- Email: [email protected]
Weather factor impacts on commuting to work by bicycle
Preventive Medicine 54 (2012) 122–124
Contents lists available at SciVerse ScienceDirect
Preventive Medicine journal homepage: www.elsevier.com/locate/ypmed
Brief Original Report
Weather factor impacts on commuting to work by bicycle Brian S. Flynn a,⁎, Greg S. Dana a, Justine Sears b, Lisa Aultman-Hall b a b
Office of Health Promotion Research, College of Medicine, University of Vermont, Given Courtyard Building N302, 89 Beaumont Avenue, Burlington, VT 05405, USA Transportation Research Center, University of Vermont, Farrell Hall, 210 Colchester Avenue, Burlington, VT 05405, USA
a r t i c l e
i n f o
Available online 2 December 2011 Keywords: Bicycling Transportation Weather Adults Exercise Physical fitness
a b s t r a c t Objective. Quantify the impact of weather conditions on individual decisions to commute to work by bicycle among a diverse panel of adults who commute ≥ 2 miles each way. Method. Working adults (n = 163) in a northern U.S. state reported transportation mode for four sevenday periods in 2009–2010 that maximized seasonal weather variations. Personal characteristics, trip to work distances, and commuting mode data were linked to location- and time-specific weather data and daylight hours. Analyses focused on effect of weather conditions on reports of commuting by bicycle. Results. Participants were diverse in age, gender and bicycle use, but were relatively well-educated; they traveled to work by bicycle on 34.5% of the logged commuting days. Modeling indicated that the likelihood of bicycle commuting increased in the absence of rain (odds ratio = 1.91; 95% confidence interval 1.42, 2.57) and with higher temperatures (1.03; 1.02, 1.04), and decreased with snow (0.90; 0.84, 0.98) and wind (0.95; 0.92, 0.97). Independent effects also were found for bicycle commuting distance, gender, and age, but not for daylight hours. Conclusion. Precipitation, temperature, wind and snow conditions had significant and substantial independent effects on the odds of travel to work by bicycle among a diverse panel of adult bicycle commuters. © 2011 Elsevier Inc. All rights reserved.
Introduction Concerns about health consequences of inadequate physical activity have increased interest in active transportation (American Public Health Association, 2009; Dora and Phillips, 2000). Accumulating evidence suggests potential for health benefits from greater participation in bicycle commuting, including achieving recommended physical activity levels and improving cardiovascular fitness and risk factors (Buehler et al., 2011; Oja, et al., 2011; Shephard, 2008). Utilitarian bicycle use is low in the U.S., but it is higher in countries with similar levels of development and weather conditions, suggesting the potential for increased participation (Pucher, et al., 2010). Evidence about influences on bicycle use for utilitarian travel can contribute to improved methods to support wider use of bicycling. Bicycling infrastructure is an important modifiable facilitator of participation (Dill and Carr, 2003), but understanding the impact of other factors also is important. Information about the influence of weather conditions on individual bicycle use behavior is lacking (Heinen et al., 2010; Saneinejad, et al., 2010). Relationships between aggregate bicycle use and aggregate weather data suggest significant associations for temperature and ⁎ Corresponding author. Fax: +1 802 656 8826. E-mail addresses: [email protected] (B.S. Flynn), [email protected] (G.S. Dana), [email protected] (J. Sears), [email protected] (L. Aultman-Hall). 0091-7435/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ypmed.2011.11.002
