Wind engineering—Past, present and future

ARTICLE IN PRESS

Journal of Wind Engineering and Industrial Aerodynamics 95 (2007) 843–870 www.elsevier.com/locate/jweia

Wind engineering—Past, present and future C.J. Baker School of Engineering, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK Available online 23 March 2007

Abstract This paper firstly considers the history of wind engineering in five rather arbitrary time periods— the ‘‘traditional’’ period (up to 1750), the ‘‘empirical’’ period (1750–1900), the ‘‘establishment’’ period (1900–1960), the period of growth (1960–1980), and the modern period (1980 onwards). In particular it considers the development of the discipline in terms of the socio-economic and intellectual contexts of the time. This leads to a description of the current state of the discipline and a forward look at possible developments, again taking into consideration the likely socio-economic and intellectual changes in the next few decades. r 2007 Elsevier Ltd. All rights reserved. Keywords: Wind engineering; History; Wind loads; Wind tunnels; CFD; Future developments

1. Introduction This paper, written at the request of the organisers of the 4th European and African Conference on Wind Engineering, considers first of all the history of wind engineering and attempts to summarise the current state of the discipline, before then attempting to consider the prospects for the subject over the next few decades. Now it is relatively easy to write a history of the subject that simply lists the ‘‘facts’’ and applications of the subject to various situations, and this is the approach that has been taken by several authors recently—see for example Aynsley et al. (1977). Cook (1985), Davenport (1999), Surrey (1999), Meroney (1999), Murakami and Mochida (1999). Indeed parts of this paper draw heavily on the work of these authors. However, all the developments within wind engineering have taken place within specific socio-political and intellectual contexts and Tel.: +44 121 414 5067; fax: +44 121 414 36785.

E-mail address: [email protected]. 0167-6105/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.jweia.2007.01.011

ARTICLE IN PRESS 844

C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

cannot properly be considered without a consideration of such contexts. In particular future developments will inevitably be driven by the prevailing social and intellectual conditions of specific times and places, and thus if the future prospects of wind engineering are to be properly considered, some attempt must be made to predict the contexts in which the developments will take place. Sections 2–6 of this paper thus considers the history of wind engineering, with due regard to the prevailing social conditions and intellectual environment. To structure this discussion we consider five periods—the ‘‘traditional’’ period (up to 1750), the ‘‘empirical’’ period (1750–1900), the ‘‘establishment’’ period (1900–1960), the period of growth (1960–1980), and the modern period (1980 onwards). These periods are to an extent arbitrary, and should not be regarded as fundamental in any way—but simply serve as a useful framework for discussion. But first some justification of what follows is required. This paper approaches the consideration of wind engineering history from a fundamentally European point of view— and indeed many readers may well judge that the viewpoint is even more restrictive and that the approach is excessively anglo-centric. If an apology is needed for such a perspective, then this emphasis arises of course because of the author’s experience and background. Further by way of justification, at least in the European context, it may be argued that the UK experiences the force of the North Atlantic weather systems and generally higher wind speeds than most other European countries. This has driven many of the developments in wind engineering that have found application elsewhere. This point being made however, developments in wind engineering techniques have of course taken place throughout Europe and around the world, even if all of these are not fully represented in this paper.

2. The traditional period (up to 1750) The historic period up to 1750 of course covers a vast range of different social and intellectual contexts. In wind engineering terms, in most parts of the world, the style of structures to withstand prevailing wind conditions evolved by experience and the development of tradition (Aynsley et al., 1977). These styles were thus inevitably localised and varied significantly from place to place. I would suggest however that these structures were also heavily influenced by considerations that are foreign to the modern secular mind—those of religion and ritual. As an early example of this we can consider the Iron Age of the Atlantic seaboard of Europe, where a particular style of structure developed known as the round house, with circular plan and a conical roof (Pryor, 2003). Now the people of this time and region had a highly developed ‘‘ritual’’ system that made extensive use of solar and lunar observations, and based much of their daily and seasonal routine around such systems. In this, circular geometries were of major importance—for example in burial mounds and the earlier Neolithic and bronze age henge monuments, such as Stonehenge or Avebury. The round houses themselves (Fig. 1), including their internal arrangement, were part of such ritual systems, and are very often oriented with major solar alignments. Nonetheless it is possible to conjecture that this style arose at least partly because in this area, the windiest part of Europe, the conical roof form experienced significantly lower loadings than the rectangular roof forms, that are more common in Central and Eastern Europe.

ARTICLE IN PRESS C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

845

Fig. 1. A reconstructed Round House at Flag Fen in Cambridgeshire, UK (Flag Fen, 2004).

