Diagnostics for Evaluating the Impact of Satellite Observations
The adjoints of the numerical weather prediction (NWP) model and data assimilation system may be used together to objectively determine the observation impact – or whether a given observation platform or observing system improves or degrades the subsequent NWP forecast.
The observation impact is a very specific measure of forecast impact, as it depends upon the choice of forecast metric, the suite of observations assimilated, the data assimilation system, and the NWP forecast model. This chapter presents an overview of data assimilation adjoint theory, the observation impact calculation, and the appropriate choices for the forecast metric. Several applications of the observation adjoint technique are presented to illustrate its usefulness to help identify systematic problems with the observing network, to quantify the value of different observing platforms, to monitor the quality of the observing network, and for channel selection for satellite radiometers.
KeywordsForecast Error Data Assimilation System Background Error Background Error Covariance Sensitivity Vector
Unable to display preview. Download preview PDF.
- Baker NL (2000) Observation adjoint sensitivity and the adaptive observation-targeting problem. Ph.D. dissertation, Naval Postgraduate School, 265 pp. Available from the Naval Research Laboratory, Monterey, CA 93943.Google Scholar
- Daley R (1991) Atmospheric data assimilation, Cambridge University Press, 457 pp.Google Scholar
- Errico R (2007) Interpretation of ad adjoint-derived observational impact measure. Tellus 59A: 273–276.Google Scholar
- Gelb A (1974) Applied optimal estimation. MIT Press, 374 pp.Google Scholar
- Hogan TF, Rosmond TE, Pauley RL (1999) The navy operational global atmospheric prediction system: Recent changes and testing of gravity wave and cumulus parameterizations. Preprints, 13th Conf Numerical Weather Prediction, Denver, CO, Amer Meteorol Soc, pp 60–65.Google Scholar
- Ide K, Courtier P, Ghil M, Lorenc AC (1997) Unified notation for data assimilation: operational, sequential and variational. J Meteorol Soc Japan, 751B: 181–189.Google Scholar
- Joly A, Jorgensen D, Shapiro MA, Thorpe A, Bessemoulin P, Browning KA, Chalon J-P, Clough SA, Emanuel KA, Eymard L, Gall R, Hildebrand PH, Langland RH, Lemaitre Y, Lynch P, Moore JA, Persson POG, Snyder C, Wakimoto R (1997) The Fronts and Atlantic Storm-track Experiment (FASTEX): Scientific objectives and experimental design. Bull Amer Meteorol Soc, 78: 1917–1940.CrossRefGoogle Scholar
- Langland RH, Rohaly GD (1996) Adjoint-based targeting of observations for FASTEX cyclones. Preprints, 7th Conf Mesoscale Processes, Reading, UK, Amer Meteorol Soc, pp 369–371.Google Scholar
- Langland RH, Baker NL (2004a) Estimation of observation impact using the NRL atmospheric variational data assimilation adjoint system. Tellus, 56A: 189–201.Google Scholar
- Langland, RH, Baker NL (2004b) A technical description of the NAVDAS adjoint system. NRL/MR/7530-04-8746. Available from the Naval Research Laboratory, Monterey, CA, 93943, 62 pp.Google Scholar
- Morneau J, Pellerin S, Laroche S, Tanquay M (2006) Estimation of the adjoint sensitivity gradients in observation space using the dual (PSAS) formulation of the Environment Canada operational 4D-Var. Proceedings, 2nd THORPEX Intl Science Symp, Landshut, Germany, 4–8 December 2006, pp 162–163.Google Scholar
- Rodgers, C. D. (1996) Information content and optimization of high spectral resolution measurements. Optical Spectroscopic Techniques and Instrumentation for Atmospheric and Space Research II, SPIE Vol. 2830, 136–147. Published by the International Society for Optical Engineering, PO Box 10, Bellingham, Washington, 98227-0010, USA, 2830.Google Scholar
- Rosmond TE (1997) A technical description of the NRL adjoint modeling system, NRL/MR/7532/97/7230 Available from the Naval Research Laboratory, Monterey, CA 93943, 55 pp.Google Scholar
- Rosmond T, Xu L (2006) Development of NAVDAS-AR: non-linear formulation and outer loop tests. Tellus, 58A, 45–58.Google Scholar
- Ruston B, Blankenship C, Campbell W, Langland R, Baker N (2006) Assimilation of AIRS data at NRL. 15th Intl TOVS Study Conf, Maratea, Italy, 4–10 October 2006.Google Scholar
- Xu L, Rosmond R (2005) Development of NAVDAS-AR: formulation and initial tests of the linear problem. Tellus, 57A, 546–559.Google Scholar