House of Commons - Science in the Met Office - Science and Technology Committee

Science in the Met Office - Science and Technology Committee Contents

3 Science to service

The science strategy

17. The Met Office science strategy 2010-2015 outlines the "top-level science strategy" for the Met Office, responding to the increasing demand for "seamless prediction systems across all timescales" for the atmosphere, oceans and land surface.[57] The strategy focuses the Met Office research agenda around four major science challenges:

i. forecasting hazardous weather from hours to decades;

ii. water cycle and quantitative precipitation forecasting on all scales;

iii. monthly to decadal prediction in a changing climate; and

iv. sensitivity of the Earth system to human activities.

The strategy proposes a new research structure within the Met Office, which aims to deliver efficiencies and set in place "mechanisms for greater integration and innovation in the science base". It also advocates "a more strategic approach to partnerships [...], delivery of the necessary infrastructure for research and services, improved processes for staff recruitment and development, and better methods for communicating and disseminating [Met Office] science".[58] These themes are discussed throughout this chapter.

18. Broadly speaking, the science strategy has been very well received across the meteorology community.[59] However, Research Councils UK (RCUK) commented that it "would have appreciated a greater opportunity to be consulted on [the strategy's] development".[60] Professor Julia Slingo, Met Office Chief Scientist, was surprised by this.[61] She explained that there was consultation through the Met Office Science Advisory Committee (MOSAC),[62] and that the director of the National Centre for Atmospheric Science (NCAS)—who was the "most obvious" representative from the Natural Environment Research Council (NERC)—was a member of MOSAC.[63] Professor Slingo was confident that the level of consultation was appropriate for the job of developing a strategy that was suitable for the Met Office, as an organisation that has a specific "public task" to fulfil.[64] John Hirst, Met Office Chief Executive, added that he would follow-up with RCUK its concerns about particular areas of the strategy it would have liked to be consulted on.[65]

19. RCUK also expressed concerns that:

The strategy contains a series of "recommendations" rather than a description of what will be done. In this respect it has something of the feel of an internal recommendation to the Met Office board rather than a set of goals towards which the Met Office is committed. Implementation is only addressed at a very high level.[66]

A formal implementation plan was in fact later published by the Met Office, during the course of our inquiry.[67] Professor Slingo acknowledged that the role of partners, such as NERC, would be "critical" in the implementation phase.[68] The importance of collaboration and partnership is discussed in greater detail later in this report (paragraphs 60-72).

THE NEW RESEARCH AND DEVELOPMENT STRUCTURE

20. The science strategy "recognises the unique position of the Met Office in having world-class weather forecasting and climate prediction in one place".[69] While the Met Office considered it "essential" in the short-term to maintain clearly identifiable programmes in both weather and climate research, it was proposed in the science strategy that integration across these programmes could and should be improved. This would be achieved by:

Bringing together all research and development under a single Director of Science;
Forming a new directorate in Foundation Science; and
Establishing a programme of integrating and innovating activities.[70]

The lack of integration across weather and climate research was noted in the past as a key weakness that "did not readily facilitate common developments" across these areas.[71] Professor John Pyle, Chair of the Hadley Centre Science Review Group and a member of MOSAC, told us that the new structure looked like a "good model but only time will tell".[72] The National Oceanography Centre considered that the new structure may also "foster a closer research partnership" with the broader research community in the UK and abroad.[73] Professor Pyle particularly welcomed the establishment of the new Foundation Science directorate, which he considered might "ease some of the tensions in developing a single Unified Model [...] for both weather forecasting and climate research".[74]

OVERSIGHT OF MET OFFICE SCIENCE

21. MOSAC was set up 15 years ago to oversee Met Office science, following the change to Trading Fund status (see paragraph 4).[75] We took evidence from the Chair of MOSAC, Professor Sir Brian Hoskins, who is also the atmospheric scientist on the Met Office Board.[76] We questioned Sir Brian about whether there was a potential conflict of interest between his role as Chair of MOSAC and as a member of the Met Office board, particularly given that, as the Chair of MOSAC, he is required to report to the Met Office Board.[77] In response, he told us that the current arrangement provided a "very good conduit" from the science to the board, and that as "a pretty independent sort of guy" he felt "no conflict whatsoever".[78]

22. Sir Brian explained that he was responsible for selecting the other members of MOSAC, in consultation with the Met Office.[79] Its membership comprises top atmospheric scientists in the UK and the equivalent of chief scientists from a number of Met Services around the world.[80] Sir Brian told us that while this may look as if MOSAC was "parading [its] programme in front of [its] competitors", he personally ensured that MOSAC got valuable input from all its members.[81]

23. In the past, MOSAC's remit covered only weather prediction.[82] Under the new science strategy, the remit has been expanded to include both weather and climate science.[83] The strategy also states that existing Science Review Groups (SRGs) for reviewing specific areas will "continue for as long as required by the relevant Customer Groups" and that the Chairs of those groups would also be members of MOSAC.[84] The Government said in its memorandum that:

In considering changes to the remit and terms of reference of MOSAC we would encourage the Met Office to reflect on the independence of the Committee and the Principles on Scientific Advice to Government, noting the revised Code of Practice for Science Advisory Committees (CoPSAC) to be published in the Autumn. This identifies best practice guidelines and provides practical advice on the operation of Science Advisory Committees.[85]

We were not clear, however, from the conversation we had with the Chairs of MOSAC and the Hadley Centre SRG, whether they adhere to the principles set out in CoPSAC, although Sir Brian did inform us that they "act in a very independent manner".[86] We are also unclear as to whether either group's terms of reference are published in the public domain as a matter of course.

