Traffic accident increase attributed to Daylight Saving Time doubled after Energy Policy Act
CCurrent Biology (2020) (7) R298-R300, doi: https://doi.org/10.1016/j.cub.2020.03.007 Licences CC-BY-NC-ND Traffic accident increase attributed to Daylight Saving Time doubled after EnergyPolicy Act
Jos´e Mar´ıa Mart´ın-Olalla ∗ Universidad de Sevilla. Facultad de F´ısica. Departamento de F´ısica de la Materia Condensada. ES41012 Sevilla. Spain (Dated: February 6, 2020)On January 30, 2020 Current Biology released the report “A Chronobiological Evaluation of theAcute Effects of Daylight Saving Time on Traffic Accident Risk” (doi: 10.1016/j.cub.2019.12.045by Fritz et al. where it was argued that fatal traffic accident risk increases by 6 % in the US due toDaylight Saving Time spring transition. This manuscript is a 1000 word correspondence showingthat the bulk of this reported risk comes from the transition dates mandated by the Energy PolicyAct in 2007.
Keywords: circadian, time zones, latitude, summer time, seasons, circadian misalignment, meridian, enter-tainment (light), transportation, road accidents, traffic accident
This is the accepted manuscript of a correspondence published by Current Biology Volume 30, Issue 7, PagesR298-R300. ISSN 0960-9822, ESSN 1879-0445.For the authored, published version please visit doi https://doi.org/10.1016/j.cub.2020.03.007The impact of Daylight Saving Time (DST) transitionson the human circadian system and on everyday life iscurrently subject to close analysis. Fritz et al. [1] stud-ied recently large scale United States (US) registry data(1996 to 2017) on fatal motor vehicle accidents (MVA)and reported the incidence rate ratio. The authors reportresults for data before 2006, for data after 2007, and forthe whole observation period. They also report morn-ing, afternoon and whole-day results. The discussion andconclusions are extracted mainly from the whole-day fig-ures and those from the entire 1996 to 2017 period. Yetthe breakdown illustrates a most interesting fact: theamendments in the Uniform Time Act made by the En-ergy Policy Act doubled the increase of fatal morningMVA attributed to the spring transition.Figure 1 shows a simplistic scenario at 40 ◦ N —the lati-tude of New York City and Madrid— in which the yearlyevolution of the solar altitude z at the latest sunrise time—here after the H-hour— is plotted (black line). Solaraltitude is appropriate for understanding a key parame-ter related to the rate of traffic accidents: illuminationconditions (see Figure 1B in Ref. [1]). The H-hour isa function of latitude only if expressed as a mean solartime or as a distance to solar noon. It impacts humansocial life since an activity starting at or after the H-hourwould certainly occur in the photoperiod irrespective ofcalendar date. People at 40 ◦ N latitude would see theSun crossing the horizon in January 5th at the H-hourand would see the Sun as high as z = 30 ◦ above thehorizon in June at the same hour of the day, if noth- ∗ [email protected]; ing else changes. As the Sun gets higher in the sky inspring/summer more people find natural to advance theiractivity aiming to mitigate their exposure to the highestinsolation: at 40 ◦ N insolation efficiency climbs up to atropical 95 % (see right axis in Figure 1). Likewise, as thesolar altitude decreases in autumn/winter people delaytheir activity aiming to mitigate their exposure to morn-ing darkness. This circanual cycle is observed in tropical,pre-industrial societies[2]. It is also observed in extrat-ropical, industrial societies synchronized by clocks[3, 4]and which have adopted year-round time schedules whereDST transition dates regulate the mechanism. Chronobi-ologists largely criticizes[1, 5] the practice, yet they even-tually acknowledge that people are prone to this naturalbehaviour[6].In Figure 1 the spring transition brings the humanactivity at the H-hour from the black line to the blueline. In fall, the shift is reversed. Vertical colored stripbands display transition dates. The Energy Policy Actadvanced in 2007 the spring transition by three weeks.In 2020 z will change from z st ∼ ◦ to z dst ∼ ◦ at 40 ◦ Nin the US. It changed from z st ∼ ◦ to z dst ∼ ◦ before2006. In Europe at 40 ◦ N, it will change from z st ∼ ◦ to z dst ∼ ◦ this year.Either of the last two cases keeps the Sun high enoughover the horizon so that illumination conditions at the H-hour are unaltered. Therefore an increase of traffic acci-dents due to changing illumination conditions is expectedto be slight. However, the changing in illumination con-ditions is dramatic for the current US regulations: a nicedeal of traffic volume is shifted close to dawn after thetransition.Several cons and pros must be further considered: first,a fraction of population commutes before H-hour in theTypeset by REVTEX a r X i v : . [ phy s i c s . pop - ph ] A p r -18°0°18°36°54°72°90° 31%59%81%95%100% T w ili gh t D a y li gh t Zenith H-hour (ST)H-hour (DST)
Latitude=40°NH-hour=07:22JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC S o l a r a l t i t ude , z I n s o l a t i on e ffi c i en cy , × s i n ( z ) ST to DST DST to ST ☀
1° 8° 18° 14° 18° 18° 14° 9° 4° -2° 2° U S ( a ft e r ) E URU S ( - ) U S ( - ) U S ( ) E UR ( ) U S ( a ft e r ) E UR ( a ft e r ) / U S ( be f o r e ) E UR ( be f o r e ) ☀ ☀ ☀ ☀ ☀ ☀ ☀ ☀ ☀ ☀ ☀☀ FIG. 1.