The ritual/religious context to structural form and to the effects of natural phenomena, although alien to modern minds, nonetheless persisted well into historic time. As an example consider the following two entries from the ‘‘Anglo-Saxon Chronicle’’ that records events in England from the dawn of the common era to around 1150 (Whitlock, 1961). 793

1122

In this year fierce, foreboding omens came over the land of Northumbria and wretchedly terrified the people. There were excessive whirlwinds, lightning storms and fiery dragons were seen flying in the sky. The signs were followed by great famine and shortly afterwards in the same year the ravaging of heathen men destroyed God’s church at Lindisfarne through brutal robbery and slaughter. Thereafter, the Tuesday after Palm Sunday, was a very great wind on March 22nd. Thereafter came many signs far and wide in England, and many illusions were seen and heard. The night of July 25th there was a very great earthquake over all Somerset and Gloucester. Later, on September 8th, St Mary’s day, there was a very great wind from morning to black night. That same year passed away Ralph, Archbishop of Canterbury.

The use of religious language for specifying dates is obvious, but note that wind storms are seen here as divine omens—in the first case of the coming of the Vikings to the north east of England, and in the second of the death of the Archbishop of Canterbury amid the tumultuous events in the reign of Henry I of Normandy and England that led to strained relationships with the Pope. In a rather more secular vein consider the entry for 1114.

ARTICLE IN PRESS 846

1114

C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

This year were very great winds in the month of October, but it was immeasurably great on the night of the Octave of St Martin, November 18th, and that was seen everywhere in the woods and villages afterwards.

The observation that trees and homes suffer most from windstorms is of course as valid now as it was then. Even within this traditional and religious framework of society however developments took place. In the 12th and 13th centuries across Europe, the desire to glorify God in architecture (mixed with not a little rivalry between religious communities) led to ever larger and grander churches being built, which would inevitably experience greater wind loads than on earlier, lower structures. Roofing methods evolved to ensure that roofs were not lost in wind storms—and in particular the use of lead on roofs became widespread. However perhaps the major evolution can be seen in spires (Encyclopedia Britannica, 2004). These were originally simply tall pyramids with four faces. However, over the centuries they evolved into polygonal and eventually conical structures (such as in Fig. 2). Such structural form would experience significantly lower wind loads that the sharp edged pyramid structures (because of the narrower wakes at supercritical Reynolds numbers) and it must have been observed fairly quickly by those responsible for the building of such

Fig. 2. The spires at Lichfield Cathedral, Staffordshire, UK.

ARTICLE IN PRESS C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

847

structures that polygonal spires blew down less frequently than pyramid shaped spires. Over the following centuries forms of spire also developed with significant gaps in the structure—which again would have reduced drag (such as the central spire on Fig. 2). Another example of the development of traditional forms can be found in the development of windmills. These seem to have been first introduced in Persia around 950, and were very much based on the design of water mills, although these were vertical axis machines with horizontal blades (Hill, 1984) (Fig. 3). The windmill seems to have been introduced into Europe around 1180, and records of mills exist in Normandy and in England at around that date. The number of mills increased rapidly around this time, and in the 13th century there were 120 mills in the vicinity of Ypres alone. Mills are also recorded in Italy at around this time, and the first illustrations begin to appear. These illustrations however show a considerable evolution from the original design, with all the mills being horizontal axis post mills of the type that was to become very familiar across Europe for the next few centuries (Fig. 3). Towards the end of this period, the intellectual atmosphere across Europe changed significantly with the onset of the Renaissance and the period known as the Enlightenment. In the 17th and 18th centuries we see the birth of modern science, with contributions from Newton, Euler and Bernoulli that were of very wide significance and were eventually to play a major role in the discipline of wind engineering. The influence of organised religion declined significantly and societies across Europe became increasingly secular. Building techniques were evolving into more modern forms, as the religious drivers of structural form were becoming of much less significance. As an example of how such forms coped with strong winds it is instructive to consider the effects of the ‘‘great storm’’ that affected Western Europe in 1703. RMS (2003) quotes the work of Defoe (1704), the author of ‘‘Robinson Crusoe’’, who wrote a comprehensive account of the storm’s effects in the UK. In the rest of Western Europe the War of Spanish Succession was underway, and newspaper reports there were politically coloured, with claims that the opponent’s ports

Fig. 3. Vertical (early Persian) and horizontal (Medieval Dutch) axis windmills (Hill, 1984).

