Note: Where available, the PDF/Word icon below is provided to view the complete and fully formatted document
Long-term meteorological forecasting

CHAIR —I welcome the representative of Weather Risk Management Services. Although the committee does not require you to give evidence under oath, I advise you that these hearings are formal proceedings of the parliament and, consequently, they warrant the same respect as proceedings of the House itself. It is customary to remind witnesses that the giving of false or misleading evidence is a serious matter and may be regarded as contempt of parliament. We thank you for your submission. I now invite you to make a brief opening statement before we proceed to questions.

Dr Werner —Good morning. Thank you very much for the invitation to participate in the hearing today and make a submission to this I believe long overdue inquiry into the services rendered at the moment by the Australian government through its arms of the Bureau of Meteorology and CSIRO. I will provide you with a brief background. Weather Risk Management Services provides consulting services to weather and climate sensitive industries here in Australia and overseas as well. We generate the products that the Bureau of Meteorology and CSIRO are producing and far beyond. It is not just season outlooks but very detailed risk assessments. That is basically how we have built over the years our reputation here and abroad as well.

What the public and you may have heard presented already by various eminent scientists in Australia, especially by the Bureau of Meteorology and CSIRO, is a little bit of what is possible. There is far more possible than what is in the public arena. I do not really want to criticise the government, but because it is a large organisation it just takes longer to adjust and bring forward the latest knowledge and techniques to provide products and services on which we can then base business decisions. Basically our business on a daily basis is to provide such information to weather sensitive industries here in Australia.

Looking at the terms of reference for this inquiry, you would be targeting more the tourism and emergency services but there are whole other sectors of the Australian economy that may be affected by weather. There are the farmers, the commodity traders—maybe through the AWB—the national electricity market and the Victorian gas market which are highly weather sensitive. For example, tariff case submissions depend on some historical analysis as well as looking forward. If a vertically integrated energy business—and we have some of those in Australia in Victoria and here in Sydney—were charged with acquiring a hydrogenerator, how would they go about that, for example?

We are participating in these sorts of risk assessments with significant value at risk: we are talking about hundreds of millions if not billions of dollars either on the traded national electricity market on a daily basis, in Australia or abroad, or in these large-scale acquisitions. So a lot of human activity and business activity depends on reliable and trustworthy short- to long-term meteorological information, therefore it is absolutely pivotal. I hope that this inquiry comes up with some recommendations at the end of the day about how we move forward from here. If we were just to continue with business as usual then we might give $100 million for a new computer and a few more resources, but what would we really gain? Would we actually have the products which enable us to make sound economic decisions within an uncertain environment? Uncertainty was, is and always will be part of the decision-making process and of scientific processes as well. Again, Weather Risk Management Services is sitting right at the interface when it comes to discovering the latest science and then applying it in the business arena. I think that is one of the things which is still missing because the climate scientists are in one camp and very often when it comes to the application of their latest findings too much time goes by before they are published. When do we get a new, more powerful computer so we can tackle these grand challenges of science and engineering?

Climate and weather forecasting is a grand challenge. The National Science Foundation in the US recognised in the 1990s that one of the grand challenges to human survival on this planet was to master and understand how the climate systems work, otherwise we would not know what is happening in 10 or 20 years. Through the IPCC mechanisms we have some understanding, yet it is currently taking too long before new results are publicised and are being applied in the business arena. Then we hear in the news that some of these findings are already superseded by observations, which is very scary. So time is really running out. We should not rush through this, but we have to change a few structures, I believe. We have the building blocks already, but we have to arrange them in a clever way to move forward.

CHAIR —Thank you. I have a couple of questions but we will go first to one of my colleagues.

Mr RAMSEY —Dr Werner, are you saying that the technologies are there but that, to your mind, the bodies that control the technologies are being overly cautious and wanting to test them to the nth degree before they are making them public?

Dr Werner —There are various approaches to how to go about this, to be more cautious, and obviously everyone would not just be using a better version of a model, for example. Maybe I should take one step back. We have various ways of going about addressing and solving the current challenges of long-term meteorological forecasting. You have probably heard already that we have all these different highly sophisticated tools. Then we have great believes in two camps, the statistical models and the dynamical models, so that is what the inquiry is now about. So we already have these tools, and you would have heard or read about the different levels of skill or fidelity in terms of how they perform. Again, there are different ‘believers’—sometimes things develop into a religion for some people within the scientific environment because there is a lot of funding at risk, let alone their positions or jobs. We have to develop an objective approach to this. What is really available? How do we go about this? Maybe we need to look overseas at the way it is structured and funded as well.