24. Professor Pyle raised the question about whether Met Office science was being "over-reviewed". He explained that there have been a number of ad-hoc reviews recently and in addition to being formally reviewed by MOSAC, the Met Office is also overseen by the Hadley Centre SRG. Unlike MOSAC, the Hadley Centre SRG is not a Met Office committee; the role given to it by DECC and Defra is to ensure that the Met Office is delivering science that is appropriate to the needs of Government. However, it operates in much the same way as MOSAC, with a rotating membership of scientists from the UK and overseas. While the Hadley Centre SRG deals with work under the climate programme, Sir Brian indicated that the equivalent in weather was the Public Weather Service Customer Group (PWSCG) (see paragraph 13). He added that PWSCG representatives were present at MOSAC discussions and were sent a copy of the MOSAC Chair's report.[87]

25. Given the move towards integrating weather and climate science, and with the Met Office Science Advisory Committee's (MOSAC) remit being expanded to include both areas, we question whether it is sensible to impose additional scrutiny by the Met Office Hadley Centre Science Review Group (SRG). We recommend that the Met Office consult with DECC and Defra to determine whether the Hadley Centre SRG is required in its current form. Our view is that it would be more sensible to formally review all science under MOSAC, whilst retaining a Hadley Centre Climate Programme Customer Group, as described in paragraph 14, to ensure that customer needs are being met.

26. We recommend that the Met Office publish MOSAC's terms of reference on its website. We also advise MOSAC to consider the Code of Practice for Science Advisory Committees (CoPSAC) at its next meeting, specifically considering whether MOSAC would benefit from adhering to the principles contained within it.

Weather and climate forecasts

27. Creating weather and climate forecasts is a complex process involving the application of science and technology to predict future atmospheric conditions from observations.[88] These observations are recorded around the world, from a variety of sources (from land, at sea, in the air and from space). Each day the Met Office receives and uses approximately half a million observations. This includes data on temperature, pressure, wind speed, wind direction and humidity. The process of 'data assimilation' then uses these observations to provide a "best estimate of the current state of the atmosphere".[89] How this state evolves over time is then calculated using a computer model, producing an estimate of the state of the atmosphere at some point in the future—a 'forecast'.[90] Forecasting involves making billions of mathematical calculations; therefore powerful supercomputers are required in order to carry out these calculations as quickly as possible.[91]

OBSERVATIONS

28. The importance of observational data in modelling future weather and climate is clear.[92] Indeed, the Government told us that it:

recognises that the robustness of Met Office models is contingent on the accuracy and adequacy of supporting observational data. Observations directly input to models and to model development (through enhancing our scientific understanding) and are the only means of verifying model outputs.[93]

However, the UK generates and owns less than four per cent of the observational data on which it relies to deliver the Public Weather Service; this value drops to less than one per cent if satellite data are included.[94] As a result "international collaboration is essential to provide the observations on which the Met Office depends".[95] This is a subject we will return to in paragraph 68.

29. In addition to the vast quantities of data available from international sources, there is also a network of volunteer observers in the UK. Stephen Burt, a freelance science writer and Fellow of the Royal Meteorological Society, told us that the voluntary observing network has declined by approximately 50 per cent: in 1975, there were 6,220 rainfall observing sites in the UK; by 2010, the figure had dropped to 3,214.[96] He attributed this decline, in part, to "very little ongoing support from the Met Office".[97] We understand, however, that the Met Office is now beginning to address this concern. During our visit to Exeter, Met Office officials demonstrated the new Met Office Weather Observations Website (WOW), which "is helping to co-ordinate the growth of the weather observing community in the UK, by asking anyone to submit the observations they are taking".[98] The Met Office WOW includes guidance on setting up a weather observation site and submitting observations online.

MODELLING

30. The Met Office uses essentially the same "Unified Model" (UM) for modelling "across all timescales from daily weather forecasting to centennial climate change predictions, and for all space scales from the local to the global".[99] Many of the organisations we heard from regarded the Met Office's models very highly.[100] The European Centre for Medium-Range Weather Forecasts (ECMWF) told us that there were "extensive objective international comparisons carried out continuously regarding the skill of global numerical weather prediction models from the Met Office and [others]. The ECMWF and the Met Office models are in the world-leading category".[101]

31. A testament to the quality of the Met Office's models is that the UM is licensed to other national meteorological services for operational use. Current users include: Norway, Australia, South Korea, South Africa, India, New Zealand and the US Air Force.[102] The Australian Bureau of Meteorology wrote to us explaining that it was "particularly well informed to be able to comment on the point of model robustness, having done a deep 'due diligence' on the Met Office and its modelling system before seeking a collaborative arrangement".[103] The Bureau explained that its decision was "strongly driven by [its] assessment that the Met Office systems are state-of-the art in modelling" and that it had "no reservations in stating that its decision has been completely vindicated by its first-hand experience operationally with the Unified Model over more than two years. These are excellent models, well-conceived, well built, and well up to the task".[104] As a result of adopting the UM, Australian forecasts have improved.[105] The accuracy of UK forecasts is discussed in paragraphs 45-51.

