The potential of high density observations on Numerical Weather Prediction : A study with simulated observations
L'influence d'observations à haute densité sur la prévision numérique du temps : Une étude réalisée avec observations simulées
Liu, Z. ; Rabier, F.
Année de publication
2003
The skill of numerical weather prediction depends to a large extent upon<br> the quantity of globally available observations. Only a fraction of the<br> available observations (especially high-density observations) is used <br>in current operational assimilation systems. In this paper, the <br>potential of high-density observations is studied in a practical <br>four-dimensional variational assimilation context. Two individual <br>meteorological situations are used to examine the impact of different <br>observation densities on the analysis and the forecast. A series of <br>observing-system simulation experiments are performed. Both direct <br>observations (temperature and surface pressure) and indirect <br>observations (radiance) are simulated, with uncorrelated or correlated <br>errors. In general, it is verified that a small reduction (increase) of <br>the initial error in a sensitive area can produce a considerable <br>improvement (degradation) of the targeted forecast. In particular, the <br>results show that increasing the observation density for the <br>uncorrelated-error case can generally improve the analysis and the <br>forecast. However, for correlated observation errors and the use of a <br>diagonal observation-error covariance matrix in the assimilation, an <br>increase in the observation number such that the error correlation <br>between two adjacent observations becomes greater than a threshold value<br> (around 0.2) degrades the analysis and the forecast. Posterior <br>diagnostics of the sub-optimality of the assimilation scheme for <br>correlated observation errors are analysed. Finally, it is shown that a <br>risk of using high-density observations and poor vertical resolution is <br>that deficiencies in the background-error statistics can lead to <br>unrealistic analysis increments at some levels where no observations are<br> present, and so produce a degradation of the analysis at these levels. <br><br>Copyright © 2003 Royal Meteorological Society</div>
Accès à la notice sur le site du portail documentaire de Météo-France