Aeolian features observed on the surface of Mars provide insight into current, and potentially past, surface wind systems. In some cases the features are clearly transient and related to the lifting and settling of atmospheric dust. Other features, like dunefields, yardangs, and ventifacts, are more persistent and likely require significant time to form. In this study we analyze the observed directions of selected aeolian features with the aid of the Geophysical Fluid Dynamics Laboratory Mars general circulation model (GCM). Initially, we examine bright and dark streaks which have been observed to form in association with global dust storms. The ability to match these features with Mars GCM wind directions provides an important validation of the model. More important, we are able to define best fit seasons and local times for both types of features which provide the basis for extension and modification of the Veverka et al. <1981> model of bright and dark dust streak formation. In addition these best fit times correspond well with the dark streak wind storm model of Magalh¿es and Young <1995>. The primary focus of this paper is to provide constraints on the range of mechanisms proposed to explain inconsistencies between current wind direction patterns and long-term wind indicators (for example, the misalignment of rock tail and ventifact orientations at the Mars Pathfinder landing site). Specifically, we assess whether changes in planetary obliquity, precession, or global dust opacity could significantly alter patterns of surface wind directions. In all cases we find the seasonal and annual average wind direction patterns to be highly invariable. While changes in the dust loading (hence the partitioning of solar absorption between the surface and atmosphere) and in the surface latitude of maximum solstitial insolation cause the vigor of the large-scale circulation to increase (especially the Hadley cell), the spatial patterns of the surface wind orientations remain essentially unchanged. In the case of perihelion during northern summer (opposite of the current perihelion position), the northern summer Hadley cell remains weaker than the southern summer cell despite the strengthened heating in the northern hemisphere. Taken together, these results cast significant doubt on orbital explanations for surface wind changes. It is thus suggested that significant changes in topography (e.g., Tharsis uplift, true polar wander) or climate (e.g., the existence of a significantly thicker atmosphere or an ocean at some point in the past) are more likely explanations for long-term wind indicators such as the ventifact orientations at the Mars Pathfinder landing site. ¿ 2001 American Geophysical Union