precipitation (Brandenburg et al., 2007; Buckley, 1982; Dill and Carr, 2003; Emmerson and Ryley, 1998; Nankervis, 1999; Nelson and Allen, 1997; Niemeier, 1996; Parkin et al., 2008; Pucher and Buehler, 2006; Winters, et al., 2007). At the individual level, Hanson and Hanson (1977) analyzed travel mode and weather diaries kept for 39 days to demonstrate correlations between bicycling to work and morning temperatures and cloud cover. The present study was planned to quantify the impact of specific weather conditions on individual use of bicycles for travel to work. Methods This longitudinal study documented reports of travel to work by bicycle among a convenience sample of adults on 28 pre-specified days over 10 months, and linked these reports to objective weather data. The study was conducted in five counties with populations ranging from 44,513 to 156,545 in the rural northeastern state of Vermont, U.S.A. Each study site included a main population center in a valley and smaller towns in rolling hills. We planned to recruit 100 participants using email lists from community organizations and workplaces, with sub-targets of one-third women, one-half over age 40, and all-season bicyclists comprising ≤one-quarter. Other criteria were: age ≥ 18, bicycle commuting frequency of > 2 annually, and commuting to work ≥ 2 miles. Full-time students were excluded. The study was approved by the University of Vermont Institutional Review Board. Weather data for the pre-specified days were obtained from four regional National Weather Service (NWS) stations based on morning commuting hours (5–9 a.m.) and proximity to participant residences: mean temperature
B.S. Flynn et al. / Preventive Medicine 54 (2012) 122–124
(Fahrenheit); mean wind velocity (miles/h); and precipitation (rainfall present/absent). Snow depth (inches) was measured at 7 a.m. by 18 NWS cooperating stations. Daylight hours were calculated for each community latitude and commuting day (www.usno.navy.mil/USNO). Age, gender, and reported distance typically traveled one-way to work by bicycle (miles) were obtained from baseline surveys. The outcome was a report of bicycle commuting or not on days that required a trip to the workplace, recorded by participants on web-based log forms. Baseline surveys were administered during May–July 2009. Log forms were completed during seven-day periods in each of four months: September 2009 (mean temperature 62°F), January 2010 (18°F), April (43°F), and July (71°F). Daily log records were linked to baseline data by unique identifiers, and to weather and daylight hours by location and date codes. The unit of analysis was person-day records for commuting days. We used a generalized linear model to estimate the independent impact of each weather factor on bicycle use for commuting days, while simultaneously controlling for other independent variables. Our dependent variable was ‘BIKED’ (yes or no) with a binary distribution assigned to the model. To account for correlation among observations from participants on multiple days, a “repeated” statement was included in the model.
Results Of 210 individuals who responded to recruitment activities, 185 met inclusion criteria. Sufficient data for modeling were obtained from 163, including 102 men and 61 women. Participants were more likely to be ≥40 years than younger (Table 1). College degrees were reported by 93% (compared to 33% for Vermont adults); 90% reported very good or better health (versus 61% in Vermont). Notably
Table 1 Participant characteristics for analytic sample by bicycle commuting distance each way (n = 163), Vermont, 2009–2010. n
Bicycle commute to work 2–5 miles (n = 72) %
>5 miles (n = 91) %
All (n = 163) %
Age 40 or under 40 +
66 97
50.0 50.0
33.7 66.3
40.5 59.5
Gender Male Female
99 64
58.3 41.7
65.3 34.7
60.7 39.3
Education b 2 yr. degree 2 yr. degree 4 yr. degree > 4 yr. degree
7 5 67 84
7.0 1.4 48.6 43.1
2.0 4.1 34.7 59.2
4.3 3.1 41.1 51.5
Health Excellent Very good Good Fair or poor
79 70 13 1
41.7 50.0 6.9 1.4
53.1 50.0 9.2 0
48.5 42.9 8.0 0.6
Children in household Yes 61 No 102
26.4 73.6
45.9 54.1
37.4 62.6
40.8 59.2
34.4 65.6
≥ 1 bike ride in the previous year for… …training or competition? Yes 56 23.6 No 107 76.4 …recreation or leisure? Yes 161 97.2 No 2 2.8 …other utilitarian purpose? Yes 140 86.1 No 23 13.9
100 0 85.7 14.3
98.8 1.2 85.9 14.1
123
