Selling Ice-Cream

This is about water, and water use efficiency, crucial for Australia's survival and sustainability.

About 10 years ago I met a wine grape grower near Mildura, on the Murray, who was irrigating his grapes by pumping water for 48 hours every 7 days. Then he started to measuring the moisture in his soil, and soon realised that after 2-3 hours the water was pouring out of the bottom, and was just being wasted. He changed his irrigation schedule to 2-3 hours every 3 days. He saved a lot of water, and grew better grapes. He even got a bonus because, by controlling the water, he was able to get the sugar content just right.

If you measure soil moisture, at different depths in the soil, you can find out...

This is how it works.

As a plant grows it sucks water out of the soil. If it sucks out 1 litre of water in a day it must be using 1 litre per day. So if we measure changes in soil moisture we can calculate how much water the plant is using.

A hole is drilled into the ground near to a tree or plant, and sensors that measure the water in the soil are installed at various depths, at 20 cm near the surface right down to about 1 metre or even deeper.

The soil moisture is then read and recorded on a computer, every day.

The change in total water in the ground, which we are measuring, is the daily water use of the plant.

The soil moisture is displayed on the computer screen as a time graph and a depth graph. Time graphs show the soil moisture varies over time, and depth graphs show how it varies with depth in the soil. The depth graph is very much like looking at a pit dug in the soil, and you are looking at the side of the pit. Only this is a lot less work.

A plant will use more water near the surface, and the soil moisture measured at the top of the soil profile will fall as the roots extract water. But lower down, below the root zone, it will not be using any water at all and there will be no change in soil moisture.

On the time graph you will see the top reading falling, but the bottom reading will be constant - a straight line horizontally across the graph.

On the depth graph you will see a vertical line rotating from right-to-left, like a wind-screen wiper in a car, fixed at the bottom and the top rotating to the left.

This fixed point, the depth that it is rotating about, is the maximum depth the plant is using water - the rooting depth.

When some water is applied - rain or irrigation - the soil moisture at the surface will increase. The depth graph will rotate across to the right, like a wind-screen wiper. And then, as the plant starts to use water it will slowly rotate back from right-to-left. And this goes on all season - a slow sweep to the left as the plant uses water, followed by a fast sweep back as irrigation water is applied.

If too much water is applied the depth graph will over-shoot, and go much further to the right. The bottom reading (which hasnít changed up to now) will start to increase and will also move to the right. (It is as though the wind-screen wiper has come adrift.) Water is flowing out of the bottom of the profile. After a few days the excess water will drain away and the bottom reading will return to normal.

This Ďnormalí - the maximum amount of water that the soil can hold before through-drainage takes place - is called the full point. Imagine putting a sponge into a bath of water, and then pulling it out and holding it up in the air. Water will pour out, but eventually the dripping will stop. The sponge is now at the full point.

If you forget to water, the depth graph line will rotate even further to the left, and eventually the bottom reading will also start to fall and move to the left. The plant roots are working harder, and are trying to extract water at a lower depth. The daily water use will probably reduce, with a corresponding drop in yield from the crop. This is when you should have irrigated - you have reached the refill point.

And thatís it - the actual crop water use, the full point and the refill point. From the daily water use the computer works out when the soil will reach the refill point (the date when you should be applying more water), and you should only apply just enough water to bring the soil back to the full point. The exact amount of water this plant needs, with no wastage.

I was at a citrus farm in South Africa when the farmer saw his first soil moisture measurements taken on his farm. He spent a few minutes in front of the computer screen working out what all these graphs meant, and then he put his arm round his foreman and said Ďitís magic, itís like an x-ray of the roots, for the first time we can see whatís going on down thereí. {By this time they were both kneeling in front of the computer, but that was going too far}.

Improve Quality and Yield

As a bonus, once you know just what is going on down there, you can also improve the yield and quality. Good wines are made from grapes that have been water stressed at certain critical times in the year; and sugar cane production increases dramatically if you can keep the moisture in the soil just below the full point.

Measure, or rely on a model

Australia is probably the leading country in the world with these techniques, and has several manufacturers of soil moisture sensors that are sold throughout the world. Essentially we are exporting our experience of how to use water efficiently.

But very few irrigators in Australia are like the grape grower on the Murray - most do not measure.

They will have been told by their lecturers, or Department of Agriculture extension officers, to calculate how much water their plants need. The temperature and humidity recorded from a weather station will be used to work out how much water a standard crop should be using. This is then adjusted for their crop using factors from a text book or manual. If you use this method every grape vine in your valley is presumed to be using the same amount of water each day.

We know this isnít true, so why is this method used ?

Selling ice-cream

A friend of mine decided to retire and open an ice-cream shop at Manly (structural adjustment, redundancy package.)

