Are we in the middle of an extreme, unusual climate event that caught authorities napping, or is this purely a political machination of some diabolical state-capturers? There have been a lot of social media posts stating that contrary to what some scientists have said, we are not facing an extreme climate event. I came across the results of some analysis done by Anton Sparks, a professional engineer at Aurecon, which support the statements that the last 3 years are in fact very likely to be significantly worse than a 1 in 100 year event.

Comparing the combined inflows over the past three years (Preliminarily estimated from the equation “Monthly Inflow” = “Monthly Increase in Storage” + “Monthly Demands on Dams” which can later be refined using evaporation and rainfall on the dam’s surfaces when the data is available) into the 6 dams that feed Cape Town with potable water (note we’re not looking at rainfall, but the inflows into the storage systems that are the main reservoirs we use) and with similar data for the period from 1928 to 2004 (Derived by DWS as part of the study to assess the availability of water in the Berg River (see http://www.dwa.gov.za/Documents/Other/WMA/19/Reports/BergAssess23Jul10.htm), one can note that:

- The annual inflows from Nov 2016 – Oct 2017 was about 100 million cubic metres (1 cubic metre of water = 1 kilolitre) LOWER than the LOWEST of any 1 year period during 1928-2004.
- The total flows over the TWO consecutive years from Nov 2015 – Oct 2017 were 80 million m
^{3}LOWER than the LOWEST two consecutive year period during 1928-2004. - The total flows over the THREE consecutive years from Nov 2014 – Oct 2017 were 180 million m
^{3}LOWER than the LOWEST three consecutive year period during 1928-2004.

This can be illustrated graphically using a “Percent Exceedance” graph (below). To produce a percent exceedance graph for say the annual inflows all that is necessary is to rank and plot the annual inflows in descending order, so that the largest annual flow occurs on the left hand side and the smallest annual flow occurs on the right hand side. None of the flows are greater than the leftmost flow so it has a percent exceedance probability of 0%, while all the flows are greater than the rightmost flow which has a percent exceedance probability of 100%. Similarly, the middle flow will have an exceedance probability of 50%. This graph enables one to visually view the distribution of flows and say to pick out the annual flow which is exceeded say 10% of the time, if that was of interest.

Looking at the first graph then, the far left of the red line shows that the very wettest 1 year period had inflows of about 1300 million m^{3}/a (at 0 percent exceedance) and the driest inflow was 380 million m^{3}/a (at 100 percent exceedance). The red dot representing the inflow from November 2016 to October 2017 only has inflows of 280 million m^{3}/a, 100 million m3/a less than any record in from 1928 to 2004. Expanding the sample to 2 year and 3 year consecutive periods shows provides data for two and three year long droughts.

What is very interesting is to then look at the values for the periods 2016, 2015 to 2016, & 2014 to 2016. In all of these cases, there is NOT ONE equivalent period from 1928 to 2004 where the inflows into the 6 CoCT supply dams are lower.

Aurecon then simulated the probabilistic drawdowns on the dams from November 2014, using a stochastic modeling technique developed by Professor Geoff Pegram of UKZN (http://sahg.ukzn.ac.za/people/) and the hydrology data available up until September 2005 and the actual system water demands from November 2014. The 2nd graph shows the results of the simulations, with the ACTUAL dam levels superimposed as black dots. The light green line shows how things would have looked if we’d had average inflows to the dams and average drawdowns – everything just fine, thanks very much. The dark green line shows a 95% exceedance probability (or 1 in 20) risk of lower storage volumes. Lines for 98%, 99% and 99.75% exceedance probability are shown. In October 2015, the dam levels were above the 1:20 risk level, in October 2016 at the 1:20 risk level and in October 2017 below the 1:400 risk level.

Again, Aurecon modelled the probability of different storages going forward from November 2017 which are shown in the third graph. The graphs show what will happen if we cut our usage, in this case Urban use by 45% and Agricultural use by 60%, both with respect to the average consumption from 2011 to 2015. The “spreadsheet simulation” is based on the worst case annual inflows from November 2016 to October 2017. If we cut our usage sufficiently, we could make it through and remain above the “Day Zero” storage level when one would have to queue for water. If we don’t, and the augmentation projects are delayed, or like Beaufort West, do not produce what is predicted, we’re in big trouble.

The red dots track the current storage. Note that despite the light rains experienced in early summer which have helped to elevate the storages slightly, the storage below the projected storages because the targets are not being achieved.

We have no choice, we HAVE to ensure the City as a whole uses less water. Given that the severity of the current drought has deepened because of the last winter we should try to reduce our consumption to 50 litres per person per day or less if at all possible. This means even less/shorter showers or sharing baths, and utilizing grey water to flush!! 50 litres per person per day is still twice the “Day Zero” allowance.

For drought information and tips and advice on water saving methods consider joining the Water Shedding Western Cape Facebook group :

**www.facebook.com/groups/watersheddingsa**

My thanks to Anton Sparks of Aurecon for patiently taking me through the graphs and answering questions about them.

Dave Gale

Water Shedding Western Cape

2018-01-15