ARTICLE IN PRESS 848

C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

were destroyed. The storm consisted of a severe depression of around 95 mbar that travelled across the English Midlands on December 7 and 8, 1703 and caused major damage over a wide area. It was at its peak in London at 6:00 a.m. and in Copenhagen at 11:00 a.m. Estimates of maximum wind speeds are in the range of 45–50 m/s. There is some evidence of ‘‘streaky’’ damage (which will be discussed further below). The return period can be estimated to be around 200 years. Defoe writes that in London: ‘‘the streets lay so covered with Tiles and Slates, from the Tops of the Houses, especially in the Out-parts, that the quantity is incredible, and the Houses were so universally stript, that all the Tiles in Fifty Miles round would be able to repair but a small Part of it’’ London at the time was a relatively new city, having been rebuilt after being destroyed by the Great Fire of 1666, and thus much of the construction was ‘‘modern’’. Nonetheless around 2000 chimney stacks were broken down in the London area and there was major gable end damage. Around 16–20 roofs of large buildings were lost. In other words there was very serious structural damage. In London there were 21 fatalities and around 200 serious injuries, with around 120 fatalities nationally. In the UK outside London the damage was equally as severe with major roof damage, over 1000 outhouses or barns destroyed and around 400 windmills toppled over. The total loss in London was estimated to be around £2 million pounds. Given that the estimate for the total value of the buildings in all of England and Wales was around £100 million at the time, this represents a very major impact indeed. In the rest of Western Europe there was similar damage, particularly in Belgium and Holland, with rather lower levels of damage in Denmark. Thus at the close of the period, it is evident that the vulnerability to wind storm damage was increasing significantly. 3. The empirical period 1750–1900 Around the start of this period the social phenomenon that became known as the Industrial revolution began in earnest. Some writers would see this as finding its first manifestation in Coalbrookdale in Shropshire in the UK, where the ironmasters of the area became ever more daring in their art, culminating in the building of the first iron bridge across the River Severn. This structure was constructed in a traditional style that would have been used for timber bridges. As such it was in modern terms heavily over engineered (see Fig. 4) and would most certainly not have experienced aeroelastic oscillations! Close by, in Birmingham James Watt began with Matthew Boulton to construct steam engines of ever increasing power. (It is of interest to note that the ‘‘heirloom’’ of office of the Head of Engineering at the University of Birmingham is a chair once owned by James Watt himself). These and industrial developments elsewhere led to major economic development across Europe, both at a local and regional scale, through the provision of much improved road systems and bridges, the development of the railway, etc. On a wider scale this was the period of the growth of European empires, which was again driven by the new possibilities that were arising from the industrial revolution, all within a context of largely unrestrained capitalism. Intellectually this period saw the development of classical hydrodynamics building on the work of Euler, Newton and Bernoulli and later through Navier’s formulation of the fundamental equations of fluid flow in 1845. The use of the techniques of potential flow

ARTICLE IN PRESS C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

849

Fig. 4. The Ironbridge in Coalbrookdale (Shropshire Tourism, 2004).

were extensively studied mathematically, and very significant work carried out on the flow around a variety of body shapes—although with the inconvenience that the solutions that were obtained did not always seem to match reality. Scientific experimentation was also gaining in respectability and Hadley and Smeaton carried out the first fluid mechanics model experiments in 1759. Towards the end of this period, work of immense significance was carried out by Osborne Reynolds in Manchester which revealed the differences between laminar and turbulent flow in pipes, and, in retrospect, began the undermining of the edifice of classical hydrodynamics. It was in this context that the effects of wind loads on transportation systems came to be of significance. In Naval terms Admiral Beaufort derived the familiar Beaufort Scale in 1808, to enable reasonably accurate weather information to be given to the military and merchant fleets. The first long span bridges that were built to improve communication links inevitably suffered from adverse effects of the wind leading to some spectacular downfalls—such as the 1836 collapse of the Brighton Chain pier due to aeroelastic oscillations, and, most famously, the collapse of the Tay bridge in 1879 (Fig. 5). This resulted in perhaps the most appalling poetry in the English language from William McGonnegal. (The full text of which is given in Appendix as it seems to the author that it is essential reading for all wind engineers.) So the train mov’d slowly along the Bridge of Tay, Until it was about midway, Then the central girders with a crash gave way, And down went the train and passengers into the Tay! The Storm Fiend did loudly bray, Because ninety lives had been taken away, On the last Sabbath day of 1879, Which will be remember’d for a very long time. More positively the Tay Bridge collapse led to the experiments of Benjamin Baker in 1884 (Baker, 1884), carried out as part of the building of the Forth Bridge, who measured the wind forces on rectangular plates, and found that smaller plates experience

ARTICLE IN PRESS 850

C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

Fig. 5. The Tay Bridge collapse.

