chapter 4 summery Flashcards
Model domain refers to:
- Model’s forecast area of coverage
Numerical models divide into:
- Global models
- Covering the whole globe
- Regional models
- Covering a more limited area
Limited area models (LAM) boundaries
- Horizontal (lateral)
- Top and bottom (vertical)
Global models boundaries:
- One vertical boundary (by nature cover the entire earth)
Domain of an NWP model can be viewed as:
- 3D array of cubes
Describe the NWP model 3D array cubes:
- Each cube encompasses a volume of the atmosphere corresponding to a model grid point
- Forecast values for met variables in each cube are derived from
- The current values within the cubes + from the surrounding cubes
Information needed to provide forecast values for the meteorological parameters. Why?
- Cannot be determined using only the data contained in the model
- Because he cubes on the boundaries are not surrounded by other cubes on all sides
How to solve the problem?
- The information from the outside boundaries must be supplied from another source (boundary conditions)
Global model merits:
- Global coverage
- Don’t require boundary conditions
- Necessary at longer lead times when weather at a location is effected by distant weather system
Global model drawbacks:
- Coarse spatial resolution
- Need parameterization of sub grid scale physical processes
- On a regular latitude-longitude grid, the grid boxes will become smaller in the longitude direction near the poles (high possibility of CFL condition violation)
Regional models merits:
- Higher resolution
- Higher spatial resolution
- Do not require parameterization of some physical processes
- Higher spatial resolution
Regional models drawbacks:
- Require boundary conditions
- (for boundary conditions) regional models depend on global models
- Parameterization required for physical processes smaller than the grid size
Requirements to solve the forecast equations:
- Accurate information
- for all forecast variables and
- Along each model boundary
- Lateral
- Top
- Bottom
How is the lateral boundary conditions data supplied to LAM?
- Using large-domain models
Boundary values are obtained from:
- Observed data
- Data assimilation system
- Forecast values from a current or previous cycle of a large scale model (LBC in LAM)
- Climatological or fixed values
- For specifying some surface characteristics such as
- Soil moisture
- SST
- Vegetation type
- For specifying some surface characteristics such as
Favorable source of boundary values is:
- Observed data
Favorable source of lateral boundary conditions values is:
- Previous run of a large domain model
Factor effecting the quality of LAM predictions
- Quality of predictions produced by the model supplying the LBC
Consequence of errors in forecast from large domain model:
- Error will move into the LAM’s forecast domain and amplify
LBC control:
- Position
- Evolution of features that cover the entire forecast domain
Longwave patterns are largely determined by
- Boundary conditions
Weaker impacts are noted on
- Jet streaks and fronts
- Regions far downwind from the upstream boundary
Synoptic scale model supplying the boundary conditions determine
- The placement and
- Timing of synoptic scale features
Influence of boundary conditions:
- Spread away from (downstream) of the boundary
- Effects amplify downstream
Preferable area of primary forecast concern:
- As far from the boundaries as possible
- Especially the upstream boundary
Boundary influence is carried by:
- The wind
…………………….. and ……………………. Will vary from one flow regime to another
- The wind and
- The direction of greatest forecast impact
Forecasters should pay attention to
- How long it takes a trajectory to move from the model boundary to the vicinity of their forecast area
How to reduce the influence of boundary errors within the area of interest:
- Placing the upstream boundary well outside the forecast area
One-way interaction:
- If information flows in one direction, from larger domain model to the smaller domain model
Nested model:
- Some LAM (eg. MM5) are run with small-area, finer-resolution grids nested inside of coarser-resolution grids within the same model
Nesting is necessary because:
- Computer memory ad speed limitations prohibit fine-resolution grids from covering the entire model domain
How information from outer boundaries are supplied to nested grid models:
- Supplied from an outside source using one-way interaction
Interface between the grids inside the nested grid model are determined from:
- The forecasts within the model itself
The forecast variables for the coarse grid are updated based on:
- Fine grid prediction
- Where the fine grid covers the coarse grid
Two way interaction:
- Coarse-grid prediction affects the fine grid prediction by
- Supplying boundary conditions on the mesh surface
- (this information flows both ways)
- Supplying boundary conditions on the mesh surface
Weather forecasting is:
- An initial value problem (IVP):
- The outcome is significantly determined by the conditions given at the start
To produce a forecast for tomorrow you need:
- To start by determining, as precisely as possible, the state of the atmosphere today
The initial conditions describe:
- The state of the atmosphere (at grid points) at the starting point of the forecast
Initial conditions include:
- The 3D fields of the forecast variables (u,v,w,T,p..) over the model domain
The initial conditions are obtained from:
- The observed data and
- Previous forecast
Through a process known as data assimilation
Local winds upstream and downstream:
The same direction of local wind is downstream boundary while if it is opposite to the local wind it is upstream
