ABSTRACT This study presents the development of a new dataset of homogenized temperature for use in trend analysis and monitoring climate change in Canada. This dataset contains daily data for 780 locations across the country: 508 locations with an active station (current observations) and long record (starting prior to 1990); 53 locations with an active station and short record (starting after 1990); and 219 locations with no current observations (station closed) but with more than 30 years of data. Daily observations from nearby sites were often merged into a single record to create a long time series. This new dataset includes observations taken at Reference Climate Stations and from the Canada Aviation Weather Services, which are used to extend past climate observations into recent times. First, the data were quality controlled. The daily minimum temperature was adjusted for the change in observing time at principal stations in 1961. Parallel daily data were used to detect non-climatic shifts when the observations from nearby sites were merged. Series of annual and seasonal mean temperatures were tested for homogeneity. Daily temperatures were adjusted using a quantile-matching procedure if needed. Two main causes of data inhomogeneity affecting the trends over the 1948–2018 and 1900–2018 periods were identified. First, the change in observing time in 1961 introduced a cold bias in the annual means of the daily minimum temperatures after 1961. Second, merging observations from airport stations with older records has often created an artificial decreasing shift in the unadjusted data because of the better exposure of the instruments at airport stations. This new homogenized dataset shows a slightly stronger warming than the unadjusted data: the trend in the annual mean temperature for Canada has changed from 1.69° to 1.74°C for 1948–2018, and the trend for southern Canada has changed from 1.32° to 1.62°C for 1900–2018 because of all the adjustments applied to daily temperature in this study.

中文翻译：

---

用于趋势分析和监测加拿大气候变化的第三代均质温度

摘要 本研究介绍了用于趋势分析和监测加拿大气候变化的新均质温度数据集的开发。该数据集包含全国 780 个地点的每日数据：508 个地点有活动站（当前观测）和长期记录（从 1990 年开始）；53 个有活动站和短记录的地点（从 1990 年开始）；和 219 个没有当前观测（站关闭）但有 30 多年数据的地点。来自附近站点的日常观察结果通常合并为一个记录，以创建一个很长的时间序列。这个新数据集包括在参考气候站和加拿大航空气象服务处进行的观测，这些观测用于将过去的气候观测扩展到近期。首先，数据经过质量控制。1961 年，根据主要站点观测时间的变化调整了日最低温度。当合并附近站点的观测值时，使用平行日数据来检测非气候变化。测试了一系列年度和季节性平均温度的均匀性。如果需要，使用分位数匹配程序调整每日温度。确定了影响 1948-2018 年和 1900-2018 年期间趋势的数据不均匀性的两个主要原因。首先，1961 年观测时间的变化导致 1961 年后每日最低气温的年均值出现冷偏差。在机场车站暴露仪器。