To answer your question, yes, the building blocks are already there. We do not need to be overly cautious; it is more like we are being a little bit protected here in Australia as to how the funding is being done. We already have the building blocks, which were developed in the seventies and eighties. The Bureau of Meteorology, as you may have heard or read, was absolutely leading in this and we had absolutely great scientists. We still have some of those, but they are reaching retirement age. There was very little done throughout the 1990s in terms of nurturing young scientists going through the universities here in Australia and having broader-scale funding—not necessarily more, just not only for greenhouse, for example. That was the big buzzword in the 1990s and there was very little research funding available in Australia for aerosol particle research, for example. You may have heard or read about that; it is now a big thing in the IPCC. One of my PhD supervisors was involved in that. But there was virtually no funding at all for that here because the big focus was on greenhouse, with millions of dollars committed to that.

As for these building blocks, it was very tough for me in the 1990s, when I was studying towards my PhD, to get any funding to actually do this. As it turns out, yes, it is a big thing now. Air pollution and volcanic eruptions all play a role. Yes, we do have these different building blocks here as well but we have to make some more publicly available, so it is something like open source rather than spending a lot of money—and we are spending tens of millions of dollars. There is building a new model, for example, and if we use a dynamic model rather than a statistical model. We are spending a lot of money and it is sort of reinforced. But there are other opportunities to participate in this. As I said earlier on in my opening brief, we are running out of time. For example, three months, six months or 12 months in the business world is a long time, and it is so even with the climate because changes are happening very rapidly. We do not have all the tools to master this but we have got some building blocks. We have got what I call the passion and the persistence to tackle these problems successfully. Yes, we do have the building blocks but we have to rearrange them as to the way we are doing business in a climate sense, otherwise we will see a repeat of the 1990s, for example, whereby one group will get a lot of funding. In the university sector, for example, there are a lot of young brains just scraping by with very small funds. Sometimes you do not necessarily need big funds, but that is just the way it is organised. You may have already heard that there are collaborations. But given the way I see things evolving here in Australia having those is not enough. I went through this; it is very tough out there.

There is a big thing to develop this new earth system simulator. There are different European countries developing it. There is already a lot of open source material as well. Maybe Australia needs to go down this path and provide some funding for infrastructure whereby young scientists and established eminent scientists here in Australia can participate in this and really make a quantum leap forward. That is very important, because time is running out here in Australia. We would not like to see a repeat of the bushfires that we had recently in Victoria, for example. If the population and emergency services had been alerted earlier, we may have been able to prevent some of the deaths and casualties and loss of property which occurred unfortunately during those bushfires.

CHAIR —If I could interrupt you for a minute, Dr Werner, the previous witness was from Firewatch Australia and he was talking about a device that has been developed and installed in Germany and other parts of the world. Its practical application is to actually detect fires well before a human eye can see them. The question I want to ask you is what I asked the previous witness. It is really about funding, isn’t it? I think a lot of what you are saying is about that and about this issue of time running out, which goes to the core that either we do not have the human resources available or we are not cultivating them in enough time. Is that right?

Dr Werner —Yes.

CHAIR —That is in order to be able to do that. We are aware that we are falling behind in this area, because we were told that yesterday.

Dr Werner —Correct.

CHAIR —You made the point about, and you drew the example of, the Victorian bushfires and said that if we had better systems of forecasting there may have been a different result. My question is: would something like Firewatch be sufficient as a detection device to deal with the issue of climatic conditions that could lead to a high fire danger and catastrophes? Otherwise, is it the actual detection process and forecasting the place where you spend the money to develop expertise and precision? Do you see both working together?

Dr Werner —It think it is probably about the last point. They have to work together. It is all about how information is exchanged—and this is very important—freely. So it is not about firewalls or about a cone of silence being put around experiments. People talk about the technology—and you may have heard or read about it—and say the new forecast models are experimental. That is how it all starts, with some research, for example, but it is all about information and how information is exchanged so we should not charge just one organisation or one body with having the sole purpose of monitoring and providing.

At the end of the day, somebody has to have the legal right to make an announcement, such as: ‘This is what we believe is happening this summer.’ Weather Risk Management Services cannot issue a fire warning. You may not like what I am now going to say but, from July last year, we knew about the drought conditions leading up to the bushfires. We knew about the heatwaves. Over the last two summers, the southern states have been having significant heatwaves. We knew about that in the previous July-August period. Again, the way information is exchanged is not part of our business. The technology is already there. We are operating in different spaces.

CHAIR —You have the capacity to be aware of those conditions?

Dr Werner —That is correct.

CHAIR —But you are telling me that you could not disseminate that information?

