As the Earth moves around the Sun, the length of the day (defined as the time between sunrise and sunset) changes. The extent to which it changes depends on latitude, as shown in the graph below:

As you can see, the length of a day changes far more during the year at higher latitudes than at lower latitudes. (Latitudes beyond 66°33′ are not shown because the Sun does not always rise or set at these latitudes.) The graph runs from one winter solstice to the next, with the two equinoxes clearly visible in March and September.

It’s quite interesting to look at by how much the length of a day changes every day. This graph would have the same shape as the previous one, but not if we look at *percentage* change. In a way, this gives an impression of how quickly it appears that “the nights are drawing in”.

At higher latitudes the length of day changes quite noticeably in early January and mid-November. In some situations two adjacent days are different in length by nearly five minutes, and at some points the day loses nearly fourty minutes over the course of a single week.

OMG! I just got excited after seeing this post here. That sounds great to know statistically that what is the rate of change of day length with latitude. I’m thankful to you for making this informative post.

Thanks, always wanted to know this. But can you explain WHY the rate of change is faster either side of the winter solstice also the curious shape of the curve mid spring and mid autumn? Answer must be topological

I’m not sure I understand what you mean by “answer must be topological”?

Hi, I was finding your comment “It’s quite interesting to look at by how much the length of a day changes every day. This graph would have the same shape as the previous one” most confusing – because clearly when daylength is max/min, rate of change of daylength is close to zero – and at the equinoxes, rate of change is clearly max. In other words, the graph you allude to ought to be a graph of the first derivative, not an identical graph.

However, that equates to a sine/cosine situation ie: d/dx sin x = cos x – so it’s not the same graph but the same shape of curve moved along 1/4 year.

I think this could do with clearing up.