Some limits to the variable mesh solution and comparison with the nested LAM one
Caian, M. ; Geleyn, J.-F.
The solution at high resolution of a global <br>spectral model with a conformally-variable mesh, obtained through a <br>stretched coordinate, is compared as fairly as possible to the <br>classical' solution of a limited-area nested model. We compare the <br>global, variable-mesh, ARPEGE model at hyper-stretched values (which is <br>integrated operationally at Météo-France with a low' value of <br>stretching) with its limited-area version, the ALADIN model. Having the <br>same physics, the same grid-point dynamics and a high <br>code-compatibility, they allow clean comparisons on real situations. to <br>conclude on their efficiency and accuracy, comparisons are made choosing<br> configurations of the two models having either the same resolution or <br>the same computational price. The limits <br>of the variable-mesh strategy on the sphere are also investigated <br>through a set of experiments with increasing stretching. We found that <br>there is indeed a limit to the stretching factor in the interval [7, 9]:<br> a stretching factor <em>c</em> = 7 keeps an accurate spectral solution <br>and gives satisfying results, even when compared with the limited-area <br>model at the same computational price. We show that the value <em>c</em><br> = 10.5 cannot be reached without paying the price of a spectral <br>deformation of the fields which becomes too obvious. Additional tests <br>showed the occurrence of this deformation, even if less pronounced, at <br>the stretching value <em>c</em> = 8.75. At the same resolution ALADIN does not manifest any similar weakness. Although<br> the solution remains stable for at least 72 hours, its accuracy when <br>increasing the stretching above the critical value becomes questionable.<br> This problem seems to be due to a particular behaviour of the <br>truncation error, which does not decrease with increasing resolution <br>through stretching, thus leading to an energy accumulation in the tail <br>of the spectra. As a conclusion of this <br>study, it follows that one can extend the idea of Courtier and Geleyn, <br>currently used in operations at Météo-France for <em>c</em> = 3.5, to medium' stretched configurations up to <em>c</em> = 7 at least (at the current computer's power). Above a value found to be about <em>c</em><br> = 9, our results show that increasing the resolution does not produce <br>any improvement and, hence, that the use of a limited-area model appears<br> a better choice for very high resolutions.</p>
Accès à la notice sur le site du portail documentaire de Météo-France