low bicycle use for training or competition was reported, indicating that few were dedicated competitors. Baseline surveys also indicated that nearly 20% rode bicycles in winter months and that biking comprised about 35% of all trips to work (median 96 days annually; range: 5–288). The total number of daily log reports for days requiring a trip to work was 2554 person-days. In aggregate, participants reported biking to work on 881 (34.5%) of these commuting days. Weather conditions for logged days matched expectations for normal seasonal variation (Table 2). Temperature was correlated with snow depth (r = −0.72) and daylight hours (0.80). Precipitation and wind were not correlated with other weather factors or daylight hours. The dependent variable for bicycle commuting was regressed on independent variables temperature, wind, precipitation, snow, daylight, distance, age and gender. Nearly all factors had a significant independent relationship with bicycle commuting. Odds ratios showed participants were nearly twice as likely to commute by bicycle when there was no morning precipitation (Table 2). A similar effect was found for temperature, where a one degree increase raised the likelihood of biking by about 3%. A one mile per hour increase in wind speed decreased biking likelihood by about 5%. One inch of snow on the ground reduced the likelihood of biking by about 10%. Discussion This study makes a unique contribution to quantification of weather influences on bicycle commuting. A bicycle commuter panel was engaged over an extended time with good variation in participant characteristics and weather conditions. Results provided evidence for independent weather factor effects on individual decisions to travel to work by bicycle. Precipitation and temperature appeared to be strong influences on the odds of commuting to work by bicycle, consistent with other research. The odds of bicycle commuting nearly doubled when no precipitation was recorded for the morning commuting hours. Bicycle commuting decisions similarly appeared to be sensitive to temperature. Contrary to expectation from preliminary research, increased wind speed diminished the odds of bicycle commuting modestly. Snow depth had a dampening effect that might be expected when most participants did not ride bicycles in winter months. The study was conducted in a region with adverse winter conditions, but most participants commuted by bicycle throughout a significant part of the year, and some even in the coldest months. Improvements in infrastructure and maintenance, better equipment, and greater acceptance of cold weather bicycling could extend the range of conditions in which bicycle commuting becomes feasible for others (Bergstrom and Magnusson, 2003; Pucher and Buehler, 2006). A limitation of the current study was that the geographic areas covered by weather station data lacked local detail that might influence commuting decisions, potentially weakening the relationships studied. These data focused on morning commuting hours and did not account for participants with other work schedules. The snow depth measure reflected general winter conditions rather than daily road conditions. These community-recruited participants were diverse in terms of gender, age, and bicycle use, but were relatively well-educated compared to the general population, perhaps due to recruitment methods and inclusion criteria. Precipitation, temperature, wind speed, and snow depth had independent effects on the odds of commuting by bicycle to work among a diverse panel of adults who bike to work at least 2 miles. These results may support development of methods to mitigate adverse effects of weather on bicycle commuting and encourage greater use of active commuting. Increased use of active transportation for routine purposes can increase opportunities for physical activity and achievement of related health benefits.
124
B.S. Flynn et al. / Preventive Medicine 54 (2012) 122–124
Table 2 Odds ratios for regression of likelihood of bicycle commuting on hypothesized factors (n = 163), Vermont, 2009–2010. Effect variables were treated as independent factors in a generalized linear model. Effect
Odds ratio point estimate
95% confidence interval
Temperature (°F) Wind speed (mph) Precipitation (no vs. yes) Snow (inches) Daylight (hours) Distance (miles) Age (years) Gender (men vs. women)
1.03 0.95 1.91 0.90 1.00 0.92 1.02 2.65
1.02 0.92 1.42 0.84 0.99 0.89 1.01 1.78
1.04 0.97 2.57 0.98 1.00 0.96 1.04 3.99
Range
Mean
− 3.2–79.2 0.0–20.0 0.0–0.4 0.0–23.0 9.0–15.4 2.0–30.0 22–66 –
45.3 5.1 b0.1 2.6 12.4 7.7 43.8 –
Note: Temperature and wind were represented by mean values, and precipitation as rainfall present or absent through 9 a.m. on logged days at the National Weather Service (NWS) regional station nearest to a participant's residence. Snow depth was measured at 7 a.m. on these days at the nearest NWS cooperating station.