Dr Werner —That is the way our business model is structured. Who would support our research into the conditions leading up to heatwave activity as such? It is my understanding that, by law, the Bureau of Meteorology, emergency services and CSIRO are charged with this research in Australia. My previous point was: it is about how information is exchanged. For example, there will be one organisation, and so again it boils down to funding. We do not receive any federal funding to do any of our research. Funding of our research is based purely on a fee structure. We charge our clients here and overseas to provide them with operational services, risk assessments. Some of this in-house money will be spent on research, because we cannot spend any time on writing grant applications to do something. That takes too long.

CHAIR —What did you do with the information you had in relation to the possibilities of what happened in Victoria?

Dr Werner —It was provided to our fee-paying customers.

CHAIR —It was provided to fee-paying customers?

Dr Werner —Yes. The Australian government or any of the state emergency services are not part of this. Yes, it was sold privately. For example, I would have conveyed to friends in Victoria that there was an increased risk of bushfire this coming summer. With reliable long-term climate forecast information, it was not too difficult to figure out that that was going to happen. It was not just the temperature; it was a straightforward meteorological event. My understanding is that there is another inquiry into the bushfires in Victoria. I am not following that inquiry closely at all, but we—

CHAIR —It is actually a royal commission.

Dr Werner —Sorry. I am looking to it. Again, we already have these building blocks but we have to sit back and think: how do we actually do business in Australia? We have to do it in a smarter way. So how do we exchange this information? I do not receive any state or federal government funding. For example, I have customers who pay a fee to obtain certain information about heatwaves. Heatwaves in the southern states would indicate an increased airconditioning load coming up in the summer. This is basically what happened in the previous two summers. The interest was in that, and a by-product of it would obviously be bushfires. Bushfires also impact on the hydro catchments of AGL and Snowy Hydro. If you have a bushfire going through those catchments, it changes the catchment’s characteristics and that can have an impact on the availability of electricity during critical periods.

So, again, the building blocks are already there. The way that we exchange information has to change in this country, because it is basically one of the road blocks. I would be unfair to my customers, my clients, who are paying a lot fees to obtain such sensitive information, if I exchanged it when we do not have any formalism to do so. It is not my bad intention to withhold information. It is basically a business case as such. So how do we do business in Australia? Are we actually allowed to exchange information? How do we cooperate with the Bureau of Meteorology, for example? The Bureau of Meteorology is allowed through various acts to issue the warnings, whereas we as a privately owned business and, as I understand it, other providers in Australia are not in a position legally to issue warnings. We may burn somebody. So there has to be a better formalism as to how we exchange this information.

CHAIR —How do you see that ‘formalism’ for an exchange of information? Would you sell your information to the bureau? How would you see it working? How would it happen?

Dr Werner —That would be one avenue to pursue. We are generating this information after having assessed that it fits the purpose of providing quality, long-term forecasts. That would be one mechanism by which to proceed. We incur a cost to generate this information. We are not obtaining this information from a third body. We are actually generating in-house short-term to long-term assessments. So there is actually a cost to our business.

CHAIR —You would sell it to the bureau.

Dr Werner —That would be one mechanism of doing it. If you asked me obviously we felt bad about this, what was going to happen in Victoria, but my understanding is some conflict of interest with the Bureau of Meteorology if I would have issued a press release, for example. I do not know what the full legal ramifications are if all of a sudden Weather Risk Management Services starts to issue warnings.

Mr RAMSEY —Given that you predicted a high risk of fire because it was going to be a hotter than average summer, did you have the ability to predict the extended heatwave as such? I think we had 14 days, which was unprecedented at that time of year. We have had that large blocking high sitting there maybe a month later, maybe six weeks later, but at that time of year we had never experienced a high, in recorded history at least, that had blocked the weather up for that long and allowed the northerly infeeds to continue day after day after day over a two-week period. Did your model pick up that there was a high risk of that occurring as well, or were we just looking at the hotter summer with higher fuel loads?

Dr Werner —The short answer is yes, we knew about it. The mechanisms for this last heatwave are slightly different and somewhat more concerning, because the heat in the interior of Australia normally during the summer period, looking at heatwave activity in South Australia and Victoria, is such that most of the hot air is sourced from northern Western Australia or the northern parts of Australia. This time the heat was actually sourced from the central parts of Australia and northern South Australia. Very scary. I have seen it in some of the climate change simulations that the centre of the heat has actually moved further down, so the actual source of the hot air was much closer to the impact area. You needed much less to tip it over basically and produce this heatwave activity. To answer your questions, yes, we did.

