Official Report 245KB pdf
This is the first oral evidence session in our inquiry into flooding and flood management. It will provide a useful context for the remainder of the inquiry. I welcome the two witnesses from the Met Office: Professor John Mitchell, who is director of climate science; and Stephen Noyes, who is director of operations and customer services. Committee members have seen the submission from the Met Office, which we received after issuing our call for written evidence. If the witnesses wish to make an opening statement, I ask that they keep it to a maximum of five minutes. Thereafter, we will ask questions.
Good morning and thank you for inviting us to give evidence to your inquiry. It is good to be here this morning. I will give you a brief overview of what the Met Office does. As you might be aware, the Met Office looks after weather forecasting and warning for the whole of the United Kingdom as well as climate change and predicting future climate for the UK and the rest of the world. The Met Office is world leading in weather forecasting capability and accuracy and in advising Governments globally about climate change.
Is the roof going to stay on?
The very severe winds are due to decrease during the day. As you might expect, the weather over the Highlands is pretty severe with temperatures around -2°C, quite a bit of snow and winds of between 70 and 100mph. The exceptional and awful weather up there will gradually improve as the day goes on. However, I am afraid that it looks like the rest of the week will continue unsettled, with more strong winds, more rain and snow over the higher ground.
Does that mean that my walk at the weekend is off?
Forthcoming questions about specific forecasting ability might or might not relate to this weekend.
I saw the paper but did not have time to read the article.
It was about adaptation to global warming, with particular interest in the food and drinks industry. Of course, that specific issue need not detain us, but I wonder whether you have any comments on one paragraph that relates to the general issue of the weather that we might expect in future. The article states:
I expect that in Scotland there will be an annual mean warming of 1°C to 3.5°C over the next 100 years. Some of that uncertainty arises from the emissions scenario and from uncertainties in the modelling itself.
So the figures that I have quoted are roughly in line with what you have been looking at, with perhaps a slight overestimation in order to make a point.
The figures are slightly towards the upper end.
Okay. That was quite useful.
I thank the witnesses for coming along this morning, and I especially thank Mr Noyes for his impromptu weather forecast. I half expected a map of the British isles to fall down behind him and was quite disappointed when it did not.
The situation in winter is simpler. Almost all the models show an increase in rainfall in mid to high latitudes, with an expected increase in intensity. However, the situation in summer is slightly more complicated. Most models show an increase in rainfall in high latitudes in the northern hemisphere, with a drying around the Mediterranean and southern Europe. The uncertainty arises because the boundary between that increase and decrease in rainfall falls across the middle of the United Kingdom.
What does that mean in layman's terms? To be perfectly frank, I could not quite follow what you said. Are you saying that rain will be less frequent in summer but that when we do get rain it will be heavier?
That is right. It is more complex than—
That is what happened in England this summer.
Yes. Steve Noyes will probably say a bit more about what happened in England in the summer but, in simple terms, it will be wetter in the winter with increased intensity and probably dryer in the summer but also with the possibility of increased intensity, although that is less certain.
Thank you for your written evidence, which is fascinating, and for what you have said so far. I pursue Jamie Hepburn's point about the intensity of rainfall. I will ask about one dimension of that, but I would also like to know what the implications are for public policy. We will consider the impact of climate change on the infrastructure needs of our nation and whether we have the correct systems to cope. When we consider investment, we have to think in the very long term. Investment may need to last for 50, 60 or 70 years.
I will take the second part of your question first because it is probably easier to deal with. Looking at the predictions, I think that it is likely that the climate will continue to warm over the next two to four decades no matter what we do on emissions. Along with that, we will get an increase in precipitation in rainfall and snowfall in winter. You have to take that into account in long-term planning. There is a range of predictions. We expect an increase of anything between 0 and 40 per cent, but that will become a bit clearer when the next UKCIP report comes out.
In that regard, we are looking at a two-spiked problem. The problems are the intensity of summer rain and the sheer additional volume of winter rain. The intensity of summer rain is more unpredictable, but it is potentially more dangerous because of the greater runoff on land that will be dry then. Can you give us a handle on that? Is it 99 per cent certain that that will happen? Obviously, the legislation that we are contemplating would require huge public investment. We need to be certain that the changes are definitely going to happen. As the experts, can you reassure us that they will happen? What event should we try to cater for as the worst-case scenario?
