Objective: To compare diary-reported trips and Global Positioning System (GPS) recording trips and identify the reasons for the discrepancies between two methods. Methods: The survey was conducted in Jiangning district in Nanjing city during July to September and November to December, 2015. Both the diary records and GPS method were used to investigate the travel behavior of 33 retired adults for 5 consecutive days. The GPS traces were display in Google Earth and then split into trips to compare with the diary records according to time and location. χ(2) test was used to analyze the influence of trip characteristics on misreporting rates of each method. Results: A total of 1 087 trips in the survey can be compared between the diary (n=909) and the GPS method (n=912). 7.3% (79/1 087) of the trips were only recorded by GPS method, and 11.8% (128/1 087) were only reported in the diary. Of the remaining 880 trips recorded by the both methods, 86.7% (763/880) matched each other, while 13.3% (117/880) did not. For the matched trips, the difference between the trip durations recorded by diary and GPS method was 2.0 (quartile was 6.0) minutes and the diary method overestimated about 25.0% trip durations when compared with the GPS method. The accuracy rates were 84.8%(903/1 065) and 86.9%(925/1 065) for diary and GPS method, respectively. Both methods were more likely to misreport the trip under 5 minutes. The misreporting rates of diary method for trips under 5 minutes were 6.8 times higher than trips over 30 minutes (21.7% vs 3.2%). The reporting accuracy was also significantly different among trips by different travel mode (P<0.05) for both methods, diary method had the lowest accuracy in reporting vehicles recorders (69.3%, 133/192), while GPS method was more accuracy in both biking (91.9%, 136/148) and vehicles recorders(89.6%, 172/192). The main reasons for misreporting by diary method were forgetting or unwilling to record and failing to record trips according to travel mode, while forgetting to carry GPS device, bad GPS signal and failing to locate the position were the main reasons for misreporting by GPS method. Conclusion: The majority of the trips reported in diary and GPS method are well consistent with each other. Both of the methods should be used simultaneously in the survey of time-activity pattern to improve the data quality.
目的： 为提高空气污染暴露评估的准确性，开展出行信息调查，比较日志记录与全球定位系统(GPS)法调查人群出行信息的一致性，探讨造成两种方法调查结果不一致的原因。 方法： 于2015年7—9月及11—12月，以南京市江宁区33名退休人群作为研究对象，同时采用日志法及GPS法进行为期5 d的出行行为调查。使用Google Earth进行GPS数据的提取和展示，计算数据的完整性，并依据时间和位置与日志记录进行匹配分析，计算两种方法一致率；采用χ(2)检验分析不同行程特征对方法漏报、误报率的影响。 结果： 两种方法共监测到1 087条出行记录（GPS法912条，日志法909条），其中7.3%（79/1 087）的行程仅被GPS记录，11.8%（128/1 087）的行程仅被日志法记录。而在两种方法同时监测到的行程记录（880/1 087）中，86.7%（763/880）匹配，13.3%（117/880）不匹配；对匹配的行程单独分析发现，日志法记录的行程持续时间大于GPS法（P<0.001），差值中位数（四分位数间距）为2.0（6.0）min，与GPS法相比，日志法高估了25.0%的行程持续时间；日志法和GPS法的行程准确率分别为84.8%（903/1 065）和86.9%（925/1 065），两种方法均倾向于漏报误报持续时间≤5 min的行程，其中日志法对≤5 min的行程漏报误报率（21.7%）远高于>30 min行程（3.2%）的漏报误报率。不同交通方式的出行监测准确率差异存在统计学意义（P<0.05），日志法对机动车出行监测准确率最低（69.3%, 133/192），而GPS法对骑车（91.9%, 136/148）和机动车（89.6%,172/192）出行监测的准确性均较高。造成日志法漏报、误报的主要原因为忘记记录、主观上不愿意报告和没有将采用不同交通方式的行程分别记录；而造成GPS法漏报、误报的主要原因为未携带GPS设备、GPS信号质量不良和因搜索信号造成的时间延迟。 结论： 日志法及GPS法总体上具有较好的一致性，但均存在一定的漏报和误报等问题，将两种方法结合使用将有助于提高监测结果的准确性。.
Keywords: Air pollution; Environmental exposure; Global positioning systems; Time-activity pattern.