Log of Very Wet Days at Manilla.

Graphical log of days with over 50mm rain

In the 130-year record of very wet days at Manilla, NSW, extreme rainfalls have not become more common recently.

Data

I arranged all daily rainfall readings for Manilla, NSW, from March 1883 to December 2014 in order of rainfall amount, and selected only the 125 readings greater than 50 mm. I plotted the values against the date, expressed in years, to two decimal places. (See Note below.)

Result

The five highest readings

The five highest readings, greater than 110 mm per day, include events that gave rise to two floods and the filling of a reservoir newly-built to store water for irrigation. The highest daily reading, 142.7 mm, came with the highest flood known at Manilla, in 14/01/1964. Thus, the highest flood matches the highest daily rainfall. That is because nearly all the flood-water came down the Manilla River, which flows in a semi-circle, with none of the catchment area far away from the rain-gauge.
These five highest readings seem to fly in an arc above the rest, with a peak near the middle of the graph. The rise and fall of this arc may have no meaning, for there are very long gaps between the events. All the same, it is a fact that there were no readings above 110 mm per day in the decades before 1910 or after 1998.

Periods with no daily readings over 80 mm

There are periods when the maximum daily rainfall observed is higher or lower than usual. I have chosen to mark those periods of a decade or more when no day had rainfall above 80 mm.
There are four such periods:

  • 1883 to 1895 (12 years);
  • 1910 to 1928 (18 years);
  • 1937 to 1954 (17 years);
  • 1999 to 2015 (16 years).

They have occurred rather regularly: with one more in a gap, there could have been a 30-year cycle.

Trend lines

Fitted trend lines explain practically none of the variation.

Twenty-first century deluges?

We are warned to expect a trend to more extremes of rainfall with predicted climate change. Australia is said to be particularly vulnerable. This data set shows that, for daily rainfall deluges, one place in Australia shows no such trend. For more than a decade, the heaviest daily rainfalls have been below normal, as they were at intervals earlier in the record. In fact, the last decade of the record resembles the first, more than a century earlier.

Later work

For further analysis, with incidence of flooding rains by decade, see here.

That led to further speculative work relating days of heavy rainfall to the observed frequency of drought.


Note.

It is not a simple matter to display events of known date on a scale of years extending back to 1883. For dates from 1900, I have converted the date to day number, divided by 365.24, and added 1900. For earlier dates, I have manually entered the year number, and added a decimal equal to the day number in the year, divided by 365.24.

Advertisements

2 thoughts on “Log of Very Wet Days at Manilla.

  1. Hi xmetman. Thanks for your suggestion.
    I have called up “Insert Julian dates” in “Microsoft Excel Help”. I then selected “Convert to a Julian date used in astronomy”. That gives a formula, but it says: “This formula only works for dates after 3/1/1901, and in workbooks using the 1900 date system.” So I can’t use that for dates going back to 1883.
    As I see it, because they use the 1900 date system, I can’t use the types of graphs in “Excel Charts” that are intended to have time on the x-axis: the Column, Bar, and Line charts. Instead, I use the “X-Y Scatter” type of chart. Then I have to hash up years and decimal parts of years as I said in my note.
    Do you think there is another way?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s