| description abstract | For a propagating mesoscale system whose intensity and structure is not changing with time, relatively coarse horizontal profiler resolution is sufficient to resolve the feature since the circulation would pass by the profiler sites quickly enough to construct a three-dimensional analysis. This is generally not true for a thermally forced mesoscale system. For mesoscale systems generated by surface inhomogeneities in surface heating (e.g., land-sea contrasts, nonuniform soil wetness, etc.), such propagation is often slow. Therefore, ideally, if thermally surface-forced systems are to be directly resolved by a profiler network, a necessary condition is that their spacing be close enough to adequately resolve the motion field of the mesoscale system. As concluded from the analyses in this paper, higher spatial resolution is required to directly monitor the horizontal wind field than the temperature field, since the horizontal wind is proportional to the horizontal gradient of temperature. Similarly, even higher resolution of vertical velocity is required since ascent and descent are proportional to the horizontal gradient of the horizontal velocity. The use of mesoscale numerical models as analysis tools, however, offers the opportunity to obtain fine-scale horizontal resolution with only relatively coarse atmospheric data. Such fine scale resolution is obtained because the surface thermal forcing can be resolved with high spatial accuracy and, through nonlinear advection and the pressure gradient force in the numerical model, fine-scale atmospheric structure can be produced. Finally, stringent data initialization requirements would result if one attempted to insert mesoscale resolution profiler-derived temperature or wind data into a model. Even if 10-km horizontal resolution were obtained with a profiler network and if relative errors in the temperature measurements were only 0.24°C through a depth of 2 km or so, a fictitious 1 m s?1 h?1 acceleration would result. For the same resolution, for winds from one profiler of 0, 5, and 10 m s?1, an error from the adjacent profiler of 2.4, 0.5, and 0.3 m s?1, respectively, would result in the same erroneous acceleration. | |
