Aerosol and Clouds
High-resolution ensemble surface insolation estimates through assimilation of coarse-scale retrievals into a simple physical model: 2. Ensemble implementation and Shortwave Radiation Budget data assimilation
Article first published online: 26 MAY 2007
Copyright 2007 by the American Geophysical Union.
Journal of Geophysical Research: Atmospheres (1984–2012)
Volume 112, Issue D10, 27 May 2007
How to Cite
2007), High-resolution ensemble surface insolation estimates through assimilation of coarse-scale retrievals into a simple physical model: 2. Ensemble implementation and Shortwave Radiation Budget data assimilation, J. Geophys. Res., 112, D10219, doi:10.1029/2006JD007873., and (
- Issue published online: 26 MAY 2007
- Article first published online: 26 MAY 2007
- Manuscript Accepted: 24 JAN 2007
- Manuscript Revised: 4 DEC 2006
- Manuscript Received: 2 AUG 2006
- surface insolation;
- data assimilation
 A simple surface insolation model was used in conjunction with the Ensemble Kalman Filter (EnKF) to ultimately derive an ensemble of high-resolution posterior insolation fields conditioned on the GEWEX Continental Scale International Project and GEWEX Americas Prediction Project (GCIP/GAPP) Shortwave Radiation Budget (SRB) product. When compared to ground-based observations, the ensemble of prior estimates was shown to reasonably capture the space-time variability and uncertainty in surface insolation except in cases where the Visible Infrared Solar-Infrared Split Window Technique (VISST) product did not properly detect clouds. The EnKF was chosen to provide a systematic framework for merging the insolation information contained in the prior estimates with the well-validated SRB product. When compared to independent ground-based observations, the posterior ensemble mean estimate showed a significant reduction in error relative to the prior. The resulting posterior ensemble mean error statistics are of similar magnitude to other recent studies at similar spatial resolution. The reduction in error is highest in cloudy-sky conditions (where prior uncertainty is largest) and lowest in clear-sky conditions (where insolation is much easier to predict). A significant value added through application of the data assimilation scheme is a posterior estimate that provides an ensemble that captures the complex space-, time-, and state-dependent uncertainty that would otherwise be very difficult to obtain using a more traditional forward modeling approach.