FENNER CONFERENCE ON THE ENVIRONMENT

The Water Resource Observation Network (WRON)
Dr Stuart Minchin

Stuart Minchin joined the CSIRO Division of Land and Water in 2005 as leader of the Environmental Sensing, Prediction and Reporting (ESPR) research theme. He has a strong interest in all areas of environmental science, particularly in water resource management. His recent employment as Principal Scientist for Water Assessment and Research, and previously as Manager for Water Resource Monitoring, involved the direction of research priorities as well as the specification and management of large-scale natural resource management programs. These were often conducted by third parties in both the private and public sector. He has an interest in the management and modelling of environmental data, as well as the potential for online delivery of data, modelling and reporting tools for improved natural resource management.

There was a fair bit of discussion here yesterday about the Prime Minister's plan for water security. One of the lesser known parts of that plan, or one that has not gained so much media attention, is the focus on water information, and that is what I would like to talk about today.

Water information is a bit of a basket case in Australia at the moment. We have, at last count during the recent scoping study for water accounting, identified over 600 different organisations that collect water information across Australia. They all do it differently, they manage the data differently, and they make it available in many cases poorly or not at all.

On top of that, with 600 different organisations and that includes private industry, councils, water authorities, state governments, the federal government many of them are still in the Dark Ages of IT and I think this is partly due to the fact that most of the people managing those organisations grew up when an information system was a wheelbarrow full of punch cards.

So what I would like to talk to you about today is how we can fundamentally change the way we manage water information, and some of the technical advances that are giving us the opportunity to really amp up water information availability across Australia.

I am going to talk about the Water Resources Observation Network (WRON). I am going to discuss why we need it, what the vision is, and then, to live up to my previous statement, I am going to do a live internet demonstration for you. So watch me crash and burn!

(Click on image for a larger version)

Why do we need a national water information system? Recently, things have changed fundamentally. Water is now a property right. This changes everything. When we started trading in land, we set up a whole industry around measuring that land, and that was the surveying industry. We defined entitlements, we had registers for land title. We have given water rights the same status under law as land title, but we lack the ability to accurately measure these entitlements John Langford talked yesterday about measuring water in elastic-sided buckets in this country. People are now paying a lot of money for the water, and the real challenge is that if we can't measure it, we can't manage it.

(Click on image for a larger version)

I would like to use an example here of how the importance of water information has changed radically in the last couple of years. This is an example of a water right in Victoria called a winter-fill dam licence. A winter-fill dam licence is the right to pump from a river into a dam on your property, to then use that water during the summer. So you fill the dam from the stream in the winter; you use the water in your dam during the summer. Traditionally a winter-fill dam licence had a total annual allocation and a total amount used, or extracted. As long as the total amount you extracted was less than the allocation, you got a big tick.

They had a bit of a problem, though, because in some cases in Victoria during the winter you would get five landholders all pumping at the same time, and the stream would dry up. So they thought, 'How are we going to deal with this?' They brought in a policy called Sustainable Diversion Limits, which is a pretty sensible policy. It says you can't take water from the stream unless the stream is above a certain level. That has fundamental implications for water information.

(Click on image for a larger version)

First of all, in order to know whether you are allowed to pump, you need to know what the stream flow is. The red line you see here is the sustainable diversion limit, and the trace is the flow. Now, most of our stream gauging information is not real time, so you don't know in real time what the flow is at that point in the river. So the first implication is that you need to know real-time flow.

The second implication is that you need to know when the pumping has occurred. So the metering needs to be time-tracked. You need to know when the pumping has occurred. Here, in this instance, the pumping has occurred when the flow is above the sustainable diversion limit; in this other instance, however, there has been an illegal pumping event. If you think about most of our meters, you realise they are not time-tracked, they just measure total volume. So the second implication is that as well as real-time flow you need time-tracked metering.

The third implication arises because we are now trading in this water, so we have entitlement trading. While the total volume that has been owned by a particular landholder over the year may match the total volume that they have extracted, they may not have owned the water when they extracted it. So you need time-tracked metering, real-time flow and geo-referenced and time-tracked entitlement information. These are three types of information that currently don't exist in most places, and if they do, they exist in different information systems flow will be in a state water-resource database, pumping or metering information will be in a water authority's metering record, and entitlement will be in a trading register.

So we need to be able to pull three types of information together, just to measure the impact of one little policy change. If you multiply this by all the policy changes happening in water at the moment, you can see that water information is going to explode as a need.

(Click on image for a larger version)

What is the WRON, and why are we keen on it? Fundamentally, it is about a national network of water information being made available at any node throughout the country by a system of web services. The real point here is using web services, and I am going to talk in a couple of minutes about why that is really important.

