Poor harvesting practices can accelerate soil erosion and decrease water quality and site productivity. Forestry Best Management Practices (BMPs) were developed to protect water quality, primarily by minimizing erosion during and after timber harvests. Although properly employed BMPs mitigate against the immediate potential for soil loss, little information exists regarding their long-term effectiveness. Since 1993, the Virginia Department of Forestry (VDOF) has conducted random water quality audits on forest harvesting operations. The VDOF will recommend remedial BMPs immediately if there is an active water quality law violation, and these recommendations are usually obviously clear to all parties. However, the potential for water degradation is more controversial and debates can arise over these recommendations. The VDOF, as in most states, does not have the resources to make visits to post harvest sites over time. Therefore, it is imperative that the BMPs employed at the closeout of the timber harvest be sufficient to ensure erosion control until the site has recovered, yet, BMP evaluations should represent real or potential problems. This study was undertaken to provide a quantitative analysis of erosion rates over time on VDOF random audited harvest sites and to identify key factors of erosion rates for log decks, skid trails, access roads, harvest areas, and stream crossings within each of Virginia's physiographic regions. A secondary objective was to provide a greater level of decision support for VDOF field staff, through the identified key factors which would indicate potential problem areas of erosion and water quality degradation particularly from logging activities and temporary roads. To better understand the primary causes of erosion over time the Dissmeyer and Foster Universal Soil Loss Equation (USLE) was used to establish quantitative erosion rates which officials can use to focus BMP recommendations. Such a decision support system for field staff, based on quantifiable erosion estimates, provides a proactive targeted prevention assessment prior to the development of actual water pollution problems. Analyses of logged tracts throughout Virginia revealed that estimated erosion rates were not statistically significant between physiographic provinces as well as VDOF audit classifications. Timber access roads were the greatest source of erosion in the Piedmont and Mountains, while harvested areas were the largest identified erosion area in the Coastal Plain, based upon the disturbance categories ratio to the total tract area. Estimated erosion rate trends over time indicated that the majority of disturbance categories were essentially recovered between eight to ten years after harvest. Further, disturbance categories in the Coastal Plain recovered faster than the other province. Erosion rates could not be explained by the year since harvest, since numerous variables interact to cause erosion. Overall predicted erosion rates and VDOF audit classifications of problems did not have consistent agreement. This indicates the need for additional calibration of VDOF ratings and perhaps the establishment of more quantifiable BMP inspection criteria. / Master of Science