Changes in monthly precipitation and temperature have been fairly well documented for North America. Changes in extremes, however, have not been as well documented as they pose several additional challenges which have been addressed in this research. First the homogeneity of daily precipitation, maximum temperature and minimum temperature had to be carefully assessed. The analysis uses homogeneity adjusted Canadian daily data, the results of statistical homogeneity tests on data from the contiguous United States to identify a subset of daily temperature and precipitation time series that were homogeneous and detailed station by station assessments of the homogeneity of data from stations in Mexico, Alaska, Hawaii, Puerto Rico and the U.S. Virgin Islands using a variety of graphs, examination of metadata, a statistical homogeneity test and consultation with relevant State Climatologists to identify the most homogeneous stations.

The analysis uses and builds on a suite of indices of extremes developed by the Expert Team on Climate Change Detection and Indices and its predecessors. This team has been formed jointly by the World Meteorological Organization (WMO) Commission for Climatology (CCl), the World Climate Research Programme (WCRP) project on Climate Variability and Predictability (CLIVAR) and the Joint WMO-Intergovernmental Oceanographic Commission of United Nations Educational, Scientific and Cultural Organization Technical Commission for Oceanography and Marine Meteorology (JCOMM). One of the authors is a current member of the team and another served on a predecessor team. The indices used are both straightforward and complex. An example of a simple index is one that assesses the warmest maximum temperature recorded during the course of a year. A more complex approach is used for percentile indices that examine extreme values during all times of the year. These indices have their thresholds determined by the mean of a 30 year base period for years outside the base period, but during that base period a boot strap approach is used to determine the threshold based on 29 years of data, i.e., excluding the year being assessed.

In addition to presenting changes in temperature and precipitation extremes, changes in the start versus end of the growing season will also be presented along with changes in the phenomenon of false springs, which is when the growing season starts but is then followed by a killing frost.