A New Approach to Air Temperature Analysis

Two new aspects of measured air temperature data analysis are examined: the extent of diurnal window for identification of temperature extrema and the potential for use of diurnal extrema timing as a new climate parameter. In this study we analyze a 64 year long hourly temperature data set from Toronto Pearson International Airport climate station to identify the impact of the observing window on diurnal extrema determination and evaluate shifts in annual extrema timing. Accuracy of temperature extrema identification is undeniably of vital importance for the authenticity of mean temperature calculation on all time steps. Due to a common absence of hourly air temperature observations diurnal temperature extrema are often the only available choice for the climatological analysis. While diurnal maxima are often properly characterized, frequent minima mischaracterization has been related to the climatological observing window presently in use. Climatological observing window (COW) is defined as a time frame over which continuous or extreme air temperature measurements are collected. A fixed 24-hour COW leads to potential misidentification of minima due to fragmentation of “nighttime” into two subsequent nighttime segments caused by the time discretization interval. Correct identification of diurnal air temperature extrema is achievable using a COW that identifies minimum over a single nighttime period and maximum over a subsequent daytime period, as determined by sunrise and sunset (Figure 1). Climatological Observing Window Night and Day (COW_N-D) aims at identification of “turning points” of air temperature-time function or the mathematical extrema. Specifically, mathematical air temperature extrema are the points in which the slope of a tangent on the air temperature-time curve changes its sign. Climatological studies almost invariably rely on daily air temperature extrema or their diurnal average as a parameter for detection of changes to the climate system. However, diurnal air temperature variability is truly characterized by both, the temperature extrema and their specific times of occurrence. Yet, routine recordings of diurnal air temperature minima and maxima are not available. Timing of diurnal extrema refers to two specific time points in diurnal air temperature cycle that identify the exact position of extrema with respect to both, temperature and time axes. Application of COW_N-D on hourly temperature-time series for identification of daily extrema generates diurnal maxima and nocturnal minima populations. Further, with midnight and noon chosen as the points of delineation, nocturnal minima and diurnal maxima are further divided into “before” and “after” subpopulations for the analysis of time-shifts in historical temperature-time series. Nocturnal minima timing population is subdivided into Before Midnight Minima (BMM) and After Midnight Minima (AMM) while daytime maxima timing population is further divided into Before Noon Maxima (BNM) and After Noon Maxima (ANM). The Mann-Kendall analysis reveals statistically significant positive trends in annual averages of AMM and ANM subpopulations further supported by the migration of BMM and BNM subpopulations across the delineation point (Figure 2). The results of this study suggest that shifts in annually averaged extrema timing, consistent with shifts in annually averaged temperature extrema, can be used as indicators of changes to the climate system when daily temperature extrema are identified using COW_N-D. With improved characterization of diurnal extrema and identification of extrema timing patterns this new approach to air temperature analysis offers a unique insight into the evolution of temperature-time trends.