Boston: July Days Over 90 Degrees
By mid-century in Boston, the sweltering heat of July 2010 may be thought of as cooler-than-average conditions, as more days above 90°F routinely occur. The 2010 dates on this "postcard from the future" for Boston are actual; 2050 dates are representative of average projections. Postcards are also available from the cities of New York, Philadelphia and Washington, DC.
DATA ANALYSIS METHODS
The statistics for July 2010 are based on values from the National Weather Service observation sites for each of the cities (see for example the Weather Service’s Baltimore/Washington office for data from that area). Monthly summaries are available that list departures from average, or anomalies, in the mean temperature, as well as each day’s high and low temperatures. Using high temperatures, we computed days above three thresholds: 90°F, 95°F and 100°F.
The values of historical record years for mean temperature and hot day numbers were taken from the observations available at the stations within each of the cities that offer the longest record of daily historical values. These were obtained from the National Climatic Data Center’s SDO dataset.
On the basis of the historical values of daily high, low, and mean temperatures, we computed, for each year available, the number of days in July above each high temperature threshold, and the mean July temperature. We then determined the standing records, i.e., the year with the most July days above 90°F, that with most days above 95°F, that with most days above 100°F, and that with the highest average July temperature.
Changes in Extremes
(Counts of days in July above high temperature thresholds)
To obtain the statistics for the average number of hot days (above 90°F, 95°F and 100°F) for the current climate (1981-2000) compared to the typical number of such days in the future climate by 2050, we used a process described in the computation of hot days in August. We replicated that analysis using July data instead of August.
Changes in Mean July Temperature
To compute by how much average July temperature may change in the future, and compare that change to the July 2010 departure from historical averages, we used climate model simulations of future changes for the lower 48 United States. This data has a level of geographic detail of about 12 kilometers, and we extracted the data that the models simulate at the locations of our cities of interest (corresponding to the grid point closest to the city’s latitude/longitude coordinates). This data has been obtained by a procedure that is detailed on the Climate Wizard website.
We took these so-called downscaled (more highly detailed) computer model projections for July average temperatures from 16 global climate models, and used emissions scenario A1B, which assumes that emissions will continue at a rate similar to that of today’s emissions.
We looked at the average change predicted by these 16 models by 2050, averaged the July temperatures of the twenty years straddling the future date, and compared them to current average values from the same simulations.