Standard Time, Daylight Saving Time and Local Time

The earth is round like a ball. When it's night in the US, it's morning in China. To do business beyond a local region, people need a common reference to communicate time. Standard time is the synchronization of clocks in different geographical locations within a time zone to a common time standard, usually based on the meridian at the center of the time zone.

Daylight saving time (DST) typically involves advancing clocks by an hour near the start of spring and adjusting back in fall. The primary purpose of daylight saving time is to adjust human activity schedule during day time to take advantage of the daylight hours during summer time. Not every country has adopted DST. In U.S., Arizona and Hawaii are not observing DST. In the regions where DST is adopted, the term standard time typically refers to the time without the offset for daylight saving time.

Local time means the time on the clock. In the regions not observing DST, local time is the same as standard time. In the regions observing DST, local time is standard time in winter and fall, daylight saving time in spring and summer. During the day when DST starts, there are 23 hours in local time. During the day when DST ends, there are 25 hours in local time.

In load forecasting, the load and weather data often comes from different sources. The misunderstanding of these terms often leads to data processing errors in load forecasting. Typical formats I've seen in load forecasting projects include 1) Greenwich Mean Time (GMT); 2) standard time; and 3) local time. While both GMT and standard time are observing 24-hour days, there are several variations of local time in load forecasting practice. During the first day of DST, some data sources observe a 23-hour day, while some use a 24-hour day with the 2nd (or 3rd) hour as zero or missing. During the last day of DST, some data sources observe a 25-hour day, while some observe a 24-hour day with the 2nd (or 3rd) hour as the sum of the two original readings of the same hour.

Regardless what format the raw data has, I usually convert the load and weather data to a 24-hour local time. At the beginning of the DST, I fill in the "missing" hour as the average between its adjacent two hours. At the end of DST, I average the original two readings in the same hour. For the convenience of the readers of my blog, below is a list of DST start and end days in the U.S. from 1987 to 2020.

Year	DST Start	DST End
1987	4/5/1987	10/25/1987
1988	4/3/1988	10/30/1988
1989	4/2/1989	10/29/1989
1990	4/1/1990	10/28/1990
1991	4/7/1991	10/27/1991
1992	4/5/1992	10/25/1992
1993	4/4/1993	10/31/1993
1994	4/3/1994	10/30/1994
1995	4/2/1995	10/29/1995
1996	4/7/1996	10/27/1996
1997	4/6/1997	10/26/1997
1998	4/5/1998	10/25/1998
1999	4/4/1999	10/31/1999
2000	4/2/2000	10/29/2000
2001	4/1/2001	10/28/2001
2002	4/7/2002	10/27/2002
2003	4/6/2003	10/26/2003
2004	4/4/2004	10/31/2004
2005	4/3/2005	10/30/2005
2006	4/2/2006	10/29/2006
2007	3/11/2007	11/4/2007
2008	3/9/2008	11/2/2008
2009	3/8/2009	11/1/2009
2010	3/14/2010	11/7/2010
2011	3/13/2011	11/6/2011
2012	3/11/2012	11/4/2012
2013	3/10/2013	11/3/2013
2014	3/9/2014	11/2/2014
2015	3/8/2015	11/1/2015
2016	3/13/2016	11/6/2016
2017	3/12/2017	11/5/2017
2018	3/11/2018	11/4/2018
2019	3/10/2019	11/3/2019
2020	3/8/2020	11/1/2020

 

Back to Load Forecasting Terminology.