SourcesofErrorinNWPForecasts.pptVIP

Sources of Error in NWP Forecasts or All the Excuses You’ll Ever Need Fred Carr Introduction NWP has become an indispensable tool for the forecaster, but it is important to understand its limitations. There are many sources of possible error in an NWP forecast. If you keep these sources in mind as you examine NWP products, you should be able to make more intelligent use of the products in your forecasts. These sources of error can be grouped into three categories: A. Errors in the Initial Conditions It is a complicated process to collect data from observations and get them into a form that an NWP model can use. Errors can occur at several steps along the way, as well as grow out of limitations of the data sources themselves. Intrinsic Predictability Limitations Errors in the Initial Conditions Observational Data Coverage Spatial Density Temporal Frequency Errors in the Data Instrument Errors Representativeness Errors Errors in Quality Control Errors in Objective Analysis Errors in Data Assimilation Missing Variables Errors in the Models Equations of Motion Incomplete Errors in Numerical Approximations Horizontal Resolution Vertical Resolution Time Integration Procedure Boundary Conditions Horizontal Vertical Terrain Physical Processes Precipitation Stratiform Precipitation ii Convective Precipitation Radiation Surface Energy Balance Boundary Layer Surface Layer ii Ekman or Mixed Layer Intrinsic Predictability Limitations Introduction B. Errors in the Model A model is by definition an approximation of reality, and although NWP models continue to grow in complexity, they cannot take into account all factors that affect the weather. The numerical solution of these models by computers introduces additional error. C. Intrinsic Predictability Limitations Even with error-free observations and a “perfect” model, forecast error will grow with time. There is an intrinsic limit to the range of a useful forecast. This range is short for small-scale pheno