Search Google Appliance


Select Duration

If the counts are being collected manually in the field, the length of time a human can reasonably be expected to accurately count is generally limited to about two hours. This greatly reduces the possible duration of manual counts. Unfortunately, estimates of Annual Average Daily Bicyclists/Pedestrians (AADBP) based on just two hours of counts have been shown to have as much as 60 percent error on average.  Fortunately, such estimates can be much more accurate if based on counts taken during peak hours in mid-September (assuming a Midwestern climate) as recommended in by the National Bicycle and Pedestrian Documentation Project.

However, if video is collected and watched later, this can increase the time a human can collect data manually. Even better, if automated counters are used, counts of 24 hours or longer are not unreasonable.  Multiple studies show that collecting at least one week of counts per location results in much increased accuracy. 123 Below is a graph illustrating how AADB estimation error decreases with increasing duration of short counts. For this reason, we recommend one week of counts, if possible.



Resources

Hankey and others provide useful guidance in their 2014 presentation at the Transportation Research Board Annual Meeting.

Nosal and others report similar findings in their 2014 presentation at the Transportation Research Board Annual Meeting.

Nordback and others recommend one week of short duration counts as optimal.

 


GUIDANCE FROM THE TRAFFIC MONITORING GUIDE 2013, SECTION 4.5.3

1.1.1 DURATION OF COUNTS

There is no definitive guidance on the minimum required duration of short-duration counts. The prevailing practice has been two consecutive hours on a single day, but that practice is evolving as more public agencies use automatic counters and become aware of the inherent variability of non-motorized traffic. The following paragraphs discuss several factors that agencies should consider when determining the duration of their short-duration counts.

Manual Versus Automated Collection

The use of automatic counter equipment can dramatically extend the duration of short-duration counts. If automatic counters are used, then the minimum suggested duration is 7 days (such that all weekday and weekend days are represented). Depending on several other factors (e.g., day-to-day count variability, the total number of short-duration monitoring sites, and the number of automatic counters), the preferred duration of automatic counts could be as long as 14 days at each location.

The use of manual observers will limit the duration of short-duration counts. However, the minimum suggested duration for manual observers is 4 to 6 hours and should be scheduled to coincide with the heaviest non-motorized use (typically mid-day for weekend/recreational trips and morning/evening commute times for other trips). Manual observers’ counting accuracy declines after 2 hours, so observers should be given short breaks or replaced with other observers. The preferred length for short-duration counts is 12 hours, which permits calculation of time-of-day use profiles. However, it is recognized that available resources may limit the collection of 12-hour counts.

The prevailing practice for short-duration manual counts has been 2 hours, largely because of resource and manual observer limitations. There is recognition that 2 hours of count data is better than no data; however, 2 hours of count data may lead to high error rates when annualizing counts and could lead to erroneous conclusions. If manual observers are the only possibility for short-duration counts, then agencies are encouraged to count for longer periods at fewer locations. Alternatively, the NBPD project (National Bicycle and Pedestrian Documentation Project: Instructions) has encouraged agencies to count multi-hour periods on several different days:

“We suggest that between 1 and 3 counts be conducted at every location on sequential days and weeks, based on the approximate levels of activity. Areas with high volumes (over 100 people per hour during mid‐day periods) can usually be counted once on a weekday and weekend day, unless there is some unusual activity that day or land use nearby.”

“Areas with lower activity levels and/or with unusual nearby land uses (with any irregular activity, such as a ball park) or activity (such as a special event) should be counted on sequential days or weeks at least one more and possibly two more times.”

Count Magnitude and Variability

If non-motorized traffic levels are high and consistent from day-to-day, then shorter periods and/or fewer days may be considered. However, a longer-duration count period will be needed to determine how variable the non-motorized traffic is by time-of-day and DOW. Unfortunately, there is little quantitative guidance or consensus in this area, and ongoing research will improve future guidance.

Weather

Weather can be a significant factor in the level and variability of non-motorized traffic and should be considered when developing a short-duration monitoring program. Seasonal weather patterns (such as cold winters or hot/humid summers) are expected by pedestrians and bicyclists and will result in relatively consistent patterns from year to year. However, heavy precipitation or unexpectedly hot or cold weather may introduce abnormal variations on a given time of day or day of year. These variations can both generate unusually high levels of activity (e.g., a very nice day) or depress otherwise expected levels of activity (due to very bad weather.)

If automatic counter equipment is used for short-duration counts in typical weather, then the minimum suggested duration is 7 days (such that all weekday and weekend days are represented). This duration provides an average of 5 weekdays and 2 weekend days. However, if atypical heavy precipitation or inclement weather occurs during this entire 7-day period, agencies should consider extending the duration to 14 days.

When heavy precipitation or inclement weather occurs with manual observers, the counts should be extended over multiple days at the same time. Local judgment should be used to determine whether to include inclement-weather days into a multi-day average.

Because of inclement weather’s influence on non-motorized traffic, weather conditions should be recorded in a non-motorized traffic monitoring program. The non-motorized data submittal format in Chapter 7 recommends three weather-related attributes:

3. Precipitation (yes/no): Did measurable precipitation fall at some time during data collection?

4. High temperature: Approximate high temperature for either the day (if a day or longer count) or the duration of the count (if the count is less than a day in duration).

5. Low temperature: Approximate low temperature for either the day (if a day or longer count) or the duration of the count (if the count is less than a day in duration).

Historical weather data can be obtained from several different sources and does not necessarily have to be collected at the exact count location.

 


1Nordback, K., W. Marshall, et al. (2013). Estimating Annual Average Daily Bicyclists: Error and Accuracy. 92nd Annual Meeting of the Transportation Research Board. Washington, D.C., National Academies.
2Hankey, S., G. Lindsey, et al. (2014). Day-of-Year Scaling Factors and Design Considerations for Non-motorized Traffic Monitoring Programs. 93rd Annual Meeting of the Transportation Research Board. Washington, D.C., National Academies.
3Nosal, T., L. Miranda-Moreno, et al. (2014). Incorporating weather: a comparative analysis of Average Annual Daily Bicyclist estimation methods. 93rd Annual Meeting of the Transportation Research Board. Washington, D.C., National Academies.