Things are more clear cut in the winter, but there is less certainty in the summer. The previous UKCIP report looked at just one model, which can be misleading. We know that there is some uncertainty, as climate modelling is not an exact science, particularly at a local level. One would have to take a risk-based approach by looking at the range of predictions, trying to understand what the most likely outcome is and then planning through risk management. Modellers will not say, "This is what the change will be," as that would be misleading and probably dishonest. Climate modelling is uncertain, but there are methods of dealing with uncertainty. In terms of risk management and considering what should be done, in particular with regard to the summer, that would have to be taken into account.
Your paper discusses integrated probability modelling. In the simplest terms, are you saying that one-in-200-year events might well happen once every 15 years?
That is the type of information that we can get. However, we have to be aware that the change in average rainfall is a relatively simple problem compared with changes in extremes, both statistically and in terms of physical modelling.
Perhaps I can help with that a little. One way of looking at the problem is that, as Professor Mitchell said, there is a winter issue, which is to do with the volume of rainfall, and there is a summer issue, which is to do with the potential for more intense rainfall events on a more local scale than we may get in winter.
We have not mentioned coastal flooding, sea level rise and inundation. Could you give us some predictions on sea levels and potential coastal flooding?
During the past 100 years, the global sea level has risen about 25cm. The predicted scenarios for the next 100 years range from a rise of about 10cm through 50cm to 75cm, which is the highest level. That is obviously a key contributor to coastal flooding. Again, there is a range of uncertainty in the predictions, but the global mean is typically up to about half a metre. I am sorry to introduce uncertainties again, but that is the way the world is. There is quite a wide range of regional model predictions and that is one area where we do not know what the geographical distribution can be, so it could vary locally by a factor of about three. There will certainly be increases in the sea level, which will increase coastal flooding.
How has the approach in your discipline, in which you have worked for some time, changed? Is the same shift operating in the design and engineering of flood management and flood prevention schemes? It strikes me that, originally, much of the science of long-term forecasting depended on extrapolation from past practice. Now you have moved to a situation in which you can incorporate in the mathematics external factors and systems that you could not include previously. You have been able to feed in an anticipation of the fact that the climate is changing. However, it seems that the disciplines that are associated with civil engineering are still working with an analysis that is based on the historical record—they speak of 100-year and 200-year flooding incidents. They have not been able to feed in and to respond to the new predictive science that you have been able to incorporate in your discipline. Is that a fair comment?
Yes. That is why in my opening statement I made the key point that the past may not be a good guide to the future. When we are dealing with long-term infrastructure, we must look at long-term trends. You referred to changes in science. Until recently, much of the work on climate change was about establishing that there was a human effect on climate. Global models were adequate for that purpose. Now we are looking at adaptation—how we respond to changes in climate—which requires much more local knowledge of climate change. We are pushing climate models to provide that information. I am mindful of the fact that the committee must make decisions about infrastructure today and I advise members to be cautious, because we know that there are limits to models' ability to simulate on a very local scale. We are doing our best to improve the models and are seeking more resources to increase our ability to produce regional predictions. The current science suggests that in winter there will be heavier and more intense rainfall and that we should consider the possibility of phenomena such as more extreme flash flooding in summer.
The predictions that you make are at a broad, generic level. You can make predictions for Scotland and, perhaps, even for parts of Scotland, but I suspect that it is difficult to predict how incidents on the Clyde will affect particular areas of Glasgow, for example. How robust is the modelling that can be used to translate your big picture predictions into more localised responses, perhaps by modelling river flows?
Are you talking about climate scenarios or weather forecasting for flooding?
I am linking the two. How do forecasts of climate change translate into forecasts of floods in general and of the tolerances of flood prevention measures in particular locations?
At present, many climate predictions are made using global models that have a resolution of 150km at best. That is pretty crude for the British isles. To downscale, we take the output from those models and create regional models, which currently have a resolution of 50km down to about 25km. That is still fairly crude in terms of catchment and flood engineering.
Before we discuss forecasting, I want us to continue discussing climate change.
The witnesses say that there will be a meta-analysis of a wide range of studies, in order to allow better predictions. Sea levels are predicted to rise, so there will come a time when we have to say that we will not build new developments below a certain height above sea level. We cannot wall the whole coast. Can you predict what that height will be?