Mr RAMSEY —So that is a more northerly airflow than we have had before, or is the pool of air it is sourced from closer, so it is hotter.

Dr Werner —Both. First of all the air was sitting there and the meteorological or synoptic situation was such it was not plucking in, and once it plucked in through the northerly winds and was north-westerly or northerly or even slightly north-easterly, north-north-easterly winds, it just kept rolling. Then when you have a stationary weather pattern you can have a day and night when you have warm invection, so transport of horizontally dry and warm air into the target area. Then you have the fuel load, which was another tick, and we had an ongoing drought with reduced soil moisture. So you have got to pick all this up. So when you do the modelling the way business is done currently it is unable to do this because we are providing seasonal outlooks in a three-month block and a lot of things can happen. Is it going to be warm or hot at the beginning, in the middle or at the end or for the whole block? It gets very blurry. It is a statistical techniques so you are looking for analogues. We never saw this before and we have to come to grips with this. There is going to be more of this. I am not painting this as climate change is going to happen or climate change is something that is happening in 2050 or 2100. Climate change always has happened, is happening right now and will happen again. There is no such thing as an upper limit. So we can get a heatwave again where we sustain temperatures in the low 40s to mid-40s. There is nothing speaking against this. We have seen it and it is possible, and we need to have tools which can simulate this and pick it up on the extreme days. On one of the days we were only 0.3 degrees out in Melbourne, for example, when it reached more than 46 degrees. It was reasonably easy to predict these sorts of things when you have the right tools. You have got to have the right tools. It is like as a carpenter you have to have sharp tools basically.

Mr RAMSEY —Even if a public body buys the information, it is how they change public behaviour as a result of it. What happens is that next season we will see people in Victoria being very conscious about their fire prevention measures. Over an extended period of time looking back to the Ash Wednesday fires in south-eastern Australia over 30 years ago, people were very good at looking after their properties and you could give the mornings but over a 30-year period that collective history disappears and people go, ‘Oh yeah, they warned about a hot summer last year and we never got burned out.’ It is very hard to keep warning people that they are heading into high-risk seasons and actually get them to do something about it. That is not your problem, but it is a question of how you influence public behaviour with the information you have at your fingertips.

Dr Werner —That is a very interesting point. It is basically a level of general education when it comes to meteorology or weather and climate in general. Comparing Australia with different countries, the US, for example, appear to be much more educated, if I can use that word. Therefore, they are more aware of it. Maybe the Australian psyche is, ‘Mate, we’ll be alright; no worries.’ People living in those really beautiful places forget about other terrible bushfires in Australian history for history. When it comes to weather, people very often forget quite quickly. We had a lot of brush there. We need to step back and look at volatility. We need to start—maybe through the Bureau of Meteorology—an education campaign of some kind that is part of the Australian education system.

We are living in the driest inhabited continent on this planet. We have deserts in the interior; we just live on the fringes. We need to teach people about the risks. People forget about these things. They think: ‘Real estate prices are going up. We have beautiful views.’ They are living in the middle of nowhere and that is highly dangerous. People living in those places need to take certain precautions so as to prevent this sort of thing from happening again. We cannot prevent these bushfires; we cannot prevent these meteorological situations—well, we can, but that is a different story to do with climate change. We have to take onboard the education of the public and tell them what the risks are and that it is an ongoing issue. We have to treat it in the same way as AIDS or as the swine flu virus and educate the public. Education about meteorology should be an ongoing process: these are the key points and this is what it means. It should not just be on a web site, because not everyone logs onto the Bureau of Meteorology’s web site. It is very important, especially for people living in the high-risk areas, not to become complacent.

CHAIR —On the issue of educating the public, it is quite clear to me from talking to people that a lot of people do not have any confidence in weather forecasting. The attitude of, ‘She’ll be right, mate,’ may be based in the fact that people believe that the forecasters do not know what they are doing. Can you comment on that?

Dr Werner —Yes. Unfortunately, when you look at the forecasting track record, it is somewhat true. In science, especially the meteorological science or atmospheric physics, there is a lot of skill involved. The forecasts can perform reasonably well. But if you are interested in a particular day—such as today—and ask, ‘Do I need an umbrella to come to the inquiry today?’ I see all the clouds out there. The Bureau of Meteorology is telling me that it is going to be 23 degrees or 24 degrees. Again, I must apologise; I do not want to bash them up or criticise them. But they are basically the main source of information in Australia. They do not have a terribly track record. We have to change this. We need to present information in a different way so that the lay person can understand it. That is our job as well.