32. In addition to the clear benefit to those countries using Met Office models, there are also considerable benefits for the Met Office itself. For example, Professor Ed Hill, from the National Oceanography Centre, told us that by operating the UM in different countries, with different weather and climate environments to our own, Met Office models "get tested in different regimes".[106] He added that more people using the models would allow users to learn lessons collectively.[107] The Australian Bureau of Meteorology indicated that feedback of this kind "facilitate[d] model improvements that would otherwise be beyond the focus of the Met Office".[108] The Bureau described it as:

"one plus one equals three" territory, where a strategic alignment of the intellectual capital of British and Australian meteorologists in the use and development of the UM yields added benefit in terms of improved forecasting capability to the citizens of both nations.[109]

33. As well as looking at the use of Met Office models by international partners, we were keen to find out more about the extent to which the models were used by the wider meteorology community within the UK. We heard conflicting views about this. The Department of Meteorology at the University of Reading stated that:

The whole suite of models used by the Met Office is used within the University community, and particularly at the University of Reading, in research projects. This means that the models are subjected to a very high level of scrutiny, often in ways unanticipated by the development teams at the Met Office. This level of scrutiny substantially increases the robustness of the models.[110]

However, Professor John Pyle, Chair of the Hadley Centre SRG, told us that:

The Unified Model architecture is often difficult to negotiate and many UK academic meteorologists employ other, easier-to-use numerical models for specific research projects (e.g. mesoscale modelling). The Met Office needs to consider this seriously. I believe the intellectual exchange with the academic community will be increasingly important for the Met Office; if instead, UK academics use other, rival models, this will certainly be to the detriment of the Met Office, and UK science, in the medium and longer term. More thought needs to be given to making their models "user friendly".[111]

Professor Sir Brian Hoskins, Chair of MOSAC, explained that one of the reasons for the Met Office model not being particularly user-friendly might be that it has historically been used mostly for operational purposes. He added that there would be "an overhead" associated with making it more usable by the community.[112] John Hirst, Chief Executive of the Met Office, accepted that the models needed to be easier to use.[113] It was his view that collaboration on modelling had historically been better in climate forecasts than in weather forecasts but that the Met Office was "now working with the weather academic community" on this issue.[114]

34. Throughout the course of our inquiry, we heard of one collaborative computer modelling initiative that was working particularly well. The Met Office and Natural Environment Research Council (NERC) joint supercomputing system, "MONSooN", allows scientists to collaborate on research into a number of modelling issues.[115] MONSooN has been credited with greatly improving the extent to which the Met Office collaborates with the wider research community.[116] For example, the National Oceanography Centre told us that MONSooN has been used by scientists from the Met Office and NERC to analyse and improve the common ocean modelling system, "NEMO", which is now a "world-leading system".[117] Professor Pyle welcomed MONSooN and stated that he would like to see it expanded.[118] He acknowledged, however, that this would require more money.[119] Mr Hirst told us that the Met Office would expand MONSooN, and was "already in discussions with NERC about the next phase".[120]

35. Met Office models are highly regarded across the UK and around the world. It is a testament to the Met Office that its Unified Model is licensed to other national meteorological services. Collaboration with these international partners helps the Met Office to further test and develop its models and should be encouraged. Similarly, collaboration with the wider UK meteorology community should be encouraged to stimulate the development of Met Office models. We note that the MONSooN project has been held up as a particularly good example, providing a joint supercomputing system that allows scientists to collaborate on research into modelling issues. We encourage the expansion of MONSooN and recommend that NERC work closely with the Met Office to develop plans for the next phase that are suitable for the research community's needs.

SUPERCOMPUTING

36. In addition to using supercomputers to facilitate collaborative research, they are also used for the operational delivery of forecasts. As explained in paragraph 27, forecasting uses powerful supercomputers, these are computers that are optimised to make billions of mathematical calculations as quickly as possible. The Met Office Science Strategy explained that:

The difference between operational and research computing requirements needs to be recognised. Operational delivery requires the appropriate capacity to deliver a suite of weather forecasts on a 24-hour, 7-day a week basis, without interruption. Increasingly it will also need to accommodate an operational suite of climate predictions. Consequently, operational supercomputing needs to be robust and under our control, and it needs a substantial partition for preoperational development and testing. [...]

Research, on the other hand, requires access to advanced computing capability in order to make further progress in model resolution and complexity, data assimilation and process-based research.[121]

37. In 2010 the Government Chief Scientific Adviser, Professor Sir John Beddington, published a review of government's needs for climate science advice. The review recommended that "a step-change increase in supercomputing capacity [...] would be required to most effectively meet the Government's key evidence and advice needs".[122] More recently, the House of Commons Transport Committee also stated that benefits would be realised if an additional £10 million funding were made available for supercomputing resources.[123] Many of the witnesses we heard from agreed that such a step-change in supercomputing capacity was necessary.[124] However, Sir John recognised in his review that this step-change would involve a four-fold increase in supercomputing costs, which was "not currently affordable".[125] While others also recognised affordability was an issue,[126] the point was made that the Met Office had already "slipped down the league table in terms of its computing resource" and that it would be "impossible to deliver world class weather and climate science without access to adequate computing capacity".[127] However, we have not in the course of our inquiry assessed historical investment in supercomputing resources. The Met Office told us that it currently:

has developments available which have been demonstrated in research-mode to deliver more accurate forecasts. However, it is not possible to implement these improvements in the Met Office's operational forecast model because of limited supercomputing resource.[128]

In the Met Office Chief Scientist's words, "the science is ready and waiting" and as a result, there would be "a very rapid return" on investment in supercomputing.[129] We asked the Met Office to explain what operational improvements it could deliver if additional investment in supercomputing was made. In response, it provided case study examples of how enhancements would deliver improved advice to users, and affect their response in severe weather.[130] For example, intense downpours caused localised incidents of surface water flooding in parts of Dorset on 18 August 2011, resulting in the Fire Service dealing with over 100 incidents in a two hour period.[131] The Met Office told us that the short lead-time warning and low confidence extreme rainfall alert were "very likely to have contributed to limited preparedness".[132] Enhanced supercomputing power "would probably have allowed more confident warnings, better indications of possible peak rainfall intensities, and longer lead time information on the potential risk, to be issued".[133] Other case studies highlighted by the Met Office included the snowstorms in the South of England in February 2009 and the Cumbrian floods in November 2009.[134] In addition to improved short-range weather forecasts, increased supercomputing capacity would also improve operational monthly to decadal predictions and climate services.[135]