proportionally more load than larger plates, and conjectured, quite correctly, that this was due to the greater coherence of small rather than large gusts. 4. The establishment period 1900–1960 At the beginning of the 20th century, the industrial revolution was coming of age and beginning to influence every aspect of society. Nowhere was this more true than in the military sphere and it is an unfortunate fact that many of the technological advances that were made in the 20th century were largely driven by military considerations, as the European empires vied for supremacy both within Europe and around the world. For example in the study of atmospheric dispersion early experimental and theoretical developments at Porton Down in the UK were driven by the demands to understand aspects of chemical warfare (Meroney, 1999). Large-scale technological warfare demands a collective effort and this period saw the development of the large government laboratories that were ultimately to have such an influence on wind engineering. By contrast, the increase in personal mobility led to a breakdown of established social patterns, one side effect of which was an increasing demand for high-level education by all classes of society. This led throughout this period to the establishment of more and more Universities, many with a technological bias, that were again to become active in the wind engineering field. Technological and material advances also led to attempts to construct more and more challenging structures—such as the first high-rise structures, and ever longer suspension bridges. In this context the work of Baker was taken forward at the National Physical Laboratory in Teddington under the direction of Thomas Stanton, who carried out fullscale measurements with tower arrays to attempt to find the ‘‘size’’ of wind gusts (Davenport, 1999). This work also included wind measurements on Tower Bridge in London. In terms of intellectual concepts, this period was dominated by the rise of boundary layer aerodynamics as pioneered by Prandtl and von Karman, which clearly demonstrated the deficiencies of classical hydrodynamics, together with the birth and development of the statistical theory of turbulence, first set out by Taylor in 1935. In addition statistical

ARTICLE IN PRESS C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

851

techniques were also developing rapidly and the extreme value theory that was to become a dominating paradigm in wind engineering was first elucidated in 1928 by Fisher and Tippett (1928). This era also saw the birth of what the author regards as three of the main wind engineering tools. Firstly there was the development of the wind tunnel—driven by the nascent aeronautical industry, although the first wind tunnels predated the advent of aeroplanes, with the pioneering work of Wenham in 1871 (Surrey, 1999). In 1893 Irminger measured pressure distributions on a variety of shapes using the flow through a chimney. Irminger (1895). Eiffel made his first wind tunnel measurements in 1909. In the 1930s Irmiger made measurements on building models in low turbulence wind tunnels. (Irminger and Nokkentved, 1936). Secondly there was the development of codes of practice with the realisation of the need to provide engineers with practical guidance on design to enable environmental loads such as wind to be properly defined. The first UK code of practice was published in 1944 (British Standards Institution, 1944). Thirdly this period saw the beginnings of full-scale measurements of wind loads on structures. It is in the interaction between wind tunnel and full-scale tests that most progress was made in the field of wind engineering during this period. The experiments of Irminger mentioned above were carried out in low turbulence wind tunnels, and many other similar measurements were made in the years that followed and produced loading data that was used in codes of practice for many decades after this. A similar set of low turbulence measurements were those made by Dryden and Hill on the Empire State Building in 1933. In 1940 Rathbun made equivalent full-scale measurements of the same building, using banks of anemometers that were photographed to give instantaneous pressures (Rathbun, 1940). Not surprisingly these full-scale pressures were subject to rapid fluctuations due to turbulence. Nonetheless it became obvious that these results did not agree at all well with the wind tunnel results. This was resolved in 1944 with wind tunnel measurements with an attempt to simulate the atmospheric boundary layer within the wind tunnel, and indicated the absolute need to simulate the wind characteristics in the wind tunnel if proper modelling was to be achieved and reliable results obtained. Similar conclusions were drawn from the full-scale measurements on the pressures on a railway shed made by Bailey (1933) which were compared with wind tunnel model results immersed in a deep boundary layer by Bailey and Vincent (1943). A further wind engineering ‘‘milestone’’ of this era was the collapse of the Tacoma Narrows bridge, about which a very great deal has been written. The cause is still not fully understood (Wyatt and Walshe, 1992), but the collapse illustrated the importance of a proper consideration of the aerodynamic stability of long span bridges, and investigations after the failure led to the derivation of appropriate wind tunnel scaling laws and wind tunnel methodology by Farquharson in 1949. Work on long span bridges also took place at NPL in the UK, with a full bridge study of the Severn Bridge (Frazer and Scruton, 1952), and the development of section model testing, in parallel with similar work in the USA and Japan. Towards the end of this ‘‘establishment’’ period the final foundations were being laid for the rapid development of wind engineering over the coming decades. Van der Hoven carried out the full-scale measurements that enabled him to derive the van der Hoven spectrum which, because of the existence of the spectral gap, allowed the concepts of independent small and large scale wind fluctuations to be formulated, which is of

ARTICLE IN PRESS 852

C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

Fig. 6. Scruton, Jensen (from Davenport (1999)) and Cermak.