Conflict of interest statement The authors declare that there are no conflicts of interest.
Funding sources and role in study This work was funded by the United States Department of Transportation through the University Transportation Center Program at the University of Vermont and by the College of Medicine at the University of Vermont. Co-investigators from both organizations at the University of Vermont designed, conducted, and reported the study. Acknowledgments The contributions of Anne L. Dorwaldt and many community partners to recruitment of participants are gratefully acknowledged. References American Public Health Association, 2009. At the Intersection of Public Health and Transportation: Promoting Healthy Transportation Policy. Washington D.C.. Bergstrom, A., Magnusson, R., 2003. Potential of transferring car trips to bicycle during winter. Transp. Res. Part A 37, 649–666. Brandenburg, C., Matzarakis, A., Arnberger, A., 2007. Weather and cycling — a first approach to the effects of weather conditions on cycling. Meteorol. Appl. 14, 61–67. Buckley, C.A., 1982. Bicycle traffic volumes. Transp. Res. Rec. 847, 93–102. Buehler, R., Pucher, J., Merom, D., Bauman, A., 2011. Active travel in Germany and the U.S. — contributions of daily walking and cycling to physical activity. Am. J. Prev. Med. 41, 241–250.
Dill, J., Carr, T., 2003. Bicycle commuting and facilities in major U.S. cities: if you build them commuters will use them. Transp. Res. Rec. 1828, 116–123. Transportation, environment, and health. In: Dora, C., Phillips, M. (Eds.), Copenhagen: World Health Organization, Regional Office for Europe, WHO Regional Publications No. 89. Emmerson, P., Ryley, T.J., 1998. The impact of weather on cycle flows. Traffic Eng. Control 238–243 (April). Hanson, S., Hanson, P., 1977. Evaluating the impact of weather on bicycle use. Transp. Res. Rec. 629, 43–48. Heinen, E., Van Wee, B., Maat, K., 2010. Commuting by bicycle: an overview of the literature. Transp. Rev. 30, 59–96. Nankervis, M., 1999. The effect of weather and climate on bicycle commuting. Transp. Res. Part A 33, 417–431. Nelson, A.C., Allen, D., 1997. If you build them, commuters will use them: association between bicycle facilities and bicycle commuting. Transp. Res. Rec. 1578, 79–83. Niemeier, D.A., 1996. Longitudinal analysis of bicycle count variability: results and modeling implications. J. Transp. Eng. 200–206 (May/June). Oja, P., Titze, S., Bauman, A., et al., 2011. Health benefits of cycling: a systematic review. Scand. J. Med. Sci. Sports 21, 496–509. Parkin, J., Wardman, M., Page, M., 2008. Estimation of the determinants of bicycle mode share for the journey to work using census data. Transportation 35, 93–109. Pucher, J., Buehler, R., 2006. Why Canadians cycle more than Americans: a comparative analysis of bicycling trends and policies. Transp. Policy 13, 265–279. Pucher, J., Dill, J., Handy, S., 2010. Infrastructure, programs, and policies increase bicycling: an international review. Prev. Med. 50, S106–S125. Saneinejad, S., Kennedy, C., Roorda, M.J., 2010. Assessing the impact of weather and climate on commuter trip behaviour in Toronto. Paper presented at XVI PanAmerican Conference of Traffic Control and Transportation Engineering, Lisbon. Shephard, R.J., 2008. Is active commuting the answer to population health? Sports Med. 38, 751–758. Winters, M., Friesen, M.C., Koehoorn, M., Teschke, K., 2007. Utilitarian bicycling a multilevel analysis of climate and personal influences. Am. J. Prev. Med. 32, 52–58.