(Click on image for a larger version)

What is the research that we are doing and where do we really need the changes to occur? There are four components here that I will talk about. The first is sensorisation, improving the way that we measure the data stream gauging, water quality information, getting much better at our measurement.

The second is data integration. We have 600 different agencies, all with data. How do we pull it all together? How do we make maximal use of different types of data together?

The third is modelling and forecasting. Improving the underpinning data and information will enable us to really transform our ability to model and forecast water availability across the country.

Last of all is thinking about how we actually deliver all of this to everyone in the country not just the water managers, but the public and having different ways of providing water information.

All of this needs to be underpinned by some national standards and information processes.

(Click on image for a larger version)

(Click on image for a larger version)

To look first at sensorisation: we are seeing some real advances in sensor technologies. There are some incredible changes in miniaturisation coming on. We think this is going to drive enormous improvement in the coverage and timeliness of water information across Australia.

(Click on image for a larger version)

We are in a geo-referenced world now. GPS on a chip costs $1 to be incorporated into a mobile phone, but radio-frequency identification (RFID) tags are less than 5c. So every bit of equipment, every bit of information will be able to be geo-referenced.

I found out the other day that Microsoft has already developed the software for a Whereabouts Clock. Those of you who have read the Harry Potter books would know that the Whereabouts Clock tells you where your family members are at a particular point in time. So the geo-referenced world is coming, and it is going to fundamentally change the way we deal with information.

(Click on image for a larger version)

(Click on image for a larger version)

On data integration: we have already had attempts by various states to bring data on water together, and there are some really good examples of that the Victorian Water Resources Data Warehouse, and similar examples in other states. But we also need to bring together different types of information, such as satellite remote sensing information, very fine scale, and bring it together with the water information to be used in models.

(Click on image for a larger version)

We need to think about how we can better get information out of that data. This slide shows an example of some work done at CSIRO, utilising a data assimilation process that has been very common in the climate science area but hasn't really been applied to water to date. It takes a whole bunch of different feeds from meteorological radar, in situ time series information, surface remote sensing and microwave information, and pulls it all together and gives you a three-dimensional and four-dimensional view of what the water balance is at that point in the catchment. So this is about bringing together the strengths of in situ data, which is time based but not very spatially explicit it doesn't have good spatial coverage with remote sensing data, which has great spatial coverage but no time coverage. So we need to develop our sciences in these areas.

Modelling and forecasting: we are really excited by the potential to fundamentally change the ways we forecast, through the availability of extra data. I have here a little example of a forecasting tool for irrigators, for several months ahead, which we are developing by utilising some of the technologies developed under the WRON.

(Click on image for a larger version)


(Click on image for a larger version)


(Click on image for a larger version)

We have got challenges in Australia with water forecasting, because to date it has always focused on flood forecasting. We haven't focused on the water availability forecasting in a forward-looking sense.

What we are doing here is taking short- to medium-term climate information, like the Southern Oscillation Index, with real-flow data from data sources, putting it through forecasting models and then forecasting what the dam levels will be, six months out. That can then be translated, through the usage and rules that most of the water authorities utilise, to create an individual tool for an individual irrigator to show them the probability that they are going to get their water right, six months out. That gives them a chance to actually decide to plant a crop or not plant a crop, for example.

So sorting out the data end of things enables a whole range of different tools that we haven't envisaged to date.

(Click on image for a larger version)

The last area I would like to focus on is the reporting and visualisation systems. This is where the live demo comes in. The example that we are using is Australian dam level information. To get dam level information across Australia previously, you had to visit multiple websites, each of them run by the individual water authorities responsible for the dams, and you had to delve through different web pages and access that information in the form in which it was provided. And the situation is even worse with water quality information.

(Click on image for a larger version)

The dam level monitor that I will show you harvests that information once a day, from multiple sites across Australia. It re-serves it as a web service, and what that really changes is our ability to deliver products in a whole different way. It allows different products to be served from the same data, and I will show you that in a second. It provides access not only by a human browsing that data, but also by machines and models. And you can find it at the http://WRON.net.au website.

[An internet demonstration was then shown, with a commentary as follows.]

I am just going to show you a couple of things.

The first is the WRON website, and I am going to show you this dam level data in a few different interfaces.

First of all we have the 'Google map interface', using Google Maps, which people are familiar with now. This shows you the whole of Australia, with every dam in its location. The colour allocated to the dam relates to the current level of storage. Because it is in Google Maps, you can click on a dam and it shows you the current capacity and a historical record of what it has done in storage. And you can zoom in, overlay satellite imagery, do all sorts of neat things. This data is driven completely live, from the web service that we have created out of this dam level information.

