Numerical Weather Prediction

Many observation systems like weather satellites and radiosondes are used to collect the necessary data required to make predictions using these models which are run in many countries globally. The models used are referred to as forecast models and they predict various weather variables like rainfall, pressure, temperature, wind and so forth (Richardson, 2007). A forecaster then assesses how these variables will interact to produce the day’s weather. There are mathematical models that use the same principles to come up with either short and long-term weather and climate predictions respectively. The climate predictions are instrumental in determining climate change. Later, in 1922, Lewis Fry Richardson attempted to come up with the first Numerical Weather Prediction but he was unsuccessful in doing so (Warner, 2010). The first successful attempt was in 1950, where several professionals i.

e. computer programmers, meteorologists and mathematicians used the ENIAC computer by simply approximating equations governing the atmosphere. Four years later members of the Swedish Meteorological and Hydrological Institute, led by Carl-Gustav Rossby came up with the first functional forecast. The system presently being used in many institutions like universities, private companies, NCEP, science laboratories and many more. The model has a global user base and this indicates that indeed it is a community model. By utilizing technological advancements, the model has adopted special capabilities like prediction of hurricanes, wildland fires, regional climates and so forth (Powers et al. In comparison to previous models like the MM5 this system has more numerical accuracy. The system has the capability to produce atmospheric simulations by first configuring a model’s domain, assimilating input, developing initial conditions, processing it then running the forecast model.

However, there are limitations to this option one of which is difficulty determining mild variations in temperature and wind while assessing wind convergence over the airport. This has necessitated use of more physical parameters for heat flux, near surface physics, more advanced numerical calculations for diffusion etc. In the WRF model there is a cumulus parameterization option which is not used in the Aviation Model because since the configuration of the AVM is in sub-kilometer resolution (Wong & Chan, 2013). This model will be very useful to pilots since it will provide information on sudden headwind changes during landing or when taking off. This is however partly hindered by the terrain around the airport that complicates wind convergence due to the near surface heat exchange. Therefore, when a change occurs in the atmosphere and it was not taken into account initially, it will be difficult to assimilate this data and this means the results will not be as accurate as required.

Another drawback is that the equations used to predict weather are not very precise. Moreover, the initial observations have many gaps because data may not be collected over areas like mountains and oceans. Concerning the future advancements in forecasting, one advantage is expansion of grids such that they incorporate non-hydrostatic resolutions of less that five kilometers so as to reduce the need for parameterization. Another expected advancement is to increase observational system so as to facilitate assimilation of more data and increase accuracy. Lynch, Peter (2008). The origins of computer weather prediction and climate modeling". Journal of Computational Physics. University of Miami. Powers, J. Grell, G. A. The Weather Research and Forecasting Model: Overview, System Efforts, and Future Directions.  Bulletin of the American Meteorological Society, 98(8), 1717-1737. Richardson, L. T.  Numerical weather and climate prediction.

Cambridge University Press. Wong, W. K.