As a qualified person with a PhD in theoretical physics specialising in numerical weather prediction modelling, I could go all the way and people’s eyes would just glaze over. Nobody would listen. They would say: ‘Look, mate, I don’t understand this. I have to make a decision here. I have to purchase so many megawatt hours for summer. Should I do this, yes or no?’ It has to be brought down to a level that is easy to understand. It is a large challenge to visualise it. We have on our hands a multidimensional problem. There is not just temperature; there is humidity, soil moisture, the role of clouds and rainfall. We have all these difference indices. It is a zoo of indices and terminology. But we have to start somewhere. We need to produce an information pack that explains what things basically mean.

Yes, it is true that the track record has not been good, so there is distrust. We have to build trust and maintain it on a daily basis. Yes, sometimes we can get it wrong, but I have to get it right more than 51 per cent of the time otherwise I would not have a business. It is very important to build this up, and you can only do it over a long period of time. I have been involved in this for a long time already. The other things that you have to have are passion and persistence—and what people call a thick skin as well. You have to really stick it up. You cannot give wishy-washy forecasts. We cannot predict what will happen on 25 March next year—we cannot do that—but we can provide possible scenarios. We have to move away from the notion of uncertainty and move into the arena of risk. Knowing the risk means that at the moment we probably know most of the underlying distributions of what may happen. Uncertainty means that something is completely unknown or ambiguous. We have to deal with this.

At the moment, looking through the scientific literature, you find people use, in an interchangeable way, ‘risk’ and ‘uncertainty’. For a risk manager in the energy market they are two separate shoes altogether—or pairs or whatever. This is what it comes down to again. You can hide very quickly: ‘I was uncertain; that’s why I got it wrong.’ But at the end of the day they have to make a decision. The public has to make a decision too: do we have to evacuate—yes or no?

It is a process which may take a few years. We have to gradually build up this trust. We have to run education campaigns, maybe in the school, to throw in a few more things. The other thing is that here in Australia—I obviously come from Germany—the education is slightly different. Again, that is where it is coming from—there is a long history. Whereas in Germany you can study meteorology or atmospheric physics—it is a completely encapsulating degree at the university—here it is seen as mathematics or physics. It is just a sideshow type thing. I am not saying that meteorology is not a great thing to study in Australia; you can still do it, but it is completely different. In the United States you can study just meteorology or atmospheric physics at a meteorology institute or department. By training I am actually an atmospheric physicist.

So you start from a completely different level and base of openness. That is missing here. And then you can see: you got this wrong so many times already. There was Inigo Jones. He got it right a long time ago—or maybe not. There are lot of things happening in the folk law as well. There are some private providers here in Australia too, but you have to start somewhere to build this trusting relationship again. And you have to take some risks. The risks are taken through the organisations of CSIRO or Bureau of Meteorology or maybe through some of the private net service providers taking some of the more cutting-edge building blocks and coming up with new ways of presenting the information to the public. So they say, ‘Oh, it’s actually much more possible.’

So we are not driving, anymore, in a Holden Kingswood SL—a 1970s or eighties car—because now we having much more sophisticated cars. It is the same when it comes to science and technology. We now have these supercomputers. We do not have Crays any more. We have IBM or NEC supercomputers. We already have all these building blocks and we have to move forward. But we have to move quickly as well, because one summer goes by and another summer goes by and many more things can happen.

Mr SYMON —There are several things I would like to pick up on there. I note in your submission that you talk about publicly available seasonal forecasts and how they are based on statistical models. But also, as you have just been talking about, there is a lack of ‘user friendliness’—if that is the right way to put it—in getting the information, even if it is based on the old models or on a new dynamic model, to the end user. And we have heard from a witness this morning on the same subject. Do you have any ideas as to how that message could be put across in a way that an interested user—Joe Public; maybe a small business operating in the construction industry—could be able to understand it and use it?

Dr Werner —That is a very good question, because at the end of the day we always have to look to the end users. So we are not doing it just because we love to do meteorology. At the end of the day—and you see this in all the funding reports and submissions—it has to benefit someone. On a daily basis we have to demonstrate and prove this.

And tomorrow is another day. Just like in a football team you run out and try to score again. It is a completely new day. You are not necessarily starting with a clean slate but you have to think, ‘How do we make these people some money?’ For example, for a construction business, looking at a three-month period is just too blurry. It is basically like taking off my glasses—or somebody else’s glasses—and trying to drive on the F3 Freeway. Not only would it be against the law, because I have to wear my spectacles, but it would be all blurry. So you cannot really make a decision. It is very uncertain.

