Effectively synthesizing multi-scale Earth and Sun system model output with measurements is at the core of much NCAR science. Observations are used to develop theories, confront model results, and, through assimilation techniques, adjust those results. Remote sensing from space now provides essential global-scale information on the atmosphere, and novel sensor networks are being developed that will provide new unique and dense observations, supplementing traditional observations. Model representation of difficult-to-observe processes can be improved by examining the mismatch between models and corresponding forecasts based on assimilation of observational data, particularly satellite observations and spectrally resolved images of the Sun and its magnetic field.
There is an emerging opportunity for NCAR to serve the community by developing and supporting numerical tools and strategies for integrating measurements and models. This process relies on new, flexible methods of data assimilation in which heterogeneous sets of physical measurements can be combined with geophysical models to both yield better predictions and detect model biases. This activity has two distinct benefits: models can augment the often-sparse coverage of observations, and high-resolution observations can diagnose strengths and weaknesses of a physical model and its supporting parameterizations. This frontier will also support new instrument design by providing a framework in which the community can assess the ability of novel observations to improve prediction or elucidate imperfectly understood physical processes. We plan to