Update on water systems (part II)

I wrote progress report on our water systems at the beginning of January. This weekend, we had the first solid rain since then. That is over 70 days without significant rain. This weekend, I estimate we caught about 8 kL in our tanks, so the summer dry spell is officially over.

From reading the water meter, we used 8.5 kL from town water over summer to make up the shortfall in stored rainwater. I estimate that

1. I accidentally released 3 - 4 kL from the tanks into the garden in late January
2. The tanks at the front of the house were only connected in December, which meant we only collected 1/2 of the spring rains

Were the summer to be repeated, I think our water would last. But, can we expect future summers to be similar to the 2014/2015 summer?

This summer, versus historical summers

This summer was not unusually dry -- if anything, the rain that we had at the beginning of January was unusual and is unlikely to be repeated every year. Here's what the data say for Adelaide:


All numbers are mm of rainfall for that month. the statistics [1] were calculated over all the historical rain data I have for Adelaide (since 1884). For the purpose of this analysis I've assumed the data are stationary (ie. there's been no change in Adelaide's rainfall patterns in that time).

Assuming the rainfall distribution is normal (a wrong assumption, but probably doesn't matter for this), 35% of years will be within 1 standard deviation of the mean. Note that for the months November to March this means that 1/3 of those months can be expected to have basically no rain at all! (Note that I haven't looked at inter-dependencies in the data (ie. Does a dry February tend to follow a dry January (which it probably does)?)

For the summer just gone,  here are the numbers



Every month except January was drier than average, but none by much more than one standard deviation. In other words, it was a dry summer but it was not unusually dry.

Rainwater Model

Based on this summer's data, I've revised my rainwater model a bit. Mainly I've realised that although I have 37 kL storage, I probably only have 30 kL of usable storage. I've also increased the washing machine frequency to daily, from 0.8 times per day. Looking at the historical data, we would now have run out five years since 1884.

Lessons

I set up a dripper system on the garden, and believe I wasted a lot of water there. Next summer, I want to move to wicking beds for the annual crops.
I also want to have a functioning composting toilet, which will save more water.

[1] Standard deviation is a measure of how variable the data are. In this case, it is a measure of how likely it is that a given year is a long way from the mean (average) rainfall. For normally distributed data, about 1/3 of years will be within one standard deviation of the mean, and about 7% being more than 3 standard deviations away from it. More info on wikipedia.


-----------
This article was written by Angus Wallace, and first appeared at guesstimatedapproximations.blogspot.com.au