Diesel particulate matter exposure to railroad train crews

ABSTRACT

Exposure assessments were conducted aboard diesel locomotives. Results were evaluated to determine variables that affect exposure to DPM (diesel particulate matter) and to assess use of EC (elemental carbon) and OC (organic carbon) as surrogates for DPM. National Institute for Occupational Safety and Health Method 5040 was used for collection and analysis of samples in locomotives and in nonrailroad settings.

The level of EC, but not OC, in locomotives was found to be significantly affected by position of exhaust stacks and windows. EC ranged from < 1 to 45 micrograms/m3 with a geometric mean (GM) of 3.7 micrograms/m3 and OC ranged from 4 to 4570 micrograms/m3 with a GM of 36.3 micrograms/m3. Background measurements of EC ranged from < 1 to 8 micrograms/m3 and OC levels were 4 to 84 micrograms/m3. This study confirms that train crew exposure to DPM is much lower than exposures for miners, is comparable to background urban exposures, and is lower than but comparable to exposures for truck drivers. It also indicates that EC levels are highly predictive of diesel exhaust exposure whereas OC levels are not, and that open windows and exhaust stack(s) in front of the locomotive cab have a significant effect on EC.

Sampling and Analytical Method Selection

ACGIH issued a proposed TLV of 150 pg/m* for DPM in 1995, which was revised to a proposed 50 pg/m* in 1999. This was further revised in 2001 to 20 wg/m* measured as EC. The Mine Safety and Health Administration (MSHA) also issued a regulation for DPM of 400 pg/m* based on the measurement of total carbon (TC), which is EC plus OC. The 400 pg/m* conentration limit is to be reduced to 160 pg/m3 in 2006. These MSHA limits are based on DPM being composed of approximately 80% TC.

The National Institute for Occupational Safety and Health (NIOSH) Analytical Method 5040 for Elemental Carbon (Diesel Particulate) requires an open-faced quartz fiber filter (QFF) and
thermo-optical analysis. This method indicates that a size-selective impactor may be required in environments where coal dust may interfere positively with the analysis. There was no coal dust
in the environments assessed in this study and thus, size selection was deemed to be unnecessary.

The collected QFE samples originally were submitted to the only commercial lab performing a thermal-optical method for EC. During the latter part of the study, samples also were submitted to an American Industrial Hygiene Association-accredited laboratory to expedite sample analysis. Both laboratories use the thermal-optical technique specified in NIOSH 5040 analytical method.

Sampling Strategy

EC and OC levels were monitored on two major U.S. railroads in a variety of locations, terrains, weather conditions, types of trains, and models of locomotives. No air-conditioned locomotives were included in this study. The same sampling techniques were used to collect background samples in various nonrailroad locations throughout the United States including a national park, a beach, and major city locations. Many of the city locations were taken in downtown areas near the local courthouse.

During the assessment of occupational exposures, older locomotives, steeper grades, and tunnels were selected for sampling if available. A total at 19 different locomotive models manufactured
between 1968 and 1997 by GE or EMD were included in this study. The terrain was categorized as flat, rolling, or mountainous, and it was noted whether the windows were open or closed. The crews positioned the windows for their own comfort, and the investigator had no input except when riding in an otherwise unoccupied trailing unit when the windows were opened to evaluate
maximum possible exposures. Testing was conducted on both lead and trailing locomotives. Train crews are subject to the Hours of Service Regulation enforced by the Federal Railroad Administration. Their workday is determined by the length of run worked with a maximum of 12 hours prior to 8 hours of rest. During the tests the length of time the crew was on the locomotive ranged from 2 hours 4 min to 10 hours 29 min.

One or more locomotives are used to pull freight trains. Each can be operated facing in either direction and are referred to as “long hood” or “short hood” forward (Figure 1). The short-
hood forward configuration is the more common, where the ex haust stack of the locomotive is approximately 5 m behind the operating cab. In the long-hood forward configuration the exhaust stack is approximately 5 m in front of the operating cab. Therefore, a sample collected in a short-hood forward leading locomotive would have no preceding stacks. A leading locomotive with long-hood forward would have one exhaust stack preceding the sampled cab.

Freight train runs can be categorized as “road” (from point A to point B with little or no intermediate work), "local" (from point A to point B or back to point A with intermediate stops to
pick up or drop off rail cars), or "yard” (moving, rail cars within a rail yard or industry). A typical freight train crew consists of an engineer and a conductor, which can be supplemented with a
brakeman /switchman for local runs or yard jobs. The crew typically occupies the cab of the first (lead) locomotive except when the brakeman /switchman and/or conductor dismount for switching activities, including coupling and uncoupling freight cars. The engineer may be the only cab occupant during switching and coupling activities. Before the use of a caboose was largely discontinued in the mid—1980s, road train crews consisted of up to five people, two of whom occupied the caboose. Currently, there are occasions when a nonworking crew might “deadhead”’ (travel to
another location to begin an assignment or return home) by riding in one of the trailing locomotives. In some mountainous terrain, and with certain types of trains, one or more pusher or helper
locomotives may be added to the middle or rear of the train. These locomotives may or may not be occupied by train crew members.

The samples were collected on the engineer’s operating console in each locomotive. This position is within an arm’s length of the engineer and was selected as being representative of a breathing zone sample, yet not obtrusive to the engineer. A previous study in this type of railroad setting demonstrated that there was no significant difference between fixed location samples and personal
samples. A flow rate of approximately 2 L/min was maintained, and the samples were collected for the entire work period except for time spent away from the locomotives at the beginning and end of the shift. Any samples with flow rate variations exceeding 5% during the sampling period were discarded as were samples taken with tobacco smokers in the cab.

Data Analysis

GMs, based on log transformation of the data, were determined for all samples and for a number of subgroups. Prior studies of EC found the data to be lognormally distributed.For samples reported as below the laboratory limit of detection (LOD), the LOD of the method was used to calculate the EC or OC level. All exposure estimates are expressed as the work period time-weighted average (TWA). The data were analyzed for significant effects of exhaust stack location, railroad, locomotive manufacturer, window position (open or closed), run type, and terrain using Statistical Analysis System (SAS®) software. Analysis was done using the General Linear Model Procedure on the log-trans-
formed data. Differences between means were evaluated with the Duncan’s mean test. GM values were calculated by the following equation:

RESULTS

The EC and OC levels under various conditions are summarized in Table 1. EC and OC concentrations are reported to the nearest whole microgram. "Preceding stacks” refers to exhaust stacks in front of the sample in the direction of travel. Fractional stacks are the result of a change of direction during sample collection. The statistical analysis results are summarized in Table II.

Data analysis shows that exhaust stacks preceding the sample have a highly significant effect on EC (p=0.0001). Window position (open or closed) also has a highly significant effect on EC (p=0.0001). None of the other variables have a significant effect on EC levels. Both the type of run (read, yard, or local) (p= 0.004) and railroad (p=0.01) had a significant effect on OC but not EC. All runs on one railroad were of the road type. Caution is advised in interpreting some of these OC results, as some variables had few data points. No other variables had a statistically significant effect an EC. The study was not well balanced in design in terms of run type, and more data may be necessary to fully
evaluate some of the OC results. The one OC result that is orders of magnitude higher than the others was collected during industrial switching of a wood pulp mill and is believed to be the result
of biogenic contamination.