Taking such an approach would be sensible. It is possible to consider a range of predictions and then choose a certain height as a cut-off point. Uncertainty will depend on two things, especially in the longer term. The first is the modelling, which we are working on; and the second is the predictions for emissions, which takes us back to socioeconomic scenarios. Over the next three to four decades, what we do in terms of emissions scenarios will probably make very little difference, so, over the next 30 or 40 years, we will probably be better able to consider the contribution made by climate change than the contribution made by natural variations. That may allow us a way forward.
So you are not in a position to make predictions at this point. Obviously, we do not want to wait 30 years before deciding what the minimum height above sea level for buildings should be, because by that time lots of new buildings would be below sea level.
What you can do is establish a minimum that you believe you will have to work to. However, you should plan in such a way that, if work has to be augmented, hooks have been left in. That would be better than building things and then having to start all over again, which would be more expensive. This is not my area of expertise, but there are ways of approaching the problem that take uncertainty into account.
I have one more general question before we move on to specifics. In paragraph 3 of your written submission, you say that the Met Office
What we are referring to is the ability to link forecasts to impacts. That is relatively straightforward, and the Met Office already produces forecasts of storm surges that can be connected to predictions of tides, for example. We can also forecast waves and overtopping. That allows us to build up a picture of the state of the seas along the coast. The bit that is missing is what that will mean. We have to consider what level of sea is required to cause coastal flooding and inundation, so that we know how to specify alert levels around the coast of Scotland. That has been worked on for a long time in England, specifically down the North Sea coast. Along with the Environment Agency, we can now specify alert levels. What happens is rather more sophisticated than our saying to the Environment Agency, "This is the height of the storm surge." We contact the agency only when we get close to the alert levels, at which point the agency can start to gear up its resources.
I understand that the data are missing, but paragraph 3 says:
Having the data is really important for verifying the performance of models. Without observational data—
I understand that, but either the data are there or they are not.
Yes.
Are the data there?
No. Some data exist in some places, but there are not enough data around the coastline.
What use is investment if we do not have the data with which to do modelling?
I am saying that investment is needed to put the data in place, so that we can improve the models.
But such data would start only from now. Is that what you are talking about?
Yes.
So we cannot fix the lack of historical data—we can do nothing about that. We can only begin the record now in the way that you think is appropriate.
Some data must exist in places. We could gather that together to have a common set of data that we can use.
That is a slightly different issue—you suspect that data exist but that they have not been brought together meaningfully.
That is my suspicion.
But we do not know that for sure.
No. That issue is probably for SEPA.
That is useful.
Two separate issues might be involved. One is that data have been collected but not put in a form and quality controlled in a useful way. The second issue might be—I do not know—that data are needed from more areas than are covered at present.
Thank you—that explains the comment in your submission.
Paragraph 5 of your submission rehearses arguments about weather radars and mentions your commercial arrangement with Scottish Power to improve some of your data. You say that although radar cover for the UK as a whole is sufficient, gaps in coverage exist in parts of Scotland. You highlight a gap in Moray, but the implication is that other gaps exist. What are the Scottish gaps?
Weather radar coverage over the UK meets two requirements. One is for weather forecasting and a supplementary requirement is to aid flood forecasting. For flood forecasting, it is ideal to have weather radar that detects rainfall at a high resolution—typically 1km or 2km—because that corresponds to the size of the weather systems that create the most intense rainfall. That is also the scale on which we look at local topography that leads to flash floods and at river catchment areas. For flooding, capturing good observations and rainfall data from high-resolution radar data is important.
What percentage of the Scottish weather area is covered by radar?
Almost the whole of Scotland is covered by radar for weather forecasting—the exception is the Shetland Isles. Just over 50 per cent of Scotland is covered by rainfall radar that would be ideally suited to high-intensity rainfall events.
Can you give us a comparison between the coverage in Scotland and the coverage in England and Wales? That question follows on directly from yours, convener. I accept that the witnesses might have to give us that information later in writing.
On the basis of the information that I have in front of me, I would say that, with regard to radar that is suitable for rainfall forecasting with a 2km resolution, almost the whole of England and Wales—90 to 95 per cent—is covered, whereas just over 50 per cent of Scotland is covered.
