Temperature analysis at 8 a.m. Monday, Nov. 23, 2015 (WeatherBell.com).

For the first time this fall, the temperature at Reagan National Airport dropped to the freezing mark. It came five days later than the 1981-2010 average of Nov. 18.

This year’s first fall freeze at D.C’s official observing site also comes a bit later than the last two years, 2013 and 2014, when the mercury first hit 32 on Nov. 13 and Nov. 15, respectively.

In the most recent climate period, spanning 1981 to present, the earliest first freeze on record occurred on Oct. 20, 1992 and the latest Dec. 22, 2001.

If we include the entire period of record, dating back to 1872 (including observations from 24th and M St. in the District prior to 1945), the earliest recorded first freeze occurred Oct. 10, 1895 and the latest remains Dec. 22, 2001.

During the first 30 years of the climate record, spanning 1872 to 1901, the average first freeze date was Oct. 30, meaning this marker has shifted forward about three weeks over 143 years. The later freeze dates are likely due to increasing urbanization and climate warming. There is also a possibility the change in observing site from downtown to the airport contributed a small amount to the shift.

CWG’s Ian Livingston notes today’s first freeze is the latest on record during a fall with a strong El Nino event (since we have reliable El Nino data, which dates back to 1950), but we’re not sure that signifies anything since the dates don’t vary a large amount (Oct. 20 to Nov. 18).

While Reagan National recorded its first freeze today, Dulles and BWI airports logged their first freezes way back on Oct. 18 and 19, compared to their averages of Oct. 15 and 29.

A web adaptation of a map showing the average first freeze date across the D.C. area and mid-Atlantic region. 162 locations were used to create the base map which has been subsequently further analyzed to try and present as accurate a picture as possible. We have smoothed out many of the isolated “dots” but have left some where there are multiple locations near one another which fit and as an indication that there is some variance in the shaded regions. (Ian Livingston, Justin Grieser and Katie Wheatley)