Contents lists available at SciVerse ScienceDirect
Preventive Medicine journal homepage: www.elsevier.com/locate/ypmed
Brief Original Report
Weather factor impacts on commuting to work by bicycle Brian S. Flynn a,⁎, Greg S. Dana a, Justine Sears b, Lisa Aultman-Hall b a b
Office of Health Promotion Research, College of Medicine, University of Vermont, Given Courtyard Building N302, 89 Beaumont Avenue, Burlington, VT 05405, USA Transportation Research Center, University of Vermont, Farrell Hall, 210 Colchester Avenue, Burlington, VT 05405, USA
a r t i c l e
i n f o
Available online 2 December 2011 Keywords: Bicycling Transportation Weather Adults Exercise Physical fitness
a b s t r a c t Objective. Quantify the impact of weather conditions on individual decisions to commute to work by bicycle among a diverse panel of adults who commute ≥ 2 miles each way. Method. Working adults (n = 163) in a northern U.S. state reported transportation mode for four sevenday periods in 2009–2010 that maximized seasonal weather variations. Personal characteristics, trip to work distances, and commuting mode data were linked to location- and time-specific weather data and daylight hours. Analyses focused on effect of weather conditions on reports of commuting by bicycle. Results. Participants were diverse in age, gender and bicycle use, but were relatively well-educated; they traveled to work by bicycle on 34.5% of the logged commuting days. Modeling indicated that the likelihood of bicycle commuting increased in the absence of rain (odds ratio = 1.91; 95% confidence interval 1.42, 2.57) and with higher temperatures (1.03; 1.02, 1.04), and decreased with snow (0.90; 0.84, 0.98) and wind (0.95; 0.92, 0.97). Independent effects also were found for bicycle commuting distance, gender, and age, but not for daylight hours. Conclusion. Precipitation, temperature, wind and snow conditions had significant and substantial independent effects on the odds of travel to work by bicycle among a diverse panel of adult bicycle commuters. © 2011 Elsevier Inc. All rights reserved.
Introduction Concerns about health consequences of inadequate physical activity have increased interest in active transportation (American Public Health Association, 2009; Dora and Phillips, 2000). Accumulating evidence suggests potential for health benefits from greater participation in bicycle commuting, including achieving recommended physical activity levels and improving cardiovascular fitness and risk factors (Buehler et al., 2011; Oja, et al., 2011; Shephard, 2008). Utilitarian bicycle use is low in the U.S., but it is higher in countries with similar levels of development and weather conditions, suggesting the potential for increased participation (Pucher, et al., 2010). Evidence about influences on bicycle use for utilitarian travel can contribute to improved methods to support wider use of bicycling. Bicycling infrastructure is an important modifiable facilitator of participation (Dill and Carr, 2003), but understanding the impact of other factors also is important. Information about the influence of weather conditions on individual bicycle use behavior is lacking (Heinen et al., 2010; Saneinejad, et al., 2010). Relationships between aggregate bicycle use and aggregate weather data suggest significant associations for temperature and ⁎ Corresponding author. Fax: +1 802 656 8826. E-mail addresses: [email protected] (B.S. Flynn), [email protected] (G.S. Dana), [email protected] (J. Sears), [email protected] (L. Aultman-Hall). 0091-7435/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ypmed.2011.11.002
precipitation (Brandenburg et al., 2007; Buckley, 1982; Dill and Carr, 2003; Emmerson and Ryley, 1998; Nankervis, 1999; Nelson and Allen, 1997; Niemeier, 1996; Parkin et al., 2008; Pucher and Buehler, 2006; Winters, et al., 2007). At the individual level, Hanson and Hanson (1977) analyzed travel mode and weather diaries kept for 39 days to demonstrate correlations between bicycling to work and morning temperatures and cloud cover. The present study was planned to quantify the impact of specific weather conditions on individual use of bicycles for travel to work. Methods This longitudinal study documented reports of travel to work by bicycle among a convenience sample of adults on 28 