Manilla had great droughts in 1902, 1912, 1940, 1946, 1965, and 2018.
A new graph
This new graph records all historic periods of severe and extreme rainfall shortage at Manilla NSW. Data show 25 values of duration from 2 months to 360 months. Unlike the graphs of rainfall shortage in earlier posts, this graph shows the months of onset, persistence, and breaking of each occurrence. [See the Note below: “An innovative graph”.]
Extreme shortages (up to the 1st percentile) are shown in red and severe shortages (up to the 5th percentile) are shown in grey.
The dashed lines labelled “First Good Data” and “Last Good Data” are limitations that apply to all cumulative rainfall deficiency data. [See the note below: “First Good Data; Last Good Data”.]
A more detailed graph for the years 2000 to 2019 is shown in another post.
“Rainfall Shortages” or “Droughts”?
Although droughts involve other factors, rainfall shortage is decisive. Some other factors, such as daily maximum temperature anomaly, vary in the same sense, Australia-wide.
Recognised times of drought at Manilla appear on this graph as extended periods of severe or extreme rainfall shortage .
To simplify, I have separated the graph into three parts according to the duration of the shortages: seasonal droughts, great droughts, and decadal droughts.
Droughts with extreme shortages of rainfall that lasted up to nine months happened often. Counting those with more than one red marker, there were 16 seasonal droughts in 136 years: about one in each nine years. Gaps as short as four years happened around 1920, while the longest gap of 16 years came between 2002 and 2018.
The most common season of extreme rainfall shortage was winter, with 8 cases. Autumn had 5 cases and spring 3 cases, while summer had none.
Some of these seasonal droughts became Great Droughts.
The Six Great Droughts: 1-year to 9-year duration.
According to this graph, Manilla has suffered six great droughts, attested by extreme shortages of rainfall at more than 2 of the 13 chosen duration values from 12 months to 108 months. They were:
(1.) 1902 (the Federation Drought);
(3.) 1940 (first part of the World War II Drought);
(4.) 1946 (second part of the World War II Drought);
[The absence of the Millennium Drought is discussed in the Note below.]
BoM continental rainfall maps
This graph of one-year to nine year shortages at Manilla can be compared with a set of maps by the Bureau of Meteorology. Those maps show rainfall deciles for each calendar year from 1900.
Decadal extreme rainfall shortages
Extreme rainfall shortages with durations of a decade or more were strictly confined to the period from 1910 to 1950, during which time some were continuous. They coincide with a time of observed low mean rainfall at Manilla (and elsewhere).
After 1950, decadal extreme rainfall shortages did not occur at all for at least forty years, but they may recur. The “Last Good Data” does not yet say whether decadal extreme shortages are now current.
First Good Data; Last Good Data
Dashed lines on these graphs, labelled “First Good Data” and “Last Good Data” show the limits of rainfall shortage values for which data are complete. To the right of the “Last Good Data” line, values are provisional. They are minimum values of severity of shortage, because more extreme values may appear there when future observations are included. To the left of the “First Good Data” line, it may be that severe or extreme rainfall shortages were then current, due to low rainfall in months before 1884 when there were no observations.
Note: An innovative graph
In posts in this blog, the line graphs and contour graphs that show rainfall shortages have a limitation that may mislead. Although they can show the end of a rainfall shortage (drought breaking) they never show the beginning of the shortage (drought onset).
On those earlier-designed graphs, shortages observed in a given month are plotted at that month. Yet, the shortages were calculated for durations that extended over several prior months. An observed 6-month shortage, for example, must have begun 5 months before the month when it was observed.
In this new graph, shortages are plotted to show their extension to dates earlier than the date of observation, up to the duration that is specified. As a result, this graph shows both the beginning and the end of each plotted episode of rainall shortage.
In a recent post with a line graph of rainfall shortage, I gave details of spreadsheet calculations in the note “Data analysis”. Tables on the spreadsheets had serial months in the rows and duration values (in months) in the columns. Data values in the tables included cumulative rainfall totals, percentile values, and classes of shortage.
For this innovative graph, I prepared four more tables, two for severe shortage and two for extreme shortage. For each class of shortage, the first table showed the occurrence of that class of shortage by a number (n) equal to the duration concerned, leaving blank all cells without that class of shortage.
The second table built on the values of the first. I used an “IF” function to determine the value to enter. In the data column for n months, the value n was entered in each cell for which the value n had appeared either in that cell or in any cell in the (n-1) cells directly below it. Other cells were assigned the value “#N/A” as returned by the function NA() so as not to appear as zeros that would be rejected by a chart with a logarithmic axis.
[This second table was devised by Andrew White.]
Comparison with a earlier graph
An earlier graph that resembles this one is copied here from the 2014 post “Manilla’s record of droughts”. In the earlier graph, one has the impression that the longer the duration of a shortage, the later it occurred. The new graph reveals that this was seldom the case.
Note: The Millennium Drought absent
The “Millennium Drought” was not a Great Drought at any place near Manilla. A map by the Bureau of Meteorology shows that one quarter of NSW was not involved in that drought. In the 84 months from November 2001, rainfall was not in “serious deficiency” (below the 10th percentile) in any part of NSW north-east of a line from Wollongong to Brewarrina.