NCHRP Report 765 provides alternative procedures to develop turning movement forecasts from a travel demand model. It is important to recognize that most travel demand models do not have sufficient details to produce accurate turning movement volumes directly from the model output. On the one hand, not all roadways connecting to intersections for which turning movement forecasts are desired are coded in the model network. In some cases, some intersections may include an approach that is in fact a centroid connector, either as a replacement for one of the intersections legs or a surrogate for a number of other local roadways; thus, an artificial set of turning movements is introduced at the intersection. When traffic assignment is performed, the turning movements at the node representing the intersection may vary considerably from observed intersection turning movements. On the other hand, turning movement analyses are typically required for an hour, but traffic assignment results are either 24- hour volumes or period volumes depending on the type of models being used. While directional volumes are reported in the model, they are more of a function of trip balancing than of actual directional distribution. The user is advised to start from existing turning movement counts and exercise caution when using the model output to develop turning movement volumes. Nevertheless, there are three (3) categories of procedures for forecasting turning movements from model output, and these procedures can be applied using either daily or period traffic assignment results, directional or non-directional volumes, and with or without model turning movement assignments:

Factoring Procedures

Factoring procedures require base year turning movement counts, base year turning movement assignments, and future year turning movement assignments. Factoring procedures assume that traffic patterns will remain relatively constant between the base year and forecast year. Based on this assumption, future year turning movements can be estimated by comparing either the relative ratios or relative differences between base year and future year turning movement assignments and then applying the same ratios or differences to base year turning movement counts. However, when travel patterns are expected to change significantly (for example, a major new development near one of the intersection approaches), other procedures may be more appropriate.

Iterative Procedures

Iterative procedures employ the traditional Iterative Proportional Fitting (IPF) or Fratar method, which has been widely used in practice to balance trip tables. The iterative method is based on an incremental procedure of applying implied growth between base year and future year to actual traffic counts. Growth rates are derived from the model. The iterative procedures would require observed turning movements for all intersections under study. This method is not applicable to new intersections for which base year counts are not available. The Fratar method would produce reasonable results for either developed areas or areas expected to experience moderate growth in land use.

Iterative procedures differ depending on whether directional or non-directional volumes are used for the approach links. The directional volume procedure adjusts future year turning movements based on either base year turning movement counts or future year turning movement estimates and the ratio of approach link forecasts to link counts. The non-directional volume procedure is more subjective and requires the analyst to produce a reasonable estimate of turning percentages as an input to the process. The procedure should be used mainly for planning and preliminary engineering applications, not for design.

“T” Intersection Procedures

“T” intersection procedures are used for intersections with only three approach legs. Directional turning volumes can be computed if the approach volumes and at least one turning movement are known. Where only two-way turning movements are available, a unique solution can be found if directional approach volumes are known. Table 7-1 presents a summary on procedures and input elements for turning movement volume forecasting.

Users are advised to consult the NCHRP Report 765, Chapter 6 for detailed discussions on these procedures. Users must always exercise professional judgment during and after the applications of these procedures. Reasonableness checks, involving stakeholders, and applying local knowledge are recommended to properly develop turning movement forecasts.