38. As well as producing case studies, the Met Office has also calculated the socio-economic benefit delivered by investment in supercomputing capacity. This was carried out in 2008, "in accordance with best practice as set out in the HM Treasury's Green Book: Appraisal and Evaluation in Central Government".[136] In compliance with these best practice policies, it was estimated that "for a £50 million five year whole-life cost, net UK socio-economic benefit totalling £0.5 billion would be delivered through provision of enhanced weather and climate services".[137] John Hirst, Met Office Chief Executive, told us that this ratio, a ten to one return on investment, "still exists" today.[138] This aggregate benefit ratio was calculated by "combining the PWS (13:1) and climate science (6:1) ratios", suggesting that investment in supercomputing capacity would provide greater benefits for weather forecasting than for climate forecasting.[139] The Government told us that "further economic benefit as a consequence of ongoing advances in the science will be delivered through the routine scheduled replacement of the current supercomputer currently planned for 2015".[140] However, the Royal Meteorological Society considered that "a further significant investment in computing resources is required, over and [above] the current commitments".[141]

39. The Met Office wrote to us to explain that delivering improvements, consistent with the socio-economic benefits outlined above, "would require a supercomputer with at least twice the capacity of the near one petaflop[142] facility now being implemented at the Met Office".[143] The cost of this—including associated infrastructure, depreciation, power, service and maintenance charges, and staff costs for developing modelling infrastructure—would be £14 million per year, over each of the next three years.[144] This was consistent with estimates from Professor Sir Brian Hoskins, Chair of MOSAC, and Professor John Pyle, Chair of the Hadley Centre SRG, on the required level of investment in the future.[145] Edward Davey MP, the Minister for Employment Relations, Consumer and Postal Affairs, acknowledged that "a very good case" had been made for increased supercomputing capacity and that BIS was "building a business case" for the next generation of supercomputing capacity.[146] He added that this was "happening with a degree of urgency".[147] However, we later discovered that "the current timetable sees this process taking up to 18 months".[148]

40. It is of great concern to us that scientific advances in weather forecasting and the associated public benefits (particular in regard to severe weather warnings) are ready and waiting but are being held back by insufficient supercomputing capacity. We echo the recent conclusions of the Government Chief Scientific Adviser and others, that a step-change in supercomputing capacity is required. We acknowledge, however, that affordability is an issue. The Met Office has over recent years built a good case for increased investment. However, we have not in the course of our inquiry assessed investment in supercomputing over recent decades. We recommend that the Met Office provide an overview of historical investment in supercomputing resources in its response to us. We encourage BIS to complete a formal business case on supercomputing, however, we do not consider that this process should take anywhere near the 18 months suggested by the Government. In our view, the Government should finalise the business case in the next six months.

41. Given the current economic climate, we considered it prudent to discuss with the Met Office and other witnesses what low-cost options there were for increasing supercomputing capacity. The Government told us that Sir John's review recognised "the need for greater collaboration on supercomputing resources, including internationally, stressing that long term development of modelling capability would likely require a European solution".[149] However, Professor Sir Brian Hoskins, Chair of MOSAC, warned us that Japan and Korea have their supercomputing capacity now, while there was still no sign of a European solution within the next few years.[150] In any case, a European solution would be more suitable for collaboration in climate forecasting than in operational weather forecasting, which is carried out to very tight timescales.

42. Professor Julia Slingo, Met Office Chief Scientist, suggested that collaboration on supercomputing "has been looked at over many years by the international community in weather and climate science".[151] She explained that the nature of some problems does not lend itself to, for example, distributed computing (the use of multiple computers communicating through a network). Professor Slingo explained that in weather forecasting, models have to run very efficiently on supercomputers and also gather and process huge amounts of data, all within a very short timeframe. This requires a specific type of machine architecture as well as a very large bandwidth to get the data out of the machine and on to a huge data archive. Professor Slingo added that the Met Office looks at this issue every year.[152]

43. The Met Office told us that "remote supercomputing options, such as third party facilities, grid computing and cloud computing, [were] not suitable at the current time".[153] While Sir John, in his recent review, agreed that Earth system and high-resolution models could only be run on supercomputers, he also stated that "in a limited number of instances grid or network computing may offer a viable and cost-effective approach, such as for low resolution ensembles".[154]

44. Given that supercomputing capacity for weather and climate forecasting is a recurring issue, we recommend that the Met Office work with the Research Councils and other partners in the UK and abroad to develop a ten-year strategy for supercomputing resources in weather and climate. This should include an assessment of which areas in weather and climate research and forecasting might benefit from low-cost options to enhance supercomputing capacity.

ASSESSING FORECAST ACCURACY

Weather

45. The meteorology community tracks collectively the accuracy of forecasts.[155] The Met Office told us that:

The accuracy of Met Office forecasts are evaluated against observations on a daily basis. The PWSCG [Public Weather Service Customer Group] specify accuracy targets for forecasts of maximum and minimum temperatures, rain, sun, wind speed and wind direction. In 2010/11 all targets were met. As of August 2011, on average (over a 36-month period) the percentage of forecasts accurate to within ± 2 °C is:

87.6% of maximum temperature forecasts on the day the forecast is issued (target for 2011/12 85%) and 78.5% of minimum temperature forecasts (target 76.5%);
81.1% of maximum temperature forecasts on the second day of the forecast (target 79.5%) and 71.7% of minimum temperature forecasts (target 69.0%).[156]

The PWSCG also "routinely undertakes public perception surveys to assess satisfaction with the forecast and warnings service".[157] In November 2010, it found that "nine out of ten people found weather forecasts useful and just over three quarters found them accurate".[158]

