Alternative Methods for Estimating Seasonal Factors and Accuracy of Daily Volumes They Yield

Abstract

The Federal Highway Administration requires each state department of transportation to have a traffic monitoring program for estimation of among others annual average daily traffic (AADT), a performance measure that guides fund-allocation to the states. Determination of AADT for each road section theoretically requires year-round counts to be made at all sections of a state’s monitored road network. Because this is not cost-effective, counts are made for a short duration on the majority of road sections, and these are adjusted to AADT estimates using seasonal factors (SFs) developed from data collected at the small number of road sections, called automatic traffic recorder (ATR) stations, at which year-round counts are made. Alternative methods exist for developing the SFs. One method makes use of only the most recent calendar years’ ATR data. A second makes use of multiple calendar years of ATR data, determining the final SFs as a simple mean of SFs from each of the calendar years. A third method, developed in this research, called the weighted-factor method, also makes use of multiple calendar years of ATR data but combines SFs from these calendar years taking into consideration their respective variances. The three methods were empirically investigated for the accuracy of the AADT estimates they yield, and the weighted-factor method found to yield the most accurate AADT-estimates.