That is quite a big difference.
Forgive me, convener, but I have another question that is specifically on this subject, which is of interest to Peter Peacock. You said that both Moray and Dumfries, which are high-risk flooding areas, would be excluded. Are they not covered?
They are covered with regard to the coarse resolution, but they are not covered as adequately as they could be with regard to the high-resolution radar that would be ideal for high-intensity rainfall events.
We can return to Peter Peacock. The question got slightly hijacked.
I presume that part of the issue of coverage for rainfall events is geographic—the mountains in Scotland are that much higher and we have less flat land and so on. However, that also means that the investment needed to resolve the problem is higher. What needs to happen? If we wanted to improve our accuracy of forecasting of localised weather events and intense rainfall, what would have to happen? Would the Government have to commission such work? Do you provide such work and the Government puts up the cash? Is local government involved? How do we achieve the increase in coverage that might give communities better protection?
And how much would it cost?
The most logical place to start would be to have discussions with SEPA on behalf of the Scottish Government. SEPA is working with the Met Office to identify the areas that are most at risk of flooding. Given that SEPA has the expertise in that regard, it can target the resources to the areas that most need the rainfall radar coverage. It has already indicated in discussions with the Met Office where it would prioritise investment if we could identify the funding. Additional investment in Scotland to provide flood forecasting advice around which SEPA and the Met Office can deliver services is something for the Scottish Government to think about.
If you wanted to provide coverage for the gaps in places where you know there is a high risk of flooding, how many sites would be required in Scotland?
To get you to a consistent level across Scotland, a good starting point would be to add at least three more radar. We would need to do more work with SEPA to understand the effects of radar coverage being blocked by mountainous terrain. That relates to what sites we can access to install radar. If it proved impossible in the Highlands to get an ideal site, there might have to be more than one radar.
In paragraph 4 of your submission, you say that you can run a model at 1.5km resolution and that that was effective during the floods in England last summer. Is it possible to deploy that resource in Scotland?
Yes. We do that already for the whole of the United Kingdom but for different regions, depending on where we expect the heavy rain to be.
I read that one of the constraints is the availability of the supercomputer to enable the model to be operational. Can you explain to the committee the practical constraint with regard to the supercomputer capacity?
The constraint is primarily financial. If we had the money, we would be able to run models for the whole of the UK at much higher resolution than 1.5km. There is a scientific constraint, but the primary constraint that prevents us from doing what we could do is financial. That is often the case with such scientific issues.
Am I right that the roll-out for Scotland will start from 2010, but that you have still to clarify some funding packages before you know the exact situation?
Yes.
So there is a 10-year period for Scotland—can you be any more specific than that?
A 10-year period?
Did you not say that it would be rolled out in the next 10 years?
It will come out in 2010. The life of each supercomputer is typically about five years. There is also a technology driver, in the sense that the manufacturers of supercomputers introduce new technology on a cycle of about four to five years. As such, there is a science cycle, a technology cycle and a funding cycle.
Like others, I will focus on the radar coverage for the warning and reporting of high-precipitation events and your statement, which was a surprise to me, that, while England and Wales are totally covered, only 50 per cent of Scottish land mass is covered. In response to some of my colleagues' questions, there was an assumption that that was to do with mountainous terrain. However, there is mountainous terrain in Snowdonia as well as in the Highlands. What do you consider are the reasons for the disparity in coverage?
It is largely historical. The funding has been made available for quite a long time from the relevant authorities in England and Wales with responsibility for predicting river flooding in particular. The situation predates the formation of the Environment Agency and goes back more than 10 years. Funding has been made available from the authorities that have statutory responsibility for providing flood warnings. They have supplemented the investment that the Met Office has received for weather forecasting, and that has added to the density of the network for England and Wales. That has not been the case for Scotland.
So the disparity predates devolution. You are saying that, before devolution, the coverage was rolled out for England and Wales but not for Scotland.
Yes.
It is worth noting that about 95 per cent of Wales is covered by just two radars. The topography and arrangement of Scotland means that it is a much more expensive proposition.