46. Professor Alan Thorpe, from the European Centre for Medium-Range Weather Forecasts (ECMWF), told us that "the distance that you can predict into the future has been advancing at about a day per decade".[159] For example, a five-day weather forecast today is as accurate as a three-day forecast was 20 years ago and a three-day forecast today is as accurate as a one-day forecast was 20 years ago. The Met Office attributed this increase in skill to "more sophisticated atmospheric physics, higher model resolution and more comprehensive observations, especially from meteorological satellites".[160] Professor Thorpe explained that as a result, over the years, more "local specificity" was available in forecasts.[161]

47. The ECMWF told us: "All global weather prediction models are routinely evaluated by the World Meteorological Organisation using independent and objective measures of skill".[162] The Met Office added that "a range of metrics are used and all show that the Met Office is consistently within the top three centres internationally".[163] However, similar metrics are "not yet available for longer range forecasts" because the appropriate methodologies are not in place and also because "verification statistics are much more limited due to the short length of the observational base".[164] The Met Office explained that "the quality of its performance against other centres is assured by including the Unified Model (UM) in all model comparisons and in the European Seasonal to Inter-annual Prediction (EUROSIP) ensemble of models"[165]—EUROSIP is a project that aims to strengthen collaboration on seasonal forecasting.[166] There is, however, a "common public perception" that the Met Office does not provide reliable seasonal forecasts.[167] The National Oceanography Centre (NOC) told us that this was "largely due to sensationalist media reporting and shortcomings in how 'probability' and 'risk' are understood by non-experts".[168] These are issues we discuss in paragraphs 52-56. The NOC added that:

Private weather forecasting companies are now often called upon to make these seasonal predictions, suggesting that this is an aspect of the Public Weather Service remit where the Met Office service could be improved. The accuracy of forecasts by these private companies needs to be carefully evaluated on a long-term basis.[169]

48. The accuracy of independent forecasters in comparison to the Met Office is an issue that is also of interest to the BBC. Roger Harrabin, BBC Environment Analyst, told us that BBC News initiated the Weather Test "to compare publicly for the first time the accuracy of weather forecasters in the UK."[170] The Met Office is co-operating with this initiative.[171] However, John Hirst, Met Office Chief Executive, pointed out that this was "not a trivial exercise".[172] Professor Julia Slingo, Met Office Chief Scientist, added that particularly with seasonal forecasting, which are probabilistic forecasts (forecasts that assign a probability to each of a number of different outcomes in order to allow uncertainties to be quantified), there is no right or wrong forecast and therefore a whole history of forecasts is needed to decide the level of skill and reliability.[173] Professor Slingo explained that the international community was still working out how to do this in a way that makes sense.[174]

Climate

49. We were also keen to know how accurate climate forecasts have been over the past few decades, and whether they were improving. We note that the climate model did not accurately predict the extent of the flattening of the temperature curve during the last ten years.[175] Professor Alan Thorpe, from ECMWF, told us that:

In 1990, when the scientific assessment was made, there were real-time predictions of what the climate, subsequent to 1990 going forward, would be. We are now in a position of having a record of what actually happened relative to the predictions that were made then of the climate from 1990 to the present time. Those comparisons show that the models of the day—of course, the models have improved since then—if anything, under-estimated the amount of global warming that has subsequently happened. We are able now, because we have done this climate prediction for a number of years, to start to assess that.[176]

Professor Paul Hardaker, from the Royal Meteorological Society, agreed that "what the early models predicted is largely what has come to pass in terms of our observations".[177] Mr Hirst added that there was "a difference between making a forecast for tomorrow when you experience tomorrow very quickly" and "going back and modelling how the climate has evolved in history to make sure that our models replicate what actually happened".[178] The Royal Meteorological Society added that there was "no more computationally complex problem in science" than simulating the climate.[179] Despite the difficulties in assessing climate models, both the Royal Meteorological Society and the ECMWF considered that the Met Office was widely acknowledged as a world-leader in climate modelling and prediction.[180]

Atmospheric dispersion

50. The Met Office uses its Numerical Atmospheric-dispersion Modelling Environment (NAME) model to "predict how material will be dispersed in the atmosphere and deposited on the ground".[181] In recent years, NAME has also been used for events such as the 2005 Buncefield oil storage depot incident, the 2008 Bluetongue outbreak over Europe, the 2010 and 2011 volcanic eruptions, and the 2011 Fukushima nuclear incident.[182] The Met Office told us that it was difficult to verify the accuracy of dispersion models "because dispersion events occur infrequently and it can be difficult to obtain reliable, quantifiable observations of the distribution and concentration of material".[183] This was an area of particular interest to us following our recent report, Scientific advice and evidence in emergencies, in which we questioned the suitability of the Met Office's dispersion predictions in relation to volcanic ash.[184] In response to our report, the Government told us that an independent review of the NAME model had been commissioned by the Civil Aviation Authority.[185] This review concluded that NAME "represents a state of the art dispersion model".[186] In its submission to us, the US National Oceanic and Atmospheric Administration National Weather Service agreed, stating that "NAME is robust, peer reviewed, and good for ash dispersion modelling".[187] We note that the Met Office continues to work with its American counterparts to improve collaboration during future volcanic eruptions.[188]

51. The Met Office is consistently within the top three centres internationally in weather prediction and is widely recognised as a world-leader in climate prediction. However, we note that the climate model did not accurately predict the extent of the flattening of the temperature curve during the last ten years. We have heard that the accuracy of short-term forecasts is easier to assess than the accuracy of longer term forecasts and infrequent events, such as volcanic ash dispersion. We encourage the Met Office to work with partners in the UK and internationally on developing metrics to assess the accuracy of longer-term forecasts of weather and climate and of forecasts based on infrequent events.

