The time of the phase for 17 August is 15:54. This should mean the things that happen on this spell are the same as those that happened on previous similar spells.
Unfortunately life's not like that. …
The time of the phase for 17 August is 15:54. That is very roughly similar to the times of the following phases earlier this year; the three underlined dates being the closest:
31 Jan 04:10 <
7 Feb 21:54 <
21 Feb 22:35
8 Mar 09:40
29 Apr 09:58 <
12 May 21:47
11 Jun 10:41
19 Jul 04:24
2 Aug 03:27
17 Aug 15:54
Let me try and explain the hours of a spell:
15:54 (the actual time of the next lunar phase) = (more or less) 16:00.
16:00 = 4, 10, 16 and 22 "hundred hours". (Both four o'clocks and both 10 o'clocks in 24-hour~ese.)
On the BOM Antarctic chart a few days into this spell there is what looks to be a tornadic spell following a serious earthquake:
Look at the black disc on the bottom of the continent (60 to 80 degrees W opposite the tip of South America.)
1. It's a forecast and thus can go very wrong.
2. There is a constant between cyclone intensities and earthquake magnitudes.
3. Severe tornado spells often replace expected earthquakes.
Large enough tropical storms and great enough magnitude earthquakes and extensive enough tornado cells can and will upset the routine of the time of the lunar phase.
This then, makes using the time of the phase as a means of forecasting the weather useless. But god has been kind to us. There is a clear signal available on the North Atlantic chart that indicates when things are about to go pear shaped.
It's called the Greenland High: