AN OBSERVATION OF THE CONFLUENCE OF STATISTICAL ABERRATIONS, DAM RATIOS, TIMING AND UNCOMMON RAINFALL EVENTS THAT COMBINED TO HAVE OUR LEADERS DRAW INCORRECT CONCLUSIONS. THEIR INFLUENCE ON ACTIONS TAKEN.

Author : J. V. Hodgkinson F. C. A. Chartered Accountant : August 2006 to May 2010    

The relationship of the Wivenhoe/Somerset dam system to its rainfall in the catchments has been quite different from the way it has been presented to us since the joint dams’ inception in 1986. The main rainfall supplying our water is the, season ignoring, rain depressions in the form of Cyclones, Monsoons and large scale rain depressions. They generally cover the whole of South East Queensland at the same time. They provide much more water than our December to March “wet season” which is well known to us.

Mr Rob Drury of SEQWater drew our attention to this in his Courier Mail article of the 10th February 2007. Output from the Government IQQM computer model, covering the 111 years 1890 to 2000, has proven Mr Drury correct. In essence 300mm of rain in the catchments in a few days is a flood and, in the right circumstances, can fill the dams from scratch. On the other hand 3 months of 100mm a month is a comparative trickle. These facts can be buried in official overall rainfall statistics the use of which in dam performance can and did throw up statistical aberrations.

Mr Drury called these rainfall events “uncommon events”. Bureau of Meteorology flood records from 1841 reveal that they occur on average every 3.7 years with most below that average. Therefore those above can extend well above that average such as the recent period of 2001 to 2007.

Mr Drury also provided dam level charts which were not published by the Courier Mail. They provided clear evidence of the importance of "uncommon events" and the inability of the "Summer wet season" to provide sufficient water from as far back as 1992. We will examine them.

This is my review based on official statistics and documents. These statistics and documents are presented to you so that you can arrive at your own conclusions.

************************************************************

What happened in our catchments to deplete our dams?

A consolidation of information contained in this website. It can be used as a companion to “Borumba dam expansion alternative”.

****************************************** 

Information confirming that no drought existed in the catchments of the Wivenhoe/Somerset Dams for the period February 2001 to 2007. The depletion was caused by the random nature of our main water supply “large scale rainfall events” such as Monsoons, Cyclones and large scale rain depressions. They have no relation to seasonal conditions and can happen at any time generally within a 4 year average. The Courier Mail article of the 10^th February 2007 sets out the rainfall requirements.  
Photo : Courier mail article explaining our water sources

The result of this determination is that the solution to cover the period of depletion in the dams is diametrically opposed to the solution required for a “drought”. The "drought" proposition is still retained in the "South East Queensland Water Supply Strategy" and is an integral part of that strategy. While the words are well chosen and well intentioned, the public assessment was that it "does nothing" at least for the present. The Strategy ignores the short history of our water supply incorporating the addition of the Wivenhoe Dam (1986) and its potentially severe damaging consequences dealing with the recent depletion as a "drought".

The evidence

1. Initial enquiry resulting in this email from the Bureau of Meteorology confirming that rainfall in the catchments for the three year period 01/08/03 to 31/07/06 was close to 80% of the long term average. It was compatible with being in “decile 1” being the lowest on record.

This means that Rainfall in three year periods had never been less than close to 80% of the long term average 1961 to 1990.  
Photo : Email from the bureau of meteorology

Decile maps are susceptible to statistical aberrations. For example, 10 rainfall measurements from 991mm to 999mm would place 991mm in the lowest on record decile 1 and 999mm in the highest on record decile 10. The Bureau offers no warning on the maps that this could be the case.

2. A copy of the six years 2000 to 2006 “decile map” promoted to every household in South East Queensland (SEQ) is attached. It is now accompanied by the corresponding “percentage” map that I had prepared by the Bureau of Meteorology for the same period and the same rainfall. It reads 80% of the long term average.

Photos : Decile graph showing lowest on record ; Percentage graph for the same water and the same period

It is now clear that the “decile” map is a statistical aberration and was used to convince the residents of SEQ that our depleted dams were the victims of a severe drought in the catchments. Most people of SEQ would have held a different view if the “percentage” map was the one published.

I was not satisfied and I required further verification. I retrieved from the Bureau of Meteorology all the data from all relevant rainfall stations in the catchments, some dating back to the 19^th Century with the following result:-

3. Schedule showing the distribution of rainfall for all operating rainfall stations in the catchments since records were kept. It clearly shows that the 4 summer months December to March receive slightly in excess of 50% of the annual rainfall. The other 8 months receive the balance but, being spread out over the 8 months, does not have the same ability to create inflow into the dams.

Rainfall stations in the catchments showing the difference between summer and non-summer rainfall

4. Summaries of the Bureau rainfall data for the Wivenhoe and Somerset catchments for the period incorporating most of the long term average 1961 to 1990 and the period to 2006. Years 2001 to 2006 are the period of the “worst drought in 100 years”

They show that the summer rainfall was quite normal in the catchments with 99.7% in the Wivenhoe and 91.3% in the Somerset . The 20% deficiency was in the non-summer months that normally create little inflow.

Photo : Wivenhoe and Somerset catchments rainfall showing the summer rainfall for the period of "drought" quite normal.

5. Population increase is worth noting. The attached schedule drawn from Government websites shows that the population was predictable and was as predicted. There was no sudden uplift.  

6. Attached is the Queensland Climate Change Centre of Excellence (QCCCE) report “The Drought to 2007”. It confirms that 76.2% of the long term average was the lowest on record.

Unfortunately the QCCCE did not have hydrological records of the Federation drought, 1898 to 1903, with which the comparison of the “worst drought on record” was made. The comparison was made solely on Rainfall.

Photo : QCCCE observations of the "drought". Confirming the rainfall at 76.2% of long term average.

Those rainfall records are attached. They clearly show that the “Federation drought” did not conform to the usual 50% of rainfall in the summer months. The rainfall in the summer months was down to 30% thus spreading the rainfall over the whole twelve month periods. That they created little inflow is clearly evident from the attached graphs.

Photos : Comparison of Federation drought rainfall with the "drought" period 2001 to 2006. Evidence that the current period quite normal and the Federation drought abnormal.

Not all of Queensland was involved in the Federation drought. North Queensland was battered by two cyclones in that period. The report of the QCCCE adds weight to the Bureau of Meteorology's opinion that district 40 in which the catchments reside is a stable rainfall area.

Photos : 1899 Cyclone and the 1900 cyclone in North Queensland. 112 degree Fahrenheit in Adelaide and 104 degrees in Melbourne

7. The depletion of our dams was caused by the random nature of our main water supply “Large scale rainfall events” that are Monsoons, Cyclones and large scale rain depressions. They are known by SEQWater as “uncommon events”. It is a misnomer as they occur on average every 3.7 years.

They are our main water supply as the attached schedule of “years to fill” indicates. They can be hidden in overall rainfall statistics as 300mm over a few days is a flood whereas 3 months of 100mm each is a trickle but the same measurement of rainfall.

Photo : Years to fill the dams without withdrawals for all our water sources.

The next two graphs are conclusive proof that it was the random nature of our main water supply that caused the depletion of our dams. It is worth printing them out and comparing the Wivenhoe experience in both graphs.

Their random nature is best illustrated by this graph drawn up from Bureau of Meteorology information and labelled “Frequency of large scale events”. It lists the 43 events since 1841. The pink shows the average of 4 years and the blue illustrates those in excess of the average.

It is quite clear that the last 6 year period was predictable and was not unusual. There were 7 periods in excess of the 6 year period and it is the blue periods to which our water supply security should be directed.

When read in conjunction with the attached dam level graph of the Wivenhoe the cause of our depleted dams comes into stark perspective. It was not a drought but the random nature of our main water supply “large scale rainfall events”.

8. The underlying problem is that our dams are too small to control these events for the benefit of our ecology and the residents of South East Queensland. The answer is laid out in the accompanying “Borumba dam expansion alternative” on the HOME page.

J. V. Hodgkinson   F.C.A.
December 2009