pre-specified days over 10 months, and linked these reports to objective weather data. The study was conducted in five counties with populations ranging from 44,513 to 156,545 in the rural northeastern state of Vermont, U.S.A. Each study site included a main population center in a valley and smaller towns in rolling hills. We planned to recruit 100 participants using email lists from community organizations and workplaces, with sub-targets of one-third women, one-half over age 40, and all-season bicyclists comprising ≤one-quarter. Other criteria were: age ≥ 18, bicycle commuting frequency of > 2 annually, and commuting to work ≥ 2 miles. Full-time students were excluded. The study was approved by the University of Vermont Institutional Review Board. Weather data for the pre-specified days were obtained from four regional National Weather Service (NWS) stations based on morning commuting hours (5–9 a.m.) and proximity to participant residences: mean temperature
B.S. Flynn et al. / Preventive Medicine 54 (2012) 122–124
(Fahrenheit); mean wind velocity (miles/h); and precipitation (rainfall present/absent). Snow depth (inches) was measured at 7 a.m. by 18 NWS cooperating stations. Daylight hours were calculated for each community latitude and commuting day (www.usno.navy.mil/USNO). Age, gender, and reported distance typically traveled one-way to work by bicycle (miles) were obtained from baseline surveys. The outcome was a report of bicycle commuting or not on days that required a trip to the workplace, recorded by participants on web-based log forms. Baseline surveys were administered during May–July 2009. Log forms were completed during seven-day periods in each of four months: September 2009 (mean temperature 62°F), January 2010 (18°F), April (43°F), and July (71°F). Daily log records were linked to baseline data by unique identifiers, and to weather and daylight hours by location and date codes. The unit of analysis was person-day records for commuting days. We used a generalized linear model to estimate the independent impact of each weather factor on bicycle use for commuting days, while simultaneously controlling for other independent variables. Our dependent variable was ‘BIKED’ (yes or no) with a binary distribution assigned to the model. To account for correlation among observations from participants on multiple days, a “repeated” statement was included in the model.
Results Of 210 individuals who responded to recruitment activities, 185 met inclusion criteria. Sufficient data for modeling were obtained from 163, including 102 men and 61 women. Participants were more likely to be ≥40 years than younger (Table 1). College degrees were reported by 93% (compared to 33% for Vermont adults); 90% reported very good or better health (versus 61% in Vermont). Notably
Table 1 Participant characteristics for analytic sample by bicycle commuting distance each way (n = 163), Vermont, 2009–2010. n
Bicycle commute to work 2–5 miles (n = 72) %
>5 miles (n = 91) %
All (n = 163) %
Age 40 or under 40 +
66 97
50.0 50.0
33.7 66.3
40.5 59.5
Gender Male Female
99 64
58.3 41.7
65.3 34.7
60.7 39.3
Education b 2 yr. degree 2 yr. degree 4 yr. degree > 4 yr. degree
7 5 67 84
7.0 1.4 48.6 43.1
2.0 4.1 34.7 59.2
4.3 3.1 41.1 51.5
Health Excellent Very good Good Fair or poor
79 70 13 1
41.7 50.0 6.9 1.4
53.1 50.0 9.2 0
48.5 42.9 8.0 0.6
Children in household Yes 61 No 102
26.4 73.6
45.9 54.1
37.4 62.6
40.8 59.2
34.4 65.6
≥ 1 bike ride in the previous year for… …training or competition? Yes 56 23.6 No 107 76.4 …recreation or leisure? Yes 161 97.2 No 2 2.8 …other utilitarian purpose? Yes 140 86.1 No 23 13.9
100 0 85.7 14.3
98.8 1.2 85.9 14.1
123
