by Tim Storer
Heat waves have impacts on a variety of industries and demographics throughout the world, and they are much scientific interest. In the United States, research has been done to study the trends in heat wave intensity and frequency. Unfortunately for the researchers, there have traditionally been many varying definitions of “heat waves,” in previous research, and additional work was required to standardize the data so that it could be analyzed for trends. The varying definitions are due to a multitude of interested parties, such as health researchers measuring heat waves that break a dangerous threshold, versus climate scientists tracking temperatures that are in high percentiles. Having compiled all the different definitions, the recent research showed mostly positive trends in the number of heat wave days per year in most of the United States, with the strongest increases in the Southeast and Great Plains regions (Smith et al. 2013). This approach is especially powerful, because by combining many different measures of heat waves, the study is likely to eliminate oversight that any individual measure may have.
Tiffany Smith, Benjamin Zaitchik and Julia Gohlke compiled data over a time period of 1977–2011 from the North American Land Assimilation System (NLDAS), though only the data from each warm season (May 1–Sept 30) was analyzed in their study. The NLDAS records a variety of weather indices from across the country, such as temperature, humidity, and wind speed. It is an excellent data source because of its high resolution (12 km). The researchers divided fifteen pre-existing heat wave indices (HI) into two categories: those that used an absolute measure of heat, and those with relative measurements. The absolute HI each have different threshold requirements, and each relative HI has a slightly different means of measurement, varying both in the percentile used, and the type of temperature measurement (daily average versus daily minimum, for example). All of these variations were described in detail in the study. Using the NLDAS data, the researchers recorded the number of annual heat wave days at each grid cell location with all fifteen different HI. To give more detail and insight, the continental U.S. was divided into six different geographical regions each composed of 3–12 states, with the goal that these areas would approximately match different pre-established climate regions. From this data, many standard statistical tests were used to analyze trends in heat wave days in each area over the studied period.
Depending on which HI was used, different regions (Northwest, Southeast, or Southwest) showed higher average heat wave frequency over the period studied. However, none of the HI showed the Great Plains, Midwest, or Northeast as having the highest heat wave frequency. It was also observed that relative HI had a more consistent measure of average heat wave count as compared to absolute HI, meaning that heat wave studies based on absolute values are more variable. These findings say nothing about climate change trends over time, but do show how the variation of heat wave definitions has serious implications for the observed distribution and frequency of heat waves. Therefore, when studying trends in warm temperatures, it is prudent to consider which definitions are relevant and to acknowledge the differences.
All fifteen HI showed statistically significant trends in annual heat wave day frequency in certain areas in the United States. In the case of each HI, most of the significant trends were positive, meaning that the overall number of heat wave days has increased regardless of which definition is used. These trends varied greatly by region, with the Southeast and Great Plains regions having the strongest trends both in terms of trend size and in percentage of land area in which the trend was observed. For example, ten of the fifteen HI show a higher percentage of land area experiencing increasing trends in the Southeast compared to the Northeast, and nine of those ten also show larger trends.
It was noted that these results differ from past studies in certain regards, such as findings about increased heat waves in the Southeast. The discrepancies can be attributed to the methodologies of other studies and the decision to use a much narrower definition of heat waves. This study does have some shortcomings, such as a relatively short time period studied, the inability to address concerns associated with localized urban heat islands, and a lack of multi-day analysis in addition to studying the number of single-day events.
Smith, T., Zaitchik, B., Gohlke, J., 2013. Heat waves in the United States: definitions, patterns, and trends. Climatic Change vol. 118, issue 3–4, 811–825. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3711804/