I have a specific question about your ability to forecast very localised events. It relates to conditions that occurred in Tayside on 22 December—it was a dry and well below freezing night and there were continuing freezing temperatures in the morning, but there was rainfall at the below freezing temperatures, which meant that between 9 o'clock and 9.15, everything turned to ice because the rain was freezing on impact. That was a very brief period of time within which a fairly severe event occurred—I have never seen such an event before, and most people to whom I have spoken have seen it only once before. Such an event is localised not so much geographically, but in terms of time and of the specific conditions that had to prevail to bring it about. Is such an event predictable, or is it not something that you could predict? It was not predicted—we had no warning of it. We had warning of today's high winds, but we did not have any warning of that event.
It is quite difficult to predict automatically from models. The numerical models of the atmosphere can provide information to the forecasters, which would give an experienced forecaster signals that would say that there is a risk of what we call freezing rain events occurring. Freezing rain happens every winter but, as you say, it is very localised and short lived. Across the UK, there are several such events every winter.
To go back to the original subject of flooding, what would your suggestions be? We have you here today to create a sort of Rolls-Royce model of flood warning—what would you suggest to us, as Scottish parliamentarians, as the model of choice? What should we do to give ourselves the best warning of floods? Thereafter, we have to cope with the actuality of the event, but in terms of warnings, what do you suggest?
There is an opportunity for Scotland to take a leading role within the UK to show what could be done. The advantage we have in Scotland is that there are fewer historical responsibilities and not so much infrastructure, so there is a real opportunity for Scotland to move ahead of the rest of the UK. One of the things that can be done in Scotland is investment in weather radar, as we mentioned, but perhaps an even more important thing—and something that could deliver benefits quite quickly—would be to consider setting up a joint forecasting centre for floods, in which organisations such as SEPA and the Met Office could be encouraged to work together on how we can target investment, provide consistent communication to the public and emergency responders in Scotland, and provide early warning of those events. If we were asked to do so, SEPA and the Met Office could set something up fairly quickly—probably within six to 12 months—that would significantly move forward Scotland's capability.
Do you have any estimate or indication of the costs of creating such a structure?
I do not think that the costs would be as significant as those that we were talking about with regard to the weather radar, insofar as capital cost is concerned. We are probably looking at a joint team of somewhere between 10 and 20 people, maximum. Long-term investment might be more significant—in understanding the science; in developing combined meteorological and hydrological models that can provide much earlier warning; in building on the investment that we talked about that is being put into weather forecasting; and in taking that to the next step to integrate that capability with hydrological models, so that we can forecast what will happen on the ground.
Your relationship with SEPA is one of the things that you deal with early on in your written submission. Paragraph 2 highlights the fact that Scotland, as part of the UK, is slightly different from most other countries in Europe because it does not have a single meteorological and hydrological agency. On the one hand, it has the Met Office and, on the other hand, it has SEPA. You have spoken a wee bit about that relationship, but could you give us a bit more detail on how it operates effectively and how data are exchanged between the two organisations? Just tell us in more depth how the relationship works.
The relationship with SEPA is certainly healthy; we have a lot of constructive dialogue. The work that both agencies do on behalf of the public of Scotland, the emergency services and Government authorities is probably constrained primarily by what they are asked to do more than anything else. There is a lot of capability in both organisations and Scotland could benefit more from the capability that they have.
You said that the Met Office and, I think, SEPA are constrained by what they are asked to do. Can you expand on that?
I will take storm surge warnings as an example. In England and Wales, the Environment Agency has responsibility for coastal defence and for providing warnings of storm surge events; that function does not exist in Scotland, as far as I am aware. There is a difference between what SEPA does and what the Environment Agency does. What we can do is limited by funding, so if an organisation or a group of organisations is given responsibility for providing more sophisticated flood predictions and warnings, that must go hand in hand with the provision of funding. We could do more, but additional funding would be required. That is my point.
I will look at the matter from a legislative or statutory point of view, leaving aside for the moment the complications of cross-border legislation. Paragraph 2 of your submission says:
There should be a duty on somebody to provide those services. That could be the Met Office or it could be the Met Office and SEPA working in some joint organisation, but such a duty would be beneficial. The same problem exists in England and Wales at the moment and the same issue has been identified in discussions with the Pitt review and in our evidence to the Environment, Food and Rural Affairs Committee at Westminster.
I want to pin that down. In legislative terms, and looking to future planning and strategy, is absolute clarity required on boundaries and responsibilities and the need to work together and co-operate? If not, surely we may end up again in the situation that you describe in paragraph 21, in which
That would be very helpful.
