About Meteoroloo.com

Meteorology is the scientific study of the atmosphere that focuses on weather processes and forecasting. Meteorological phenomena are observable weather events which illuminate and are explained by the science of meteorology. Those events are bound by the variables that exist in Earth's atmosphere. They are temperature, pressure, water vapor, and the gradients and interactions of each variable, and how they change in time. The majority of Earth's observed weather is located in the troposphere.

With model output approaching observational data (e.g. from satellite soundings) in resolution, the sheer size of the datasets means that data mining and data management will become equally important considerations in meteorological computing. In light of the decrease in density of surface and rawinsonde observations, new algorithms have to be developed to extract similarly accurate information from satellite data, for example about cloud type and distribution. Data management will become more global in nature, with some central archives storing a large number of numerical experiments from various institutions. These data need to have a sufficient amount of metadata attached and can then be conveniently retrieved by a WWW interface from anywhere. These new archives will alleviate the important task of comparing experiments conducted with different models, which is instrumental for their further improvement. Also, grid computing may be an interesting way to harness the power of meteorological supercomputers more effectively. Of course international cooperation is nothing unusual in modelling, but grid computing might automate the process of running a model where the right amount of computing resources are currently available and leave scientists more time for analyzing the results. Meteorological instrumentation that is used at the surface or in airplanes also has room for improvement. Radar and lidar show precipitation and clouds by their effects on emitted monospectral electromagnetic waves. If radar measurements can be used to accurately determine the amount of precipitation (which as of now is only possible with rain gauges), this would be beneficial for numerical weather prediction. Lidar can be used to study clouds that are so thin that they cannot be seen by the naked eye such as certain types of cirrus filaments. Researchers continue to find new atmospheric details such as high-altitude clouds that can form from contrails, which suggest that air travel may affect regional weather. Improvements in understanding forecast uncertainty are being made by the use of ensemble forecasting (using slightly different starting conditions to make several forecasts). This was pioneered operationally in 1992 by both the European Centre for Medium Range Forecasting (ECMWF) and the US National Centers for Environmental Prediction (NCEP) which use different methods for generating each ensemble member. These techniques have allowed forecasters to understand when features in a forecast are nearly certain or rather unlikely and also allow extreme events to be picked up further in advance. This method of forecasting is still in its infancy and its use seems likely to become more advanced and widespread in the future. Targetted observations (also known as adaptive observing systems) also seem likely to reduce errors in forecasts. Calculations are made to find where extra observations would most improve a forecast for a given time and place. Then extra observations from that region can be obtained using aircraft, satellite, dropsondes or other novell observing systems to reduce the uncertainity in the analysis of that area. When these starting conditions are used to generate a forecast then errors can be reduced in the pre-defined region of interest. Experimental programs have used this technique , but it is not in operational use anywhere to date. This method has the potential to greatly reduce errors in forecasts in any numerical forecasting system.