low bicycle use for training or competition was reported, indicating that few were dedicated competitors. Baseline surveys also indicated that nearly 20% rode bicycles in winter months and that biking comprised about 35% of all trips to work (median 96 days annually; range: 5–288). The total number of daily log reports for days requiring a trip to work was 2554 person-days. In aggregate, participants reported biking to work on 881 (34.5%) of these commuting days. Weather conditions for logged days matched expectations for normal seasonal variation (Table 2). Temperature was correlated with snow depth (r = −0.72) and daylight hours (0.80). Precipitation and wind were not correlated with other weather factors or daylight hours. The dependent variable for bicycle commuting was regressed on independent variables temperature, wind, precipitation, snow, daylight, distance, age and gender. Nearly all factors had a significant independent relationship with bicycle commuting. Odds ratios showed participants were nearly twice as likely to commute by bicycle when there was no morning precipitation (Table 2). A similar effect was found for temperature, where a one degree increase raised the likelihood of biking by about 3%. A one mile per hour increase in wind speed decreased biking likelihood by about 5%. One inch of snow on the ground reduced the likelihood of biking by about 10%. Discussion This study makes a unique contribution to quantification of weather influences on bicycle commuting. A bicycle commuter panel was engaged over an extended time with good variation in participant characteristics and weather conditions. Results provided evidence for independent weather factor effects on individual decisions to travel to work by bicycle. Precipitation and temperature appeared to be strong influences on the odds of commuting to work by bicycle, consistent with other research. The odds of bicycle commuting nearly doubled when no precipitation was recorded for the morning commuting hours. Bicycle commuting decisions similarly appeared to be sensitive to temperature. Contrary to expectation from preliminary research, increased wind speed diminished the odds of bicycle commuting modestly. Snow depth had a dampening effect that might be expected when most participants did not ride bicycles in winter months. The study was conducted in a region with adverse winter conditions, but most participants commuted by bicycle throughout a significant part of the year, and some even in the coldest months. Improvements in infrastructure and maintenance, better equipment, and greater acceptance of cold weather bicycling could extend the range of conditions in which bicycle commuting becomes feasible for others (Bergstrom and Magnusson, 2003; Pucher and Buehler, 2006). A limitation of the current study was that the geographic areas covered by weather station data lacked local detail that might influence commuting decisions, potentially weakening the relationships studied. These data focused on morning commuting hours and did not account for participants with other work schedules. The snow depth measure reflected general winter conditions rather than daily road conditions. These community-recruited participants were diverse in terms of gender, age, and bicycle use, but were relatively well-educated compared to the general population, perhaps due to recruitment methods and inclusion criteria. Precipitation, temperature, wind speed, and snow depth had independent effects on the odds of commuting by bicycle to work among a diverse panel of adults who bike to work at least 2 miles. These results may support development of methods to mitigate adverse effects of weather on bicycle commuting and encourage greater use of active commuting. Increased use of active transportation for routine purposes can increase opportunities for physical activity and achievement of related health benefits.
124
B.S. Flynn et al. / Preventive Medicine 54 (2012) 122–124
Table 2 Odds ratios for regression of likelihood of bicycle commuting on hypothesized factors (n = 163), Vermont, 2009–2010. Effect variables were treated as independent factors in a generalized linear model. Effect
Odds ratio point estimate
95% confidence interval
Temperature (°F) Wind speed (mph) Precipitation (no vs. yes) Snow (inches) Daylight (hours) Distance (miles) Age (years) Gender (men vs. women)
1.03 0.95 1.91 0.90 1.00 0.92 1.02 2.65
1.02 0.92 1.42 0.84 0.99 0.89 1.01 1.78
1.04 0.97 2.57 0.98 1.00 0.96 1.04 3.99
Range
Mean
− 3.2–79.2 0.0–20.0 0.0–0.4 0.0–23.0 9.0–15.4 2.0–30.0 22–66 –
45.3 5.1 b0.1 2.6 12.4 7.7 43.8 –
Note: Temperature and wind were represented by mean values, and precipitation as rainfall present or absent through 9 a.m. on logged days at the National Weather Service (NWS) regional station nearest to a participant's residence. Snow depth was measured at 7 a.m. on these days at the nearest NWS cooperating station.