For example, in the construction business—and, yes, in the past we worked with construction businesses here as well—if they take a roof off a data centre or a bank here in Australia we work with project managers and engineers. Engineers are a different breed to scientists, because with scientists there is always some fuzziness, uncertainty or risk around. With engineers they would like to know everything so they can calibrate and measure against something so that that they know the risk. So they know whether they have to put a tarpaulin on and what sort of tarpaulin, or whether they can proceed at all with the project on a given date.

So there needs to be a much finer level. There have to be different products. The current product is targeted at a very broad sector. From that alone it would blur, and obviously in a three-month period there are rainfall and temperature outlooks as well. Very often businesses are working on a quarterly or monthly basis, so you need a much finer temporal resolution in your products and services. Maybe you would like to issue probability distributions. You may have heard about it or read about it. There are ensemble techniques, which would be one way of assessing the risk around a certain outcome in temperature or rain or hail or snow. You are using these sorts of approaches, for example, where you would offer a probabilistic information type forecast but based on deterministics. Statistical lists are basically a data reduction technique. I have an example of 10,000 data points. What is a mean? What is the standard deviation? That is what you would be interested in. From there you would infer the most likely outcome—possibly.

This kind of avenue may work. You could say, for example, that in January, historically, so many thunderstorms occurred in Sydney or in Brisbane. Going through our dynamical forecast systems, we have generated various ensembles. Based on these ensembles we come up with these statistics, and they would be more detailed. As you get closer you would be zooming in. You would use different tools again to reassess the risk of the forecasts you have issued recently. In that seasonal outlook, you are getting closer to current days. You would have more information, closer information, but it does not necessarily mean, as you are getting closer, that your predictability will increase. That is not true. It is not true because there is generally a perception that I can predict the weather very well for today or tomorrow. There are a lot of examples where it is not the case and we would be much better performing further out—for example, if there were rainy periods or heatwave activity. So we have to disengage from this notion as well.

We have a lot of examples in our business and through our clients where we have predicted flooding and heatwave activity which occurred at the same time. For example, that happened in Queensland in the previous year’s summer. So you have this issue of seasonal outlook and you are getting closer and then you would zoom in, literally providing a product. You would go back to the customer or client and say, ‘We have to know this information. We will reassess the risk; it has changed.’ So basically you have a sense of ownership. Again, I believe they are passionate and persistent about it, but it seems to be somewhat lacking. You have this sense of ownership. It is not just file and forget—you do something with it. That is what it appears like on websites here and overseas. So there has to be a sense of ownership as well. There are livelihoods at risk, businesses at risk, earnings at risk, capital expenditure at risk. There are a lot of things. That needs to be absorbed as well and not just said in the statement. We need to be fully aware of the implications. A more entrepreneurial approach to this may be the way to go in the future.

Mr SYMON —Thank you for that. You mentioned ensemble type forecasts. Again, going back to your submission, you say that the US National Centre for Environmental Predictions runs ensemble forecasts four times per day with 21 ensembles, resulting in a total of 84 ensemble member forecasts for the next 16 days. I take it the Bureau of Meteorology is not in that league, from what I have read in your submission, although I have not asked them about this. Could you fill me in on what you understand they do?

Dr Werner —Yes. I am obviously not fully involved with the current research activities in the Bureau of Meteorology. My understanding is there was always an attempt to do ensemble type forecasting in Australia, currently the PME model—the new seasonal dynamic model. But what they are doing is what is called the poor man’s or poor person’s ensemble. Basically you do a forecast, a straight dynamical forecast, say, for six months, assuming you have a seasonal forecasting system. Tomorrow you do another forecast and so on. After a week you would have seven ensembles covering six months out. You are moving forward in this window rather than having multiple initialisations. Based on your initialisation schemes for the ensemble forecast—with the National Centre for Environmental Prediction in the US or the European Centre for Medium-Range Weather Forecasts performing—you have multiple initialisations per given day.

The way it is done is that you change slightly the initial conditions. You can maybe change parameters within the model—physical aspects. There are different ways of doing this, basically moving forward. The more you have, basically a picture is forming. It is like having a box camera where you have one hole to look through, so you would not see much, like the eye of a fly. If you had 100 holes in the camera you would see more. Take a chip in a digital camera where you are talking about megapixels—you can resolve much more. Resolution is almost everything, going down through horizontal resolution as well as having temporarily more of these ensembles because you have more of these risks addressed. Otherwise it is hugely uncertain.

Mr SYMON —Does that come down to a lack of resources in Australia or a lack of computing power or a lack of dollars? Otherwise, does it seem to be a preference for a different direction?

