This thesis is an empirical analysis of methods to enhance estimates of annual average daily traffic (AADT) on Manitoba's provincial highways. Three methods of enhancing AADT estimates are presented: (1) human-intervention techniques to evaluate the estimates, (2) developing new traffic pattern groups (TPGs) to improve the representation of provincial highway traffic patterns, and (3) developing a "transparent" methodology for assigning short-count sites to appropriate TPGs and control stations. This thesis develops human-intervention techniques for evaluating AADT estimates on Manitoba's provincial highways, including visual route consistency checks, comparison of current year estimates with historical estimates, and checking for intersection balancing. This research also investigates whether improved TPGs can enhance the quality of AADT estimates on Manitoba's provincial highways. This research develops new TPGs for an expanded study region consisting of Manitoba and Saskatchewan, based on the premise that traffic patterns are not affected by the political boundary that separates two provinces, rather these patterns are characterized by the transportation, activity and flow systems of a region. The new TPGs are found to improve the quality of AADT estimates for the majority of Manitoba's short-count sites. The TPG and control station which are assigned to a short-count site depend on the method used for assignment. An inappropriate assignment method leads to inconsistent and non-reproducible assignments, which in turn affects the quality of AADTs estimated from short-count sites. This research develops a "transparent" methodology to assign Manitoba's short-count sites to TPGs and control stations. The new method involves a documented rule-based procedure to assign short-count sites to appropriate TPGs and control stations. (Abs ract shortened by UMI.)