The seasonal variation of solar altitude at ◦ N . The solar altitude z at the hour of the latest sunrise of theyear —black, H-hour (Standard Time, ST)— and at the preceding hour –blueish, H-hour (Daylight Saving Time, DST)— asa function of calendar date for 40 ◦ N latitude where the H-hour is 07:22 mean solar time, the sunrise time on January the 5th.Coloured vertical strip bands display different arrangements of transition dates; each one expands for one week. The yellowishline displays solar altitude at solar noon. Sunrise is set at z c = − . ◦ due to solar finite apparent size and atmosphericrefraction. winter dawn, they are more prone to see the Sun belowthe horizon after DST begins in the middle of March;second, z d still worsens as latitude decreases (see Supple-mentary material); third, and although figure 1 is show-ing solar properties and thus it is insensitive to longitude,human social activity is dictated by standard clock timewhich makes z dst worsen westward of a time meridian,an effect discussed by the authors. On the other side,human activity at the H-hour in the US slightly reducedcompared to that in Europe at this time[7] which wouldallow a slightly larger extension of DST regulations inthe US.Fritz et al. [1]’s results remarkably agree with thisframework. Figure 2 of the Ref.[1] attributes to DSTspring transitions before 2006 and after 2007 a 5 % in-crease in afternoon fatal MVAs and a 6 % increase inmorning fatal MVAs before 2006. The relevance of theseresults ( 5 %) should be evaluated by comparing the out-come to the relative standard deviation (RSD) of weeklyaveraged fatal MVAs which is ∼
15 % (see Figure 1A ofRef. [1]). The attributed increase is just one third ofRSD. Also the lower endpoint of the confidence intervalfor each of these observations barely reaches 1.00–1.01.Understandably, the authors agree that this impact isslight. These observations can be plausibly linked to cir-cadian misalignment and sleep deprivation since illumi-nation conditions were unaltered here.In sharp contrast, the morning increase attributed toDST after 2007 more than doubles the preceding results:13 % versus 5 % to 6 % with the lower endpoint of the con- fidence interval at 1.06, markedly greater than 1.00 (seeFigure 2 of Ref. [1]). Indeed the morning increase after2007 is the one and only result close to RSD. It is there-fore a marked impact. Following Figure 1, a plausiblecause of this spike is the change in transition dates dueto the Energy Policy Act, which brought a dramatic andpredictable change in illumination conditions close to therush hour. Notice that DST onset has regularly occurredin the Northern hemisphere from the end of March to theend of April, only after the sun is high enough in the sky.Illumination issues could also arise at the fall transi-tion. Currently z dst decreases to ∼ ◦ in Europe and to − ◦ in the US —meaning that the latest sunrise (clock)time of the year actually happens on the first Saturdayof November— compared to z dst ∼ ◦ on the last Sun-day of September, a transition date in continental Europediscontinued in 1996. Very likely an advance of the falltransition date by some three weeks could result in a de-crease of morning MVAs, mimicking the results obtainedby Fritz et al.Summarizing, the analysis shows that the acute impactof DST transitions on incidence rate ratio for weekly av-erage fatal MVAs is close to 5 %, one third of the RSD ofobservations, a slight impact. However, the results showthat the Energy Policy Act more than doubled morningincidence rate ratio up to 13 % due to a marked changein the illumination conditions. Based on the first obser-vation, the authors call for discontinuing DST policies,despite its impact being slight. Instead the outstandinganalysis provided by Fritz et al. calls for discontinuingthe provisions of the Energy Policy Act relative to tran- sition dates. [1] Josef Fritz, Trang VoPham, Kenneth P. Wright, andC´eline Vetter, “A Chronobiological Evaluation of theAcute Effects of Daylight Saving Time on Traffic AccidentRisk,” Current Biology , 1 (2020).[2] Gandhi Yetish, Hillard Kaplan, Michael Gurven, BrianWood, Herman Pontzer, Paul R Manger, Charles Wil-son, Ronald McGregor, and Jerome M Siegel, “Naturalsleep and its seasonal variations in three pre-industrial so-cieties.” Current Biology , 2862–2868 (2015).[3] Daniel Monsivais, Kunal Bhattacharya, Asim Ghosh,Robin I. M. Dunbar, and Kimmo Kaski, “Seasonal andgeographical impact on human resting periods,” ScientificReports , 10717 (2017).[4] Jos´e Mar´ıa Mart´ın-Olalla, “The long term impact of Day-light Saving Time regulations in daily life at several circlesof latitude,” Scientific Reports , 18466 (2019).[5] Till Roenneberg, Anna Wirz-Justice, Debra J. Skene,Sonia Ancoli-Israel, Kenneth P. Wright, Derk-Jan Dijk,Phyllis Zee, Michael R. Gorman, Eva C. Winnebeck, andElizabeth B. Klerman, “Why Should We Abolish DaylightSaving Time?” Journal of Biological Rhythms , 227–230(2019).[6] Till Roenneberg, Eva C. Winnebeck, and Elizabeth B.Klerman, “Daylight Saving Time and Artificial TimeZones – A Battle Between Biological and Social Times,”Frontiers in Physiology , 944 (2019).[7] Jos´e Mar´ıa Mart´ın-Olalla, “Latitudinal trends in humanprimary activities: characterizing the winter day as a syn-chronizer,” Scientific Reports , 5350 (2018). ACKNOWLEDGMENTS
Prof. Dr. Juan C. del ´Alamo from University ofWashington and Prof. Dr. Manuel Garc´ıa-Villalbafrom Universidad Carlos III alerted JMM-O of thecommented paper. JMM-O thanks Prof. Dr. JorgeMira from Universidade de Santiago de Compostela fora critical reading of this letter. An earlier version ofFigure 1 was presented in a meeting held on 2018 at theConsello da Cultura Galega in Santiago de Compostela(Spain). The figure was produced by gnuplot . Itplots a file produced by a script in octave whichuses xplanetxplanet