But they didnít just rush in - first they did a lot of research on the internet to find out how many ice-creams they were likely to sell. There are a lot of MBAís on the economics of ice-cream, mainly in California. They found that, at $2 per cone, they could expect Ďbase-lineí sales of 1,400 cones per day. But because the vacant shop they were looking at was about 50m back from the beach, they could expect only 68% of the base sales. On the plus side they could expect increased sales as it got hotter - for every degree above 20 degrees you could expect an increase of 7%. A temperature of 25 degrees should give 35% higher sales.

They wrote a business plan based on the research, with graphs and diagrams and a nice plastic binder. The bank manager said Ďyou beauty, lots of diagrams and graphs, how much money do you want ?í

Well, it all went incredibly well. They got the shop, fitted it out, sold lots of ice-cream, well worth being re-structured.

One hot day in mid-summer, with a temperature of 32 degrees, they needed to work out what the sales had been that day so they could order more ice cream.

How should they do it ?

Base sales are 1,400, but they are 50 m back from the beach so only 68% of base; but they are 12 degrees hotter than 20 degrees, at 7% per degree.

How would you work out how much ice-cream they had sold ? Would you...

Calculate 1,400 * 0.68 * 1.72


Just look in the till ?

The calculation would tell them how many they should have sold, if the model had been 100% right

Looking in the till would tell them how many they had actually sold.

The MBAís might well be interested in the daily takings to refine their models, but if you want to know what is really happening surely itís best to just look in the till.

(Note A reader has pointed out, quite correctly, that you should really be looking in the freezer to see how much stock you had left - not the till. The auditers might be interested in any discrepancy between the stock and the till - but that is another matter.)

Back to Irrigation

If this had been an irrigator, growing apples or olives or grapes, they would have been taught to use the calculation method. The temperature would be from a local weather station, and various correction factors could be local or from research and trials done in other countries.

But, why are they told to use the calculation method ? By Ďlooking in the tillí, measuring moisture in their soil, they would know exactly how much water they have and how much more needs applying.

I think it is tied up with teaching, and exams, and the warm inner glow you get from looking up a number in a book. Academics and researchers do trials and take measurements, and growers should then do what their nice mathematical model says.

If you are a researcher and told...

We have a major crisis with water in Australia, how much money do you need to tell us how we can save water

I am sure you would reply...

We need to do some trials, another research project, and we will have the answer

Thatís what I would say if I was a researcher.

But if growers are shown how to measure and interpret their own soil water they can find out exactly what is happening to their crop, in their soil. They will know the efficiency of each and every irrigation, and at the same time can improve quality and yield.

Another example: Apples in California.

Just inland from Santa Maria in California is the Cuyama Valley, where most of the carrots for the USA are grown. But there also some apple growers, and they wanted to know how much water they should be applying to their apples.

The nearest weather station was 30km away, and a compensation factor was needed to allow for the fact that their weather was slightly different. A research team from the Department of Agriculture measured the actual water use at about 6-8 different trees, and concluded that the Cuyana Valley apple growers should apply 4% less water than normal. Good stuff, a warm glow of practical research, and they reported an excellent r^2 correlation of 0.96.

However, the published paper also had the raw data from each of the trees they had measured. The 4% compensation factor had been calculated from the average of all the trees (the r^2 had also been from the average). And there was a wide scatter in the individual trees. Some trees were actually using 50% more water than the average, and some 50% less. So although the grower was being told to apply 96% of, say, 6 mm/day, what was really happening was that some of their apples were actually using 9 mm/day and some 3 mm/day !

But the final straw was that far too much water had actually been applied to the apples. From the raw data it was quite clear that at least two of the sites were being grossly over-watered - gallons of water were just being wasted. Even though the researchers had been there all summer they had not told the grower, and it wasnít even mentioned in the paper.

Back to Mildura

Back to the original grape grower near Mildura, who 10 years ago changed his irrigations from 48 hours every 7 days to 3 hours every 3-4 days, saved a great deal of water while improving his yields and quality.

I met him again recently at a seminar - for the last 10 years we have been using his first seasons data as a demonstration of what can be done if you measure soil moisture.

I thanked him in front of the whole group, and said I assumed that all his neighbours had benefited from his work and were now also irrigating more often, but with less water each time - they do have a major salinity problem down there.

He said that they were still pumping for 48 hours every 7 days.

It seems they get big discounts for using electricity at weekends.

June 2002 - no diagrams because it was intended as a script for radio

Trevor Finch
Research Services New England
11/16 Nicholson Street
Balmain, NSW 2041, Australia
t: 9810 3563 f: 9810 3323
e: <>