ABSTRACT
1 Records covering many years can provide information additional to the probability of each of a number of alternative events occurring, such as b 1 and b 2. For instance, in addition to discovering that over, say, 100 years event b 1 (flood) occurred in 60 out of 100 years, that is, with a frequency of 3/5, and event b 2 (no flood) therefore with a frequency of 2/5, the records may reveal the following information: (i) prior to event b 1, a period of several weeks of cloudy weather – a condition we refer to as z 1 – was observed in half the number of b 1 events; (ii) prior to event b 1, a period of several weeks of mixed weather – referred to as z 2 – was observed in one-third of the number of b 1 events; (iii) prior to event b 1, a period of several weeks of clear weather – say z 3 – was observed in one-sixth of the number of b 1 events.
