Treatment of Clouds in Atmospheric Models
Clouds and precipitation consist of many types of water particles ('hydrometeors'): cloud droplets, rain drops, ice crystals, hail stones, snowflakes…. The particle sizes range from micrometres for cloud droplets to millimetres, sometimes centimeter for snowflakes.
Many things can happen in the cloud. For example, water vapour condenses on a cloud droplet; colliding drops merge and form a larger drop coagulation; snow particles collect duplets. These interactions are called cloud microphysical processes.
Cloud proceses can be modelled by two approaches.
- model variables are the number of particles for each size class ('bin')
- cloud processes can be formulated with high precision
- results are very accurate
- computing time is very high
- usually used in specialised models
- model variables are bulk properties such as total number of particles or total condensate mass per volume
- cloud processes can only be formulated approximatively and with assumptions
- results for the bulk properties are reasonably accurate but no information on microstate
- computing time is low
- usually used in e.g. weather forecast models
Clouds in a Weather Forecast Model: Using the Bulk Approach
In this example we see huge differences between the models – these are the changes introduced when using another distribution function or different predicted moments (N – number, L -- mass, Z – reflectivity).
We improve bulk models and validate their results with bin model results as benchmark.
When formulating the equations for the bulk approach, some choices have to be made:
- a mathematical function to approximate particle size distribution,
- the number and type of bulk quantities to use.
The consequences of these decisions and assumptions are best studied in simple models.