Conflict of interest statement The authors declare that there are no conflicts of interest.
Funding sources and role in study This work was funded by the United States Department of Transportation through the University Transportation Center Program at the University of Vermont and by the College of Medicine at the University of Vermont. Co-investigators from both organizations at the University of Vermont designed, conducted, and reported the study. Acknowledgments The contributions of Anne L. Dorwaldt and many community partners to recruitment of participants are gratefully acknowledged. References American Public Health Association, 2009. At the Intersection of Public Health and Transportation: Promoting Healthy Transportation Policy. Washington D.C.. Bergstrom, A., Magnusson, R., 2003. Potential of transferring car trips to bicycle during winter. Transp. Res. Part A 37, 649–666. Brandenburg, C., Matzarakis, A., Arnberger, A., 2007. Weather and cycling — a first approach to the effects of weather conditions on cycling. Meteorol. Appl. 14, 61–67. Buckley, C.A., 1982. Bicycle traffic volumes. Transp. Res. Rec. 847, 93–102. Buehler, R., Pucher, J., Merom, D., Bauman, A., 2011. Active travel in Germany and the U.S. — contributions of daily walking and cycling to physical activity. Am. J. Prev. Med. 41, 241–250.
Dill, J., Carr, T., 2003. Bicycle commuting and facilities in major U.S. cities: if you build them commuters will use them. Transp. Res. Rec. 1828, 116–123. Transportation, environment, and health. In: Dora, C., Phillips, M. (Eds.), Copenhagen: World Health Organization, Regional Office for Europe, WHO Regional Publications No. 89. Emmerson, P., Ryley, T.J., 1998. The impact of weather on cycle flows. Traffic Eng. Control 238–243 (April). Hanson, S., Hanson, P., 1977. Evaluating the impact of weather on bicycle use. Transp. Res. Rec. 629, 43–48. Heinen, E., Van Wee, B., Maat, K., 2010. Commuting by bicycle: an overview of the literature. Transp. Rev. 30, 59–96. Nankervis, M., 1999. The effect of weather and climate on bicycle commuting. Transp. Res. Part A 33, 417–431. Nelson, A.C., Allen, D., 1997. If you build them, commuters will use them: association between bicycle facilities and bicycle commuting. Transp. Res. Rec. 1578, 79–83. Niemeier, D.A., 1996. Longitudinal analysis of bicycle count variability: results and modeling implications. J. Transp. Eng. 200–206 (May/June). Oja, P., Titze, S., Bauman, A., et al., 2011. Health benefits of cycling: a systematic review. Scand. J. Med. Sci. Sports 21, 496–509. Parkin, J., Wardman, M., Page, M., 2008. Estimation of the determinants of bicycle mode share for the journey to work using census data. Transportation 35, 93–109. Pucher, J., Buehler, R., 2006. Why Canadians cycle more than Americans: a comparative analysis of bicycling trends and policies. Transp. Policy 13, 265–279. Pucher, J., Dill, J., Handy, S., 2010. Infrastructure, programs, and policies increase bicycling: an international review. Prev. Med. 50, S106–S125. Saneinejad, S., Kennedy, C., Roorda, M.J., 2010. Assessing the impact of weather and climate on commuter trip behaviour in Toronto. Paper presented at XVI PanAmerican Conference of Traffic Control and Transportation Engineering, Lisbon. Shephard, R.J., 2008. Is active commuting the answer to population health? Sports Med. 38, 751–758. Winters, M., Friesen, M.C., Koehoorn, M., Teschke, K., 2007. Utilitarian bicycling a multilevel analysis of climate and personal influences. Am. J. Prev. Med. 32, 52–58.