to be 'nudged' to match observations (data assimilation)
Why has the time taken to run a model never really changed?
Although the rate of calculations have increased, new aspects have been added making the models more complicated
How do we assess how well models can make projections?
The ability of models to capture past observations measures their ability to make future projections (if they've been accurate in the past they will likely be accurate in the future)
What it the CMIP?
The Couple Model Intercomparison Project .... pooled together more than 40 models
Was the hiatus predicted by models?
Most didn't however some ensemble members do capture it
What is a model ensemble?
A set of simulations with slightly different starting conditions
How good are models at predicting extreme weather?
Simulations has improved but with large regional discrepancies
How good are models at predicting intense rainfall?
Frequency of intense events and amount of precipitation are underestimated
How good are models at predicting extra tropical storms?
Well captured in models with significant improvements since 2001
How good are models at predicting tropical cyclones?
Well captured in models with significant improvements since 2001 (IPCC)