COMMUNICATION OF FORECASTS

52. A forecast, however accurate it is, is of little use if it is not communicated well and understood by the customer.[189] The National Oceanography Centre told us that most of the public perception of forecasts is via television broadcasts.[190] Professor Paul Hardaker, from the Royal Meteorological Society, considered that it was the responsibility of both the forecaster and the broadcaster to ensure that forecasts are accurately communicated.[191] Given the inherent uncertainty in forecasting, one of the ways in which communication could be improved would be through providing more information about probability. Professor Brian Hoskins told us that MOSAC has "always encouraged the Met Office to produce more information about likelihood" but that "media pressure" on weather forecasts meant that there was only a short period of time to communicate forecast information.[192] Sir Brian did not consider that there was an easy way to get this information across in a 30 second broadcast.[193] However, the use of the Internet and digital technology, such as the BBC's "red button" facility, means that more detailed forecast information could easily be made available to those who want it.[194]

53. Professor Hardaker told us that "many countries make much greater use of probabilistic information in their forecasts than we do, even in their broadcasts".[195] Sir Brian cited a recent example:

Let us take [...] the hurricane that was likely to inundate New York. US television was showing 12 possible tracks provided by 12 different models for the hurricane. I do not believe that that sort of information is difficult for the public to assimilate, and we should not underestimate the public's ability to take on odds and make their own decisions based on those.[196]

We asked the Minister whether he considered that the communication of such detailed information could be improved in the UK. In response he told us that the Met Office was considering the matter.[197] The trade union, Prospect, confirmed that the Met Office "has recently invested significantly in enhancing the presentation of site-specific information on its public website".[198] The Met Office is currently testing these changes on the beta version of its public website.[199] Prospect added that the Met Office "is also seeking to enhance the presentation of probabilistic weather forecast information".[200] The Royal Meteorological Society told us that "if the Met Office was able to provide more information about uncertainty in its forecasts, it may be less subject to the criticism it has seen from time-to-time from public and media alike".[201]

54. An area where the Met Office has been particularly heavily criticised in the recent past is in seasonal forecasting, for example, during the bad weather that followed its 2009 prediction of a "barbecue summer".[202] Philip Eden, a Fellow of the Royal Meteorological Society, told us that predictions like this "go to the very heart of what weather forecasts are for"; he questioned whether they were a public service, or entertainment.[203] Nick Baldwin, Chair of the Public Weather Service Customer Group (PWSCG) told us that following the barbecue summer prediction, the PWSCG was:

heavily involved in a discussion about withdrawing the previous seasonal forecasting approach. The consultation we undertook showed that people did not find it very useful in the way it was presented, and that they would rather have received a shorter-term forecast so that the three-month forecast was replaced with a 30-day rolling forecast. A lot of work has gone on since then with the Met Office, and over the next week or so it will introduce a new seasonal forecasting methodology for civil contingency communities, which includes a better explanation of the uncertainty facing us. [...] It is important that people are organised and have a good understanding of that forecast. We have been funding that information and it will be released through the Cabinet Office.[204]

While it may be the case that seasonal forecasts with an accompanying explanation of the uncertainty are more useful for civil contingencies than, for example, the general public, the Met Office is bound by clear rules that state that if it makes this information available to the Government, it must also be made available publicly.[205] It was also suggested that seasonal forecasts were useful across a wide range of industries, for example, insurance, power generation, construction, agriculture, tourism and retail.[206]

55. The Met Office should continue to produce longer term ("seasonal") forecasts as they are useful for civil contingencies and a wide range of industries. These forecasts should always be communicated carefully and accompanied by explanatory notes describing the uncertainty. We recommend that the Met Office develop a communications strategy that sets out, for example, how it intends to enhance the ways in which it presents probabilistic weather forecast information.

56. The Met Office should also work closely with broadcasters, such as the BBC, to ensure that forecasts are communicated accurately. In particular, we are keen to see broadcasters make greater use of probabilistic information in their weather forecasts, as is done in the United States. Broadcasters should also make more use of digital technology to ensure that probabilistic forecast information is available to those that want it.

ACCESS TO DATA

57. In generating weather and climate forecasts, the Met Office generate huge quantities of data. The Government told us that:

The PWSCG require the Met Office to provide a meteorological library and archive service available to anyone with an interest in the weather or climate and an approved place of deposit for meteorological information under the public records Act (1958). [...] There is also a legal requirement handed down to the Public Weather Service from the Lord Chancellor's office to archive meteorological data on behalf of the UK Public.[207]

We heard contrasting views on how well the Met Office is meeting its requirement to provide public access to library services and to historical data. Among those who thought the Met Office was doing well were Research Councils UK and the National Oceanography Centre;[208] while the Committee on Climate Change and others appeared to be less satisfied with the current arrangements for accessing data from the Met Office.[209] Issues raised by the latter group included that the Met Office charges heavily for access to electronic records and that while some older records are available in paper form for photocopying, this was not the case for more recent data, which often originated in digital format.[210] Nick Baldwin, Chair of the PWSCG, told us that the Met Office was looking at ways in which to expand the data that were publicly available but that this needed to be done in a cost-effective way.[211] Mr Hirst added that the Met Office tried to address every specific request or issue that was raised with it on data accessibility.[212] In order to gather views on how historical data could best be provided online, we heard that the Royal Meteorological Society and the Met Office were running a consultation with the wider community.[213] It was also suggested during the course of this inquiry that the Met Office could learn from other countries—including the United States and Australia—where data were considered to be more freely accessible.[214]

58. We heard from a number of witnesses that freeing up access to data might help to grow a more vibrant private sector that could, for example, develop specialist weather and climate services.[215] This view was shared by the Minister.[216] Driving innovation and growth by freeing up public data is one of the main aims of the Government's plans for a Public Data Corporation (PDC).[217] The Met Office has been consulted by the Government on the development of the PDC.[218] Further details on the PDC were outlined following the 2011 Autumn Statement by the Chancellor of the Exchequer, when the Government announced that it would establish a Data Strategy Board and a Public Data Group that would "maximise the value of the data from the Met Office" and other organisations. The Government stated that the release of Met Office data under the Open Government Licence represented "the largest volume of high quality weather data and information made available by a national meteorological organisation anywhere in the world".[219] However, the possibility of the PDC adding "an unnecessary and unhelpful level of bureaucracy" was raised by Professor Hardaker, from the Royal Meteorological Society, as a potential problem.[220]

59. We note that there are contrasting views on how easy it is to gain free access to Met Office data. While we take some reassurance from the fact that the Met Office tries to address specific concerns about this as and when they arise, we consider that the current consultation in collaboration with the Royal Meteorological Society on access to data should help the Met Office to deal with the problem in a more strategic manner. We recommend that the Met Office also look to other countries for best practice on making data more freely available. Alongside this, we welcome the Government's initiative under the Public Data Corporation to make more Met Office data available to drive innovation and growth. The Government should continue to work with the Met Office to ensure that the new arrangements are effective and do not add an unnecessary level of bureaucracy.

Collaboration and partnership

60. Throughout this report, we have touched on the subject of collaboration between the Met Office and the wider meteorology community, both in the UK and abroad. In this section we look in more detail at the ways in which the Met Office collaborates with the research base, with the Government and other public bodies, and with international partners.

RESEARCH BASE

61. Professor Sir Brian Hoskins, Chair of MOSAC, told us that "collaboration in the UK was quite rocky 20 or 30 years ago" but that this had improved significantly.[221] The Met Office now appears to be well connected to the research base.[222] John Hirst, Met Office Chief Executive, told us that a "head of [science] partnerships" post had been created and that the Met Office was acting on the need to collaborate.[223] The Met Office stated that in 2010, its scientists served on 39 committees related to the UK research base and that representatives of the UK research community were included on MOSAC and the Hadley Centre SRG.[224] The Met Office's Chief Scientist is also a member of the Natural Environment Research Council's (NERC) council.[225] The Met Office's science strategy recognises the importance of collaboration and outlines the activities that the Met Office will undertake.[226]

62. One activity we heard a lot about during the course of our inquiry was the Joint Weather and Climate Research Programme (JWCRP).[227] The JWCRP is a joint programme between the Met Office and NERC which aims "to ensure that the UK maintains and strengthens its leading international position in weather and climate science".[228] Professor John Pyle, Chair of the Hadley Centre SRG, told us that while the intentions of JWCRP were "excellent", it would be "foolish to underestimate some of the practical difficulties".[229] He explained that such a collaboration, on an institutional level, would "entail some loss of sovereignty [and] effective management of joint programmes [would] be a challenge".[230] He implied that progress with the JWCRP was understandably moving slowly,[231] but that a more strategic approach was now being taken.[232] Other activities include the Met Office's work with NERC on the cross-Government, cross-Research Council programme, Living with Environmental Change,[233] and its collaborative relationship with individual universities.[234]

63. The Government told us that it fully endorsed the Met Office's proposal for stronger partnerships and collaboration; however, it added that "the proposed science partnerships should also include representation from government to provide additional context to proposed research programmes".[235] In response, Mr Hirst told us that he didn't understand the thoughts behind this and that the Met Office needed to understand precisely what these representatives would want to do.[236] Professor Julia Slingo, Met Office Chief Scientist, noted that the Met Office needed to be careful, particularly with its academic partners, that this did not conflict with the Haldane principle (the traditionally accepted view that detailed decisions on research should be made by researchers, not government).[237] The Haldane principle does not apply to research funded by the Met Office; it does, however, apply to research that is funded by the Research Councils.[238] The JWCRP, for example, is jointly funded by the Met Office and NERC. Government representation on the JWCRP might therefore be questionable. We sought clarification from the Minister; he explained that "the suggestion that the Government should be represented on the science partnerships was, frankly, to ensure the links between the Government and policy-relevant research and that any potential research overlap is minimised" and added that he did not believe there was a conflict with the Haldane principle.[239] The Government subsequently wrote to us clarifying that it did "not necessarily envisage Government representation in all scientific collaborations engaged in by the Met Office" and that where representation was desirable, it would most likely be at the Chief Scientific Adviser or official level.[240]

64. We recommend that the Government consult with the Met Office on the need for Government representation on Met Office science partnerships. While such representation may be desirable to ensure strong links between the Government and policy-relevant research, care must be taken to ensure that there is no conflict with the Haldane principle—particularly where partnerships are co-funded by the Research Councils.

GOVERNMENT AND PUBLIC BODIES

65. The Met Office works with the Government and public bodies to deliver the operational services required by the public sector under a range of contracts (see paragraph 9). For example, under its Public Weather Service contract, the Met Office produces severe weather warnings and works with the emergency planning and responder communities (see Annex A). This is an area in which it is especially important to have good coordination with the Government and public bodies. Prospect told us that Met Office links with the emergency responder community had "greatly improved" in recent years. As a result, the Met Office has a better understanding of "how improved forecasts and warnings can help to mitigate some of the impacts from severe weather".[241] The Government added that "the Met Office [had] also developed the 'traffic light' system of four colours which highlights the weather maps and advisories it sends out. This simple system alerts emergency planners and the public to the level of risk and certainty of the weather event".[242]