Dr Werner —It is probably a sort of a mix of all of the above. Predominantly, I would say—and I mentioned this in my opening statement—Australia used to have really smart people working in the field. One of my other PhD supervisors was Professor Lance Leslie, who was back then at the Bureau of Meteorology’s research centre and was one of the best dynamical meteorologists Australia has ever produced. He was by training a mathematician. After he left the Bureau of Meteorology he became a full professor of the University of New South Wales. I had three PhD supervisors. He then moved on and went to the US. It is my understanding that Dr Bill Burke is still at the Bureau of Meteorology. He is another very eminent person.

Again, it was not nurtured but I do not know for what reasons, maybe funding or maybe there was an executive decision somewhere. It is not always an issue of funding; it is how the funding is delegated. There is the politics itself. Through the 1990s here in Australia the flavour of the day was CO2 emissions, and how they impact on climate, and the greenhouse effect. There was very little on aerosols and very little on the radiative properties of the atmosphere. Only a few people on this planet actually built these models from the ground up and know how these models work. You can use them as a black box. It is like this: instead of sitting in your car today, you are sitting in a jumbo jet and you have to get back to Canberra and you think, ‘There are a lot of buttons. How do I do this?’ With these models there is a risk of misapplying them as well. Again, you have to buy this knowledge. It used to be that it was not nurtured. We have to get back to what we had before, because we are developing some dependency. Too much is depending on it. It is just handing out the funding dollars and getting a third-party model.

There are great models out there. It is basically about a mix of politics itself. Even if we do not see aerosol particles as a red warning indicator, it still needs to be funded—and the same with radiative properties. My background is specialising in radiative transfer and theory. Very few people are actually developing these new codes within these climate models and weather prediction models. When you look at some of them you see some of the codes are dating back to the seventies and eighties, so you should get a new shell. Say you drive a brand-new car but the motor itself, the radiance field, is a motor of the climate and weather system. If you switch it off it does not really look like a weather movement. Obviously there would just be darkness. But that is the motor itself and there is very little done, especially in Australia, to progress this further. My principal PhD supervisor was Professor Michael Box, who is now retired. He was the leading person in Australia when it came to the radiative properties of aerosol particles. He is now retired and there is very little happening as to these very important key issues, so very few people have an understanding of these models.

You may have heard or read that you can freely download a climate model. If you have a large enough work station you can run the model. It is already there and it does not cost you anything. But you have got to have the knowledge itself. Otherwise it is like giving someone the keys but they do not know how to use the keys themselves. Funding is one thing but it is also about resources. But we already have the building blocks and maybe the way to move forward to make this a really clever country again in climate and weather research is by having a dual approach. We would still have the Bureau of Meteorology and the CSIRO but we would nurture by having outside arrangements like those that are happening over in the United States where you have an open-source and developed model, for example. From a risk management perspective, you do not want to put all your eggs in one basket. We are talking about maybe $100 million in funding over, say, half a decade. We do not want to spend this money, so you have to have a dual approach to nurture this human capital. That is one thing Australia can become leading in. It is about not just being able to sit at an international scientific meeting but actively contributing and looking at our risk profile and exposure. I think that is really necessary.

Mr SYMON —Thank you.

Mr RAMSEY —Yesterday we had evidence from some people in the academic world who highlighted that universities had difficulty accessing the technology that the Bureau of Met and CSIRO have in their supercomputing facilities. They highlighted problems with that silo mentality, which is exactly what you were just saying about putting all our eggs in one basket. From what you have just been saying, I gather you would concur with that it is difficult for people to actually progress another thought process or another model in Australia because everything is targeted into this one area already.

Dr Werner —Correct, unfortunately. I wish I had a machine like the CSIRO has in Victoria. It is just great. But, again, you do not have to spend tens of millions of dollars on hardware. I also made the point in my submission that new technologies are emerging. We now have new hybrid cars, for example, which may run just on electricity and not on petrol anymore, so there is a different question about where the electricity for these cars comes from.

A classical computer uses a central processing unit, or CPU. That is the brain where all the calculations are done, and you have a few thousand of them in a supercomputer. They all need to talk to each other, so they are run massively parallel. Backdoor processors were the flavour of the day a few years ago, but now we have these massively parallel computers. They are all run together, with a computer code running, to solve all these complex equations to basically produce an outlook. But that is old technology that needs a lot of energy and has absolutely huge cooling requirements. Cooling is the single biggest issue worldwide when it comes to data centres and supercomputers—they run really hot.

New technologies are emerging, for example, in GPU, or graphical processing unit, computing. We recently went to an exercise with one of the larger international computer and supercomputer manufacturers, and we discovered there is a factor of anywhere between 100 and 1,000 in terms of cost and performance. The speed-up of some of the computational fluid dynamics code, which is the code of these dynamic models—computational fluid dynamics or hydrometeorological—is anywhere between 10, 50 and 100 times. That is already established and those sorts of benchmarks on this new type of technology are already freely available over the internet.

Again, we have to pursue these different technologies now. We should not take the supercomputer away, but we have to ask serious questions. Now we have to invest tens of millions of dollars and six months later the technology is already slower, and we have to solve all these problems faster and faster. So we need to address differently how we put in place, at taxpayer’s expense, this very expensive infrastructure. We have to put these people together—computer scientists, not IT people but computer scientists, and atmospheric physicists, oceanographers and land surface modellers—to model the atmosphere, look at the ice shields and put this all together, and we need to have more access. But you are absolutely right—it is very difficult.

Back when I was studying and working at universities it was basically the universities’ own resources that were used. Now I am running my own business successfully, and we run our own smaller supercomputing facility, so we are not accessing any of these facilities. You can do it smarter, and I do not have $10 million to buy a nice fast computer!

Mr RAMSEY —Of course you have!

Dr Werner —We do not, and you have to do things smarter because you have to stretch the funding, although that does not mean they should be underfunded. We still have this pool of funding, but we need to ask: what do we do and how do we do it? That way we may lower the walls of these silos so that these young researchers can actually gain access. It is very difficult to get funding from the ARC here in Australia, but the attitude should be: ‘We have this brilliant idea; why don’t we give it a go?’ That is how innovations very often come about—with contributions that seem like completely off the wall ideas. We would still go through a peer review process and that sort of thing, but we would be fast-tracking this rather than going through all the different funding. So you would have this idea and you would sit down with your PhD supervisor and say, ‘I would like to try this.’

CHAIR —That came up yesterday—the whole issue of the structure and especially the ARC and the way they can be prohibitive to young scientists who may have different ideas and may think outside the square of what is acceptable thinking. I have listened to you speak. You have raised issues such as the human resource shortage, a diminishing capability and the fact that we once were at the forefront. We have heard before many times, especially in our previous inquiry on research, that Australia seems to be lagging behind where it used to be X number of years ago in a number of fields. Obviously, this area is one of them as well. Is the current establishment an inhibiting factor to encouraging new scientists? Or is it that they are just not available? Do they see no scope for pursuing the sorts of things they want to pursue here?

Dr Werner —It is difficult. Again, what you need is passion and persistence. It is very difficult. When I think back to the early 1990s, it was very difficult. You have to have a very understanding family and believe that eventually there will be a light at the end of the tunnel, so to speak. It is daunting as well when you know what the success rate is. But we have to step back and say, ‘Australia is a small country, so there are limited funding resources.’ I do not know what the entire science budget is—I did not read up on this—but let us assume it is $100 million. It is like a $100 million pie, so you have all these people to feed. You have to buy lots of things. Hardware is very expensive to buy. Any experiment in experimental physics is extraordinarily expensive. Just to teach physics is expensive, as is chemistry. You have to buy all the equipment.

People are going to get disenchanted if they do not have the passion and persistence. At the end of the day, only a few people really succeed and stay in the field, so we have to make sure that this is actually happening. We educated all these people to PhD level. A lot of PhDs get a lot of publications out and then say, ‘No, I don’t do meteorology anymore. I do computer code for an investment bank.’ The talent is wasted. The talent is there, so we have to change a few things and put the passion back. That is a necessity here in Australia because we are hugely exposed to the elements.

Maybe we have to change the funding structure. One of my concluding remarks in my submission was that maybe we should have a special levy or tax where we raise extra funds through the normal taxation system. I know we are already heavily taxed, but maybe it could be just a dollar per person, to raise funds, to raise the level of awareness, and those funds could be put towards the further advancement of long-term meteorological forecasting here in Australia. That is very, very important, rather than saying, ‘Okay, we already have the funding. Now we have to decide what to do with it. We have read the submissions. We have had all these public hearings’—like this hearing right now—’and now we will have business as usual.’ A few things really have to change; otherwise we will just see a repeat of the eighties and nineties and we will not have gained anything. The only thing we will have done is to spend more funding—and we will have a new computer as well. So we can see this levy as a scientific stimulus which leads to a more prosperous Australia in an economical sense as well.

CHAIR —Thank you very much. You certainly have a lot of passion and drive. It has been a very interesting discussion. Dr Werner, thank you very much.

Dr Werner —Thank you very much.

Resolved (on motion by Mr Ramsey):

That this committee authorises publication, including publication on the parliamentary database, of the transcript of the evidence given before it at public hearing this day.

Committee adjourned at 11.49 am