fundamental importance to the developments that followed over the coming decades (van der Hoven, 1957). At the National Physical Laboratory in London, Scruton further developed techniques for the wind tunnel modelling of long span bridges, together with scaling laws for the wind tunnel modelling of dynamic structures—which ultimately resulted in the widespread acceptance of a dimensionless number that bears his name (Scruton, 1963). In Denmark Jensen was formulating the scaling laws for proper boundary layer testing of low rise structures, and again deriving the scaling factor that was to bear his name (Jensen, 1958). In the USA Cermak set up the Fluid Dynamics laboratory at Colorado State University, where much fundamental work was carried out on the simulation of atmospheric flows with and without thermal effects (Fig. 6). By 1960 the foundations for the discipline had been securely laid on the much older foundations of the disciplines from which it has grown—meteorology, fluid mechanics and structural mechanics. 5. The period of growth 1960–1980 During this short 20-year period there were major changes in the nature of society in Europe and across the world, and these were paralleled with significant advances within the discipline of wind engineering. In western society at large, the after effects of the 1939–1945 war were beginning to fade, and this was a period of ever increasing prosperity and economic development. There was a widespread ‘‘social democratic’’ consensus in western countries, one side effect of which was the further development and growth of central government research facilities. One British Prime Minister of the 1960s spoke of ‘‘the white heat of technology’’ and this was a period of optimism concerning the potential benefits of technology to society (by contrast, and rather chillingly this was also the period when the fear of nuclear annihilation was at its height). High rise structures and other large infrastructure projects were the order of the day across the developed world. During this period computer technology began to develop at an accelerating pace, and there were massive developments in the design of scientific instruments and in data acquisition technology. It was within this context that the discipline of wind engineering developed, with everincreasing speed, and achieved some level of self-definition. The papers presented at the

ARTICLE IN PRESS C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

853

international conferences (see below) indicate that the primary concern has been with strong wind climatology and wind structure, wind effects on low rise and high rise buildings, bridges, masts and towers, wind damage, codification and insurance. The discipline has concerned itself to a much lesser extent with the study of wind effects on people (pedestrian comfort, air quality, ventilation, etc.), wind effects on the natural environment (trees and crops, sand and snow drifting) and to studies of wind energy, these tending to become the preserve of other, related discipline. From now on in this paper, I will not mention any individuals, as in historical terms I believe we are too close to fully appreciate any individual contribution to the subject. That is, I will not mention any individuals other than one—Prof. Alan Davenport (Fig. 7). The importance of the contribution made by Davenport to the subject cannot be over emphasised. In 1961 he elucidated the concept of the wind loading chain, which gave a conceptual framework to the study of wind effects on structures (Davenport, 1961). He was heavily involved in the construction of the World Trade Centre. The laboratory he developed at the University of Western Ontario became the dominant force in international wind engineering terms. Davenport’s major role having thus been noted, there were of course many others involved in the development of the discipline during this period.

Fig. 7. Alan Davenport.

ARTICLE IN PRESS 854

C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

Davenport’s concept of the wind loading chain, which he applied in the frequency domain, led to a range of spectral methods for calculating the loads and displacements of high rise buildings, bridges, etc. These have become very widely used throughout the discipline, and indeed have become the dominant class of analytical method, although by their nature they implicitly assume linear structural behaviour. These decades saw the development of the boundary layer wind tunnel from an essentially research tool, into a reliable and robust tool for commercial design purposes, with the increasing realisation of the need to model the turbulence spectrum as accurately as possible and with the routine use of small pressure transducers with scanivalves, and the introduction of the base balance techniques. Techniques for the measurement and prediction of atmospheric pollutants also advanced rapidly (Meroney, 1999), and in 1961 Pasquil developed his classification of atmospheric stability that was to remain in use for many decades. Around the world a number of ground breaking full-scale experiments took place—the Aylesbury house experiment (Eaton and Mayne, 1975) and the mobile home measurements in the USA (Marshall, 1975). Perhaps most significantly for the long-term future of the discipline, the series of international conference began (Table 1). Davenport (1999) noted that the first of these was somewhat different in character from the rest, with many fewer papers, and much greater industrial involvement than the conferences that followed. Perhaps something has been lost in this change. During this period, the process of codification of wind effects began in earnest, and a significant number of codes were developed by National Standards Organisations— for example the updated UK code (British Standards Institution, 1972) and the Australian Code (Standards Association of Australia, 1973). One particularly useful development that has been of long-term importance is the production of the ESDU Wind Engineering series, which has become an invaluable data source for many designers over recent decades (ESDU, 2004). Finally in 1975 the first edition of the Journal of Industrial Aerodynamics was published (later to become the Journal of Wind Engineering and Industrial Aerodynamics) and quickly established itself as the main journal of record in the field.

Table 1 The development of the International Wind Engineering Conferences (updated from Davenport, 1999) Year

Location

Papers

Attendees

Countries

1963 1967 1971 1975 1979 1983 1987 1991 1995 1999 2003

Teddington Ottawa Tokyo Heathrow Fort Collins Gold Coast Aachen London, Ont. Delhi Copenhagen Lubbock

24 42 113 70 250 120 90 264 200 290 238 (+126 posters)

300 85 215 200 350 — — 385 — — 380

20 20 26 25 25 — — 35 25 — 32

ARTICLE IN PRESS C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

855

6. The modern period 1980– Whilst all divisions of historical periods in this paper are arbitrary, this is particularly the case with any division that marks the start of a ‘‘modern’’ period. Developments within wind engineering were continuous across the boundary. Perhaps the real reason that the author has chosen this date is that 1980 is the time when he first became involved in the discipline! However, the date does seem to mark, even if in an arbitrary fashion, the maturity of the discipline, with the adoption of the Journal of Wind Engineering and Industrial Aerodynamics as the official journal of the International Association of Wind Engineering. More seriously at around that time, major societal changes were beginning. The prevailing social democratic consensus was beginning to break down, and individualism and free choice became the dominant themes of political discourse. This was the era of privatisation in a number of western countries, with the closure or commercialisation of national laboratories and industries. There was a general lessening of public confidence in science and technology and development slowed down, although there was growth in the service industries—particularly in financial services. The discipline of wind engineering was seriously affected by such changes with the closure/commercialisation of many national building institutes for example (often with a deplorable loss of corporate memory of past work). There was a reduction in the number of large developments within the developed world to be tested at model scale, although this lack of work was often compensated for by an increased workload on large high-rise structures in the developing world. These points being made there have been major developments within the subject over this period. There have been significant advances in wind tunnel testing techniques, particularly in terms of instrumentation, with the use of large number of simultaneously monitored pressure transducers and the increasingly frequent use of LDA and PIV techniques for velocity measurements. Similarly the development of the three component sonic anemometer has revolutionised full-scale measurements. Further major full-scale experiments were carried out in South Africa (Milford et al., 1991a, b), Silsoe in the UK (Richardson et al., 1995; Hoxey et al., 1995) and at Texas Tech (Levitan and Mehta, 1991a, b). At the time of writing an extensive project is underway in the USA to measure wind conditions and full-scale structural loading during hurricanes, which should yield a very considerable quantity of information that will be of significant use in design (Gurmley et al., 2005). All these developments have of course been underpinned by the rapid growth in IT techniques and computer power which makes high-speed data acquisition and the analysis of large amounts of experimental data possible. It has also led to the increasing use of what is now the fourth fundamental tool of wind engineering—CFD techniques. Computational fluid dynamics has progressed immensely over the past two decades– through the use of inviscid panel methods; then simple k–e techniques, which were afterwards refined in various ways to make them more suitable for wind engineering application; and now increasingly through unsteady flow methods such as LES, DES and discrete vortex modelling. For a full review of CFD developments in Wind Engineering, see Murakami and Mochida (1999). The last 20 years have also been extremely busy in terms of code development and revision across the world. These are well summarised in the recent series of papers produced by the IAWE Codification Initiative (Wind and Structures, 2005). In conceptual terms the period has seen an increasing application of modern analytical methods to wind engineering—particularly advanced probabilistic

ARTICLE IN PRESS 856

C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

techniques, wavelet analysis, orthogonal decomposition, etc. (see Kareem, 2003). Of particular significance has been the gradual trend towards using time domain methods in the design process. This will be discussed further below. As the discipline has developed, the regular international conferences have become focal points for the academic and research communities within Wind Engineering. At the 1983 conference the International Association of Wind Engineering was formed, whose primary role for the first 15 or 20 years of its existence was the organisation of the International Conferences (Solari, 2005). These have come to be supplemented by a regular series of regional conferences for the three IAWE regions—Americas (from 1970), Asia/Pacific (from 1985) and European/African (from 1993 onwards). Over this period a number of national or regional wind engineering associations have been formed, and this process has been greatly stimulated by the more formal constitution recently adopted for the IAWE in 2003. Other conference series have been initiated that are regularly attended by the Wind Engineering community—the Bluff Body Aerodynamics and its Applications series that began in Kyoto in 1989, and the Computational Wind Engineering Conferences that began in 1992. This period has also seen a seeming increase in the frequency of major wind disasters both in temperate and in tropical regions. There are a number of possible reasons for this including the increase in population densities in vulnerable areas, the growth of informal megacities and, almost certainly, the first effects of anthropogenically induced climate change. Whatever the reason there have been a number of storms that have had major impacts on the regions they have affected. It is appropriate to consider the effects of such storms at this point, since they illustrate something of the current understanding and preparedness for large wind storms. In particular we consider four ‘‘events’’, two near the start of the period—the 1987 storm in the UK and Hurricane Gilbert in the Caribbean in 1988—and two near the end—the late 1999 storms Lothar and Martin in Western Europe and Hurricanes Charley, Francis, Ivan and Jean in the Carribean in 2004. The October 1987 storm was a major cyclonic depression that passed across the south east of England, hitting London in the middle of the night at around 3:00 a.m. (Buller, 1988). The full extent of this storm was not well forecasted, largely due to a lack of offshore meteorological stations, and came as a complete surprise to the public. The maximum wind speeds are shown in Fig. 8. The ‘‘streaky’’ nature of the damage suggested the existence of large-scale structures within the main cyclonic depression. These have been described by Browning (2003) and Browning and Field (2004) and given the name ‘‘Sting Jets’’ because of the characteristic ‘‘scorpion tale’’ pattern of clouds that are associated with them. Gust speeds of around 40–45 m/s were recorded in the south of England, with gusts of up to 50 m/s near the coast and over open sea. The return period was around 150–200 years. In the whole of the UK there were around 18 casualties and very considerable damage and disruption was caused across the south of England. However, most of this damage was due to tree fall and power supply failure. Whilst there was significant cladding damage to buildings, particularly to older buildings that were built before the development of codes of practice, the lack of major structural damage was very noticeable (Fig. 9). In other words all the modern structures, which had been designed to be wind resistant, survived, and the majority of structures which had been designed to various generations of wind loading code of practice, performed more than adequately. In late December 1999 the storms Lothar and Martin moved across Northern France and into Germany in quick succession on December 26 and December 27. (Munich Re,

ARTICLE IN PRESS C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

857

Fig. 8. Gust speeds in the 1987 storm (from Meteorological Office, 2005).

2002). The storm tracks are shown in Fig. 10. Maximum wind speeds of 40–50 m/s were recorded, which one presumes would be near or above the design wind speeds for engineered structures. The insured losses in the two storms were 11.5 and 4 billion euros respectively, with 110 fatalities in Lothar and 10 in Martin. In both storms there was very significant fac- ade and roof damage to buildings (including chimney fall), with cranes and overhead power lines being blown down. Blackmore and Delpeche (2002) showed that much of the roof damage was to relatively old tiled roofs and chimneys, whilst a major part of the fac- ade damage was caused by flying debris. There was very significant disruption to electricity supplies with millions of consumers being without power in France for several days, and major disruption to transport services of all types. Much of this disruption was due to fallen trees and there was significant forest damage. All in all the effects were similar to the 1987 storm in the UK—much cladding and roof damage, and major disruption to urban life. However, the warnings for these storms were significantly better than for the 1987 storm, because of the enhanced meteorological networks. Hurricane Gilbert first made landfall on September 12, 1988, when it struck the island of Jamaica, with torrential rains and winds in excess of 110 mph (Fig. 11). The hurricane caused buildings to loose their roofs and destroyed power supply lines, cutting off the island’s 2.3 million residents from the outside world. The centre of the storm hit the capital of Kingston, where wind gusts up to 140 mph were recorded, seriously damaging the Kingston airport and tossing airplanes across the tarmac. In total 26 people were killed and 500,000 were left homeless. Gilbert then moved on to the Cayman Islands, and 140 mph winds were recorded on September 13. By the time it had reached the Yucatan peninsula wind speeds of 175 mph were recorded, and a historic low-pressure reading of

ARTICLE IN PRESS 858

C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

Fig. 9. Damage in the 1987 storm (Buller, 1988).

888 mbars was measured. On September 14, the eye of the storm passed over the Mexican resort areas of Cozumel and Cancun with 160 mph winds. As Gilbert passed over the Yucatan Peninsula, 23-ft waves and continuing high wind speeds were recorded, which devastated the beachfront resorts and forcing tens of thousands to evacuate the area. In total 29 people were killed and the hurricane caused more than $880 million in damage and left nearly 200,000 homeless. The storm then weakened as it moved north over the Gulf of Mexico and the USA. Whilst the meteorological forecasts were on the whole accurate, in many instances these were not communicated to those likely to be affects—and there is colloquial evidence that tour companies were still moving tourists into the path of the hurricane up to a few hours before the hurricane hit. The 2004 Caribbean Hurricanes were the most severe series of storms in the region for a number of decades. They tracked from East to West in succession over a period of a few weeks—Charley (12–14 August), Francis (1–4 September), Ivan (14–16 September) and Jean (15–17 September). Storm tracks are shown in Fig. 12. The effect these storms had on different countries very much depended upon the state of development of these countries, their preparedness for such events, and to a significant extent the political systems in these countries. In almost all countries there was some attempt to warn the population before

ARTICLE IN PRESS C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

Fig. 10. Storm Tracks of Lothar (top) and Martin (bottom) (from Munich Re, 2002).

859

ARTICLE IN PRESS 860

C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

Fig. 11. Hurricane Gilbert storm track.

the storms, and to evacuate vulnerable areas. The efficiency of this process however varied significantly across the region—with large-scale (repeated) evacuations in Florida and in Cuba (where the process was more or less enforced on the population). In other countries there was simply nowhere to be evacuated to, and the population managed as it could in local shelters. In general however warnings were given far enough in advance for preparations to be made, and communications were restored fairly quickly after the events This would not have happened a decade or so ago, and there has clearly been significant technical progress in the field of forecasts and warnings that has been of major significance. In all areas however there was major damage to structures, disruption of power and transport systems for many days, loss of crops, contamination of water supplies, etc. In the light of such major effects it has been heartening to see the involvement of wind engineers in the IDNDR. In writing of this activity Davenport (1995) began to see wind engineering in its wider context and advocated building design for hurricane resistance, hazard identification, education, legislation, insurance. This is surely a trend to be welcomed by the wind engineering community. 7. The future It has been shown in the preceding sections that to a large extent, developments in the discipline that has come to be known as Wind Engineering have been driven by various social and economic circumstances, and by the prevailing intellectual environment. This is likely to continue to be the case over the next few decades, and thus to consider future prospects for wind engineering, the likely contexts need to be considered. These contexts

ARTICLE IN PRESS C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

861

Fig. 12. Storm tracks of 2004 storms.

are likely to be different in different parts of the world. In parts of the less developed world, there will be a need for fundamental infrastructure development, in terms of transport links, energy infrastructure, etc. In other parts of the world (for example China and various countries in South East Asia) urban development will continue at a pace dependent upon the growth of local economies. In Europe and North America the socioeconomic drivers for the development of wind engineering are likely to change significantly, as the infrastructure is well developed—for example there are few long span bridges left to be built. Whilst there will continue to be a need for wind engineering involvement in the design of prestige large structures, the focus is likely to turn towards the need to sustain and maintain the existing infrastructure, to develop energy sources that are sustainable in the long term. The desire to improve the quality of life will also become an increasingly important driver. For all parts of the world the effects of climate change on the frequency and intensity of wind storm are likely to be important. In intellectual terms, it is likely that the continued and seemingly limitless development in computing power will become of increasing importance—and thus computational techniques will become of increasing practical use. Similarly the increase in computing power will mean that time series analysis of wind loading becomes increasingly attractive, and may well displace some of the well established statistical and frequency domain methods that have been used in the past, and will thus be able to deal more satisfactorily with non-linear structural behaviour.

ARTICLE IN PRESS 862

C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

Thus on the basis of the above assumptions for the likely socio-economic and intellectual contexts, I would see the following trends emerging in the discipline of wind engineering. (a) In the less developed world, there will be an increasing need for wind engineering expertise in the development of appropriate, sustainable, wind energy systems. This will involve both the development of novel turbines using locally based material, and also a significant amount of fundamental work to fully specify local wind climates. In such areas, wind engineering expertise could also be useful in the alleviation of a number of other problems—such as wind damage to agriculture, shelter belt design, etc. (b) In the developed world there will be a continued need for wind engineering input into the design of new infrastructure and buildings. In Europe and similar areas, the emphasis of wind engineering will change from involvement in the design of new structures, to involvement in the maintenance and improving the sustainability of existing infrastructure. Indeed the major impact made by recent storms in the UK (of one to two years return period) has been to cause major disruption to transport and power supply links through tree fall, flying debris, etc. The incidence of structural damage has been small (Baker, 2004). Where it has occurred, it is usually limited to cladding damage that (as noted by ABI, 2003) could usually be alleviated by better maintenance of the building stock. (c) The requirement for sustainable energy sources, and the need to reduce greenhouse gas emissions will also lead to an increasing need for the development of wind energy systems within developed countries. In addition the improvement of air quality will become of increasing importance. If wind engineers are to become more involved in both these developments, then more consideration needs to be given to the structure of low speed winds than has been the case in the past—particularly the effects of atmospheric stability. (d) A full understanding of the effects of climate change on wind storm frequency and intensity is required in order to establish appropriate design risks. Current predictions from the use of large-scale Global Climate Models are ambivalent (Munich Re, 2002; ABI, 2003) but the most recent suggest that the major effect of climate change in European terms is to weaken the high pressure region that lies over the centre of Europe in the winter period, which will allow a southward drift of storm tracks. This will result in higher extreme wind speeds over the south of England, France, Germany, etc (McDonald, 2004). In the author’s opinion one of the pre-requisites for such an understanding is to determine the reasons for the differences in the design wind speeds for different countries in Europe (as given in the National Annexes for the Eurocodem—see Fig. 13 from Miller, 2003) to establish a reliable base for understanding the likely effects of climate change. Methods based on gradient wind condition and surface pressure measurement such as that developed by Miller (2003) would seem to be appropriate here in harmonising design wind speeds across Europe (Fig. 14). (e) There is a need to revisit some of the basic assumptions of wind engineