Technical Thematic Report No. 5. - Canadian climate trends, 1950-2007
Climate data are records of observed climate conditions. They are taken at specific sites and times with particular instruments under a set of standard procedures. A climate dataset, therefore, reflects not only climate condition, but also other non-climate-related factors such as where and how the observations have been taken. For example, a change in observing procedures and/or instrumentation can introduce a non-climatic change in the data series. These artefacts in the climate data are removed, as much as possible, through the creation of homogenized data series in order to provide reliable assessments of climate trends. Trends presented in this report are largely based on homogenized data.
The temperature data used in the analysis were from 210 relatively evenly distributed stations across the country that had been rigorously checked and corrected for known sources of systematic error (such as station shifts, changes in observing procedures, and excluding stations with strong urban warming effects) (Vincent, 1998). This dataset has been used in previous studies of changes in Canadian temperature and its extremes (Zhang et al., 2000; Bonsal et al., 2001; Vincent and Mekis, 2006). Trend information was computed over the period 1950 to 2007, and daily mean temperatures were computed from the average of the daily minimum and maximum temperature.
The precipitation data used include adjusted daily rainfall and snowfall amounts observed at 495 stations across the country (Mekis and Hogg, 1999). All known inhomogeneities in the station data caused by changes in location and precipitation measurement programs were carefully minimized: wind undercatch, wetting loss, evaporation, trace events, and varying snow densities were also considered in the adjustment procedure. A subset of this dataset has been used in other studies to investigate changes in heavy precipitation events in Canada (Zhang et al., 2001b) and trends in precipitation intensity in Canada (Vincent and Mekis, 2006). The variables selected for trend analysis were annual and seasonal precipitation totals, the fraction of annual precipitation falling in solid form (expressed as a percentage), and the number of days with measureable precipitation (greater than trace amounts).
Information on trends and variability in snow cover were derived from daily snow depth observations made at climate and synoptic stations. Daily snow depth observations from manual ruler measurements have been made at most Canadian synoptic stations since about the mid-1950s. The daily observing program was extended to climatological (cooperative) stations in the early 1980s, approximately quadrupling the number of stations in the Canadian network to about 2000. However, there are only about 150 stations with more-or-less complete data from 1950 to 2007 for monitoring changes in snow cover conditions in Canada. The data for this report were taken from a recent update of the Canadian Snow CD (Meteorological Service of Canada, 2000) which includes data rescue of previously undigitized Canadian snow depth data and the reconstruction of missing values as outlined in Brown and Braaten (1998). Only stations with 47 years or more of data were included in the analysis and trends were computed over 57 snow seasons from 1950/51 to 2006/07. It should be noted that most of the daily snow depth observations are made at open sites in or near populated regions and may not be representative of the surrounding snow cover, particularly in regions with higher terrain (such as British Columbia and Alberta) and forest cover, as snow in open terrain tends to melt out faster than snow in vegetated areas. The station distribution is also strongly biased toward southern latitudes with major data gaps in areas above about 55°N. The snow cover variables presented in this report are snow cover duration (SCD) defined as the number of days with 2 cm or more of snow on the ground from an August to July snow year, and the annual maximum snow depth. SCD is computed over the fall (August to January) and spring (February to July) halves of the snow year to provide information on changes in the onset and melt dates of snow cover. An assessment of the homogeneity of daily snow depth observations was carried out by Brown and Braaten (1998) with little evidence of detectable inhomogeneities.
The timing of ice formation and thaw on water bodies are important indicators of climate condition. Tracking and analyzing these “ice-on” and “ice-off” events is known as ice phenology. Historical Canadian ice phenology data are archived in the Canadian Lake Ice Database developed by Lenormand et al. (2002) based on ice observation programs managed by the Meteorological Service of Canada and the Canadian Ice Service. Some additional data has been added from the volunteer IceWatch program (IceWatch, 2008b). While there are a large number of observations in the database, there are relatively few sites with continuous observations spanning several decades suitable for trend analysis, and observing programs ceased at many sites during the 1990s. This report therefore draws extensively on the results of published trend analyses of Canadian ice phenology data provided by Zhang et al. (2001a) and Duguay et al. (2006) and, more recently, the analysis of Latifovic and Pouliot (2007) who used visible satellite imagery to extend the ice phenology record at about 40 lake sites across Canada.
The availability of water is important for ecosystems, especially in relatively dry regions. For this analysis, the Palmer Drought Severity Index (PDSI) was used as an index of water availability and was computed from observed temperature and precipitation data. The PDSI reflects changes in long-term moisture, runoff, recharge, deep percolation, and evaporation, and is useful for drought analysis over time spans of months or seasons. A positive value indicates wetness, while a negative value indicates dry condition. PDSI modeling requires the availability of co-located and concurrent mean air temperature and total precipitation observations. For this report, PDSI was computed from the rehabilitated historical temperature and precipitation datasets described above for 80 stations with more-or-less complete temperature and precipitation data from 1950 to 2007.
The availability of heat for plant growth was investigated by calculating the number of growing degree days over the growing season. The start of the growing season was defined as the date when mean temperatures were greater than 5°C over 5 consecutive days in the spring and the end of the growing season was defined as when the inverse condition was met. The average temperature summed over this period is the number of growing degree days (a measure of accumulated heat during the growing season) and was computed at all temperature stations.
- Date Modified: