The annual pattern of rain day rainfall

In Manilla, the mean pattern of rainfall on rain days through the months of the year is simple and regular. This pattern can be worked out from the 125-year rainfall record of Manilla Post Office, Station 055031, beginning in 1883.
The graph above shows that, on the average, on a day when rain falls in January, the total in the day is about thirteen millimetres. When rain falls in July, the total in a day is about half of that: that is, six and a half millimetres. The pattern through the year is close to a perfect harmonic cycle, with a maximum in the third week of January, (four weeks after the longest day) and a minimum exactly six months later, in the third week of July. Only two of the monthly readings do not match the pattern well: January has about one millimetre more than would fit the curve, and December about half a millimetre less.
Of course, most people in the district realise that heavier rain falls in summer, but few would know any details. I do not think that the Bureau of Meteorology has ever worked through these figures. [See note below about the use of “rain days” in the Bureau.]
This very simple pattern of mean rainfall per rain day is the more remarkable because it comes from two other patterns that are not so simple.

The second graph is the pattern of monthly rainfall totals through the year. Manilla has two peaks of rainfall volume in the year. The major peak comes in the last days of December, a few days after the longest day, and a minor peak just six months later, at the end of June. Winter is marked, not by a minimum of rainfall, but by a secondary maximum. Much more detail is given in the post “A seasonal rainfall model for Manilla”
and in the post “Manilla 30-year Monthly Rainfall Anomalies”.

The final graph shows simply how many days of rain there are in each calendar month, on the average. This pattern is quite strange. Most months of the year have about six rain days. April has fewer: about four only. All three autumn months (March, April, and May) have less rain days than other months. I have drawn the graph to show that the months of winter, spring, and summer average 6.0, 6.2, and 6.3 rain days per month, while the months of autumn average only 4.6.

Note.
The Bureau of Meteorology does not seem to tabulate numbers of rain days now. There is a definition of “rain day”, under the Frequently Asked Question “What is a rain day?” on their web-site.

A rain day is recorded when there has been a daily rainfall total of at least 0.2 mm. For some recent stations the minimum is 0.1 mm. A rain day does not occur when there is only a trace of moisture in the rain gauge, or when the precipitation was observed to be solely from frost, dew or fog.
A rainfall total of 0.2 mm is quite a small amount of rain, and unlikely to have much impact on many activities. Therefore, days of rain greater than or equal to 1 mm, 10 mm, and 25 mm are often used as indicators of the number of “wet” days.

This statement misses a vital point: it is not the minute amounts of rain involved that matters, it is which days should be counted as rain days when comparing times and places. An interaction occurs when data are taken from tipping-bucket automatic rain gauges with 0.2 mm buckets. When there is little evaporation, a bucket will eventually tip by accumulating dew and minute traces of rain. Then a rain day will be recorded. There is no justification for calling this dry day a “rain day”. That is a defect in the system.
This anomaly would not occur if the definition were not “at least 0.2 mm” but “more than 0.2 mm”. Since the old criterion was “one hundredth of an inch” (0.254 mm) the difference in actual amount is minute.
I have faced this problem of “rain day” definition since the rain gauge at Station 055031 (now Station 055312) became an automatic tipping bucket (0.2 mm) gauge. In 88 days, 34 days had rain recorded but, of these, 8 showed “0.2 mm”. On these eight days, I did not see or record any rain myself, and I am sure that others did not either. Counting these days with 0.2 mm readings as “rain days” makes the monthly total of rain days jump to well over the long-term average number from manual readings.
In counting rain days using this tipping-bucket record, I intend to count only days with more than 0.2 mm recorded.
I have not had a statement from the Bureau of Meteorology either confirming or denying that the concept of a “rain day” has fallen out of use in the Bureau.