66. A good example of effective collaboration that was raised by a number of witnesses was the Flood Forecasting Centre (FFC), which is a partnership between the Met Office and the Environment Agency, staffed by members of both organisations. The FFC was set up in 2009 following the report by Sir Michael Pitt, Learning lessons from the 2007 floods, known as the "Pitt Review", which recommended that the "Environment Agency and Met Office should work together, through a joint centre, to improve their technical capability to forecast, model and warn against all sources of flooding".[243] Professor Paul Hardaker, from the Royal Meteorological Society, told us that it was important for the Met Office to be effectively joined up with other agencies that were involved in issuing warnings.[244] The Government stated that the 'traffic light' system "is now used across all early warnings distributed by the Met Office, Flood Forecasting Centre and Environment Agency".[245] Professor Hardaker considered that while it was still "early days", the FFC appeared to be working well.[246] Mr Hirst told us that the Cumbrian floods occurred just after the FFC had been established and as a result, the Met Office and Environment Agency "were able to give 24 hours' better notice than we had ever been able to do for a level of rainfall that was beyond any historic record of rainfall in this country".[247] Other examples of collaboration include the relationship of the Met Office with the aviation industry to improve the understanding of the spread and effect of volcanic ash,[248] and with the Highways Agency on the impact of weather on the road network.[249] A number of other examples were given by the Government in its submission.[250]

67. Following the success of the FFC in providing joined up scientific advice to Government and emergency responders and in recognition of the need for a similar approach to other potential natural hazards, the Met Office has set up the Natural Hazards Partnership (NHP).[251] The NHP brings together thirteen collaborative government agencies to coordinate advice through a single contact point.[252] Within a year of being established, the NHP has "piloted a multi-hazards warning service and the expertise is being integrated into the Cabinet Office National Risk Assessment process to ensure the best use of scientific evidence in planning and preparing for natural hazard events".[253]

INTERNATIONAL PARTNERS

68. In order to achieve the scientific advances required to improve weather and climate services, it is important to collaborate both nationally and internationally.[254] International collaboration is particularly important "to provide the observations on which the Met Office depends" (see paragraph 28).[255] The Met Office represents the UK in the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT), a collaboration that funds and operates a coordinated satellite network, which provides weather and climate data 24 hours a day, 365 days a year.[256] There are also "intergovernmental arrangements" between Europe and other countries, "for the real time exchange of weather and climate related satellite data".[257] While there are clear advantages in sharing the cost of obtaining satellite data, there are also some problems with relying on international partners for crucial meteorological observations. For example, on 24 October 2011, the Guardian reported that budget cuts in the United States Congress could affect a "critical weather satellite" that is relied upon by the Met Office and other national meteorological services around the world.[258] When we asked the Minister about the extent to which the UK Government could influence its international counterparts in such a situation, he responded:

I am not sure whether it is a question of Ministers phoning their counterparts and having strong discussions; it is more a question of ensuring that strong collaboration and co-operation across the globe is maintained. We certainly recognise that various Governments are under many cost pressures. The best way to apply pressure is to work together through those cost pressures and to understand their longer-term implications.[259]

We also asked how the specific problem highlighted in the Guardian article was being dealt with. The Minister was initially uncertain about the discussions occurring between the UK and the United States, but was eventually able to clarify that Mr Hirst was in discussions with his American counterpart and that exchanges were not taking place at a Ministerial level.[260] The Government also later added in a supplementary memorandum to us that "this is a US political issue and therefore Ministerial involvement is unlikely to be of assistance at this time, however support has been offered to Mr Hirst on this matter should the situation change".[261]

69. Beyond EUMETSAT, there are also close links between the Met Office and Europe through the European Centre for Medium-Range Weather Forecasts (ECMWF), a part of the "European Meteorological Infrastructure", which coordinates meteorological activities in Europe.[262] The Met Office is one of 19 Member States of the ECMWF.[263] The ECMWF provides a "complementary" service to the Met Office as it looks at longer-range forecasts.[264] Professor Sir Brian Hoskins, Chair of MOSAC, told us that MOSAC had called for the Met Office to improve its collaboration strategy with the ECMWF and that there were now moves in that direction.[265] While there is no mention of the ECMWF in the Met Office's science strategy, we note that it is referred to in the science strategy implementation plan.[266]

70. The Met Office also represents UK interests at the World Meteorological Organization (WMO).[267] The Met Office plays an "active role" in the WMO and "participates in a number of the WMO Commissions and Working Groups".[268] For example, Met Office scientists are actively engaged with the WMO's World Climate Research Programme, World Weather Research Programme and The Observing System Research and Predictability Experiment (THORPEX).[269] The WMO is also used as a forum in which to discuss collaboration on severe weather events, such as space weather.[270]

71. As previously discussed, the Met Office also collaborates with international partners on modelling: specifically, by developing and testing its Unified Model with those countries using it under licence (see paragraph 32). Amongst those working with the Met Office in this way is the Australian Bureau of Meteorology. The Bureau told us that "based on several years of experience, the Met Office is very well planned, efficient and effective in the way it structures and manages its side of the relationship".[271]

72. The Met Office collaborates well with a number of partners, both in the UK and internationally, to improve the accuracy of its modelling and predictions, and also to enhance the effectiveness of its advice to responders dealing with severe weather events. The Met Office is well respected internationally and it was able to provide us with a number of examples of how its work is feeding through to improved services.

57 Met Office, Met Office science strategy 2010-2015, November 2010 Back

58 As above Back

59 For example: Ev 61, para 3.2.1 [National Oceanography Centre]; Ev 64, para 7 [Royal Meteorological Society]; Ev w7, para 3 [Australian Bureau of Meteorology]; and Ev w10 [Rowan Douglas] Back

60 Ev w15, para 14 Back

61 Q 101 Back

62 Q 99 Back

63 Q 101 Back

64 Qq 99-100 Back

65 Q 101 Back

66 Ev w15, para 13 Back

67 Met Office, Met Office science strategy 2010-2015: Implementation Plan, November 2011 Back

68 Q 101 Back