The second part to this is that you can click on a little box at the top and the service downloads machine readable data, so that you can point a model to this page and the model can ingest that data in a readable way and populate itself to provide that information.

The same interface, the web service, allows us to serve different views of the data. If you click on 'Check the gauges' you see gauges like those on your car dashboard. They are driven live as I click on them by the web service.

And a third way of accessing the same information is through 'View as time series'. So here we have total capacity for Australia. Currently we are at 52 per cent. You can break it down to particular states and look at the particular regions. Queensland actually looks not too bad, but when you look at the south-east Queensland area you find they are very low at the moment.

(Click on image for a larger version)

The last way of accessing that information from this page is through 'Widgets'. A widget is a new way of accessing information from your desktop. Some of you may have similar things on your computer already. I have got the weather report for Canberra for the next few days, a weather radar from the Bureau of Meteorology, and we have a dam widget. The dam widget allows us to look at the same information, to look at the historical level of that dam, and you can change its location just by clicking to choose a different dam. It updates automatically over the web, and you are able to get up-to-date information on your local storage. This is all available now. You can download it from the WRON website.

The last thing I would like to show you is an experimental tool that we have developed, which is delivered in Google Earth. Google Earth has recently released a new update that is, I believe, going to fundamentally change a lot of delivery of information. That update is the little 'Time Slider' tool that you see at the top of this screen. Google Earth transformed things by bringing three dimensions into our information space, but what the Time Slider does is to allow us to bring in four dimensions. We can slide through time.

What we have here and what I will demonstrate is a visualisation of the dam data for the whole of Australia. As I zoom in to the Melbourne area, I can drag the Time Slider through time and watch the dam levels going up and down. The shadow you see is the total storage for the dam, and a little point at the bottom allows me to click on that and access data for the dam. So what we have here is that on 28 May 2002 the actual storage in Thomson Reservoir was at 49 per cent, shown by a little dot. If I want to zoom forward in time and look at what it was on 1 May 2005, I can do that.

This is, effectively, a live visualisation of a water account nationally. You can look at where the water was, anywhere in Australia, at any point in time. We think this is the way we are going to have to deliver water information in the future.

To wrap up: we are seeing some real technology advances in water information. They are now fundamentally important in improving the access, availability, forecasting and reporting of water information. And we are working with a range of partners to deliver these sorts of technologies in real systems to help with Australia's water crisis.

Discussion

Question: I have two relatively simple questions. My first question is really about how you chose the nodes for the network, and the second question is about the type of data that you use. Do you use data from organisations like Waterwatch and so on, to inform that?

Stuart Minchin: The vision for the WRON hasn't eventuated yet. We are showcasing some of the technologies. The actual operational system is what is hinted at in the Prime Minister's water plan. There needs to be an operational agency that runs it, and they will decide on the nodes that are specific. That map I showed was an exemplar.

In terms of Waterwatch data, it is very close to my heart I was involved in a Waterwatch program when I worked in Victoria. My view on Waterwatch data is that it is very useful for certain questions that you're asking. What we need to get to is that all data is made available but you are very clear on how it was collected, to what standard. That enables you to choose the level of evidence that you wish for the particular question that you are asking.

Question: How do you determine who is using your data? Do you have a hit rate? Do you monitor usage of your service?

Stuart Minchin: First of all, it is not 'our' data. This is about enabling access to everyone's data, which the public has already paid for.

I think the key point to make here is that in many organisations earlier I mentioned IT being in the Dark Ages people want to control the information flow, from the collection point right through to how it is finally used and interpreted by someone. That is the way we used to work with information. We don't work that way any more. You have seen the advent of things like Google, Yahoo!, YouTube, all of these web service oriented places. What they do is to make available interfaces to that information. They don't try and control the final graph that you print out. They make access to the information possible, make clear the metadata around that information, and then allow people to make use of the delivery mechanisms.

Question (cont.): Yes, but you would like to know who is using what you have prepared.

Stuart Minchin: Ideally, but there are other ways of tracking that, for example by looking in the literature. A good example, potentially, comes from when I was involved in setting up the Water Data Warehouse in Victoria a few years ago. We used to track who asked for the water data in Victoria, and we used to get 400 requests, on average, per year for the data that made up that database before it went online. In the first year of it being online, we had 50,000 downloads of that information. So there is a point to which you can track, but once you have made it available, people use it much more commonly for all sorts of things.

I will just hark back to a little note I made when Robert Humphries was talking this morning, about the $80 million bridge that was washed away due to poor data. I think this is costing us a lot at the moment, in over-engineering structures because we don't have the information available, in taking the precautionary principle on information. So this is not only about saving future costs but also about improving the way we currently engineer things.