Spatial Prediction of Aadt in Unmeasured Locations 3 by Universal Kriging

27 28 ABSTRACT 29 30 This work explores the application of kriging methods for prediction of average daily traffic 31 counts across the Texas network. Results based on Euclidean distances are compared to those 32 using network distances, and both allow for strategic spatial interpolation of count values while 33 controlling for each roadway's functional classification, lane count, speed limit, and other site 34 attributes. Universal kriging is found to reduce errors (in practically and statistically significant 35 ways) over non-spatial regression techniques, though errors remain quite high at some sites, 36 particularly those with low counts and/or in less measurement-dense areas. Interestingly, the 37 estimation of kriging parameters by network distances showed no enhanced performance over 38 Euclidean distances, which require less data and are much more easily computed. 39 40 INTRODUCTION 41 42 Traffic flow volumes represent key information for proper transportation engineering and 43 planning decisions. Sampling, tracking, interpolating, and extrapolating Annual Average Daily 44 Traffic (AADT) counts is fundamental to road construction and maintenance scheduling, as well 45 as to demand modeling and validating estimates of network activity. However, assembly of 46 accurate and robust traffic counts is not straightforward, due to difficulties in measurement and