Abstract:The Environmental Protection Agency (EPA) has indicated that those attainment area intersections "[operating] at LOS D, E, F or those that have changed to LOS D, E, or F because of increased volumes of traffic or construction related to a new project in the vicinity" will require detailed modeling of Carbon Monoxide (CO) impacts. The purpose of this study was to determine which, if any, intersections operating at LOS D will result in a modeled CO exceedance when controlling for worst-case meteorological conditions, intersection orientation, and intersection geometry and traffic volumes.

To conduct the technical analysis, a hypothetical four-leg, five-phase, pre-timed, signalized intersection was constructed with volumes and phasing designed to ensure LOS D operations. The intersection is representative, in terms of size and composition, to many intersections found in the urban areas of California with construction, management, and design oversight provided by Caltrans. The hypothetical intersection included an east-west main-line approach with three-through lanes and separated right- and left-turn lanes on the eastbound approach. The westbound main-line included three through -lanes, with one lane allotted to combined through- and right-turn volumes, and a separated left-turn lane. The side-street approaches comprised each of two through-lanes, one lane also used by right-turns, and a separated left-turn lane.

The analysis tasks undertaken to estimate modeled CO concentrations for intersections operating at LOS D included:

• 1. Estimating the highest (exceedance) and second highest (violation) modeled 8-hour CO concentrations by intersection orientation, a total of 288 runs of EPA's recommended model CAL3QHCR, for each of four worst-case meteorological conditions in California: Sacramento (1989), Redlands (1981), West Los Angeles (1981), and San Jose (1988);
• 2. Using the worst-case identified in Task 1, i.e., by location and orientation, analyzing the effects of increasing and decreasing traffic volumes and merging right- and left-turn lanes with through movements on modeled CO concentrations.
• 3. Using the worst-case identified in Task 1, i.e., by location and orientation, analyzing the effects of changing fleet characteristics on modeled CO characteristics.

There are several important aspects of these analyses that should be highlighted. First, the eastbound main-line approach was designed to operate with both the highest queues and the longest delays and the signal timing, phasing, and traffic volumes were adjusted to maintain a LOS D in each analysis scenario; this was done to ensure consistency with EPA's "hotspot" designation policy. In addition, the user-defined modeling parameters that were required as input to CAL3QHCR were always selected to reflect a more conservative modeling approach. Specifically, the analysis specified a pre-timed signal, likely to produce more conservative (higher) estimates of queues than the more commonly found actuated signals, receptors located at 3-m, and 1997 emission factors (i.e., an assumed vehicle fleet operating in the 1997 calendar year) with slightly higher cold start percentages (30%) than that used FTP (26%).

The analysis results clearly demonstrate that, based on CAL3QHCR predicted CO concentrations, using worst-case meteorological conditions, employing a plethora of additional conservative inputs (e.g., pre-timed vs. actuated signalization), and assuming a 3.0 ppm background, most intersections operating at LOS D will not be likely to violate the 8-hour 9.0 ppm CO standard.