I seek clarification on whether the statutory changes in respect of the Met Office require to be made at Westminster. The Met Office is, after all, owned by the Ministry of Defence.
Not necessarily, but we will need to clarify that for the committee. If a measure is required specifically for Scotland, and the Met Office and SEPA are required to work together to deliver it, approvals from Whitehall may be needed. However, I am not sure of the exact position. The issue would be worth clarifying.
It would be useful if you could do that. If we decide to look at some of those issues in our final recommendations, it would be helpful for us to be absolutely clear on that.
My question is on the emergency planning exercise. In going through some of the processes—with which I assume you are involved; please tell me if you are not—have you identified where the problems arise in the interface between the various organisations and the clarity of their roles? Do we need a real flood event in Scotland before we can find out where the problems will arise and overlaps will occur, or is it possible to find that out by way of the emergency planning exercise?
That is a useful suggestion in terms of exploring the issues and helping the committee to clarify its eventual recommendations. We participate in exercises. A couple of years ago, we established the role of the public weather service adviser. Pat Boyle, who is sitting behind me in the gallery, is one of the advisers for Scotland. She and her colleagues work closely with the emergency responder community in Scotland and participate in exercises. We also deliver training to that community. It is important that they know what we do and what our messages mean, including in terms of probability and risk. In that way, they can make informed decisions and ask us the right questions about what is going to happen. As the member rightly said, the community is a good one to turn to in terms of bounding the issues and establishing and clarifying needs.
I return to Peter Peacock's point on whether there should be a statutory duty on the Met Office to provide help to SEPA. In terms of working with SEPA, is it possible for SEPA to buy in the Met Office's services? I have read that your being at committee is also something of a sales pitch—I do not mean that horridly.
It depends on what you are asking each organisation to do, given that each of us has a different capability. Certainly, you would want the two organisations to work closely together, but other organisations can add considerable value, too.
That raises another point, which I would like to pursue. In a sense, you are saying that, because you are partly commercial, you have to adjust your activity to whoever is buying the service. Do you have priority clients? Which is the better route to follow to become such a client—commercial or legal?
Government would be the route. There is a separation between our commercial business and our Government business. The Met Office exists to provide services to the citizens and Government of the UK. We happen to have a commercial business bolted on to that, which helps us to offset the overall cost. We are talking about the Met Office acting as a part of Government, delivering services on behalf of Government to the citizens of the UK—and Scotland in particular, in this context.
I wish to raise the issue of what happened in England last summer, which you covered in paragraphs 20 and 21 of your submission. That experience indicated a degree of confusion about who was responsible for what when it came to warnings, the services that became required and the differences between different kinds of forecasting. Your submission states:
The communication problems in Scotland are pretty much identical to those in England and Wales.
So it is likely that the level of confusion that existed in England in July could be replicated in Scotland.
I suggest so, yes.
And that would be resolved by having the sort of joint centre that we have discussed.
Yes.
And that would make a huge difference to how people perceive matters. That is useful.
There are two—possibly three—things that I think we should consider. I am not sure whether this has already been covered, but I cannot see it in our paper. It strikes me, particularly in light of the evidence that we have just heard, that it might be worth hearing from the trunk roads authority and other road authorities, given the surface area that is taken up by tarmac and their responsibility for gully cleaning and so on. Secondly, in addition to planners, there is the house building industry. There is clear pressure from builders to build in certain areas. It might be worth getting house builders' take on the issues. Thirdly, it might be worth hearing from Sir Michael Pitt about what he discovered about organisational issues in England that could help us here. That is a slightly less obvious possibility, but it might be worth thinking about.
A direct approach to Sir Michael Pitt might be useful. We could write to him in the first instance.
That is different from whatever the trunk roads authority is now called—
You mean Transport Scotland.
Yes, and its agencies that deliver services on the roads. Issues such as the cleaning of gullies and culverts are critical.
That is a fair point. We will take that on board and explore how best we can get that evidence. It might be easiest to include such an organisation in the round-table evidence session, but I do not know whether it will be best to hear from Transport Scotland or one of the agencies. We will explore that.
Meeting continued in private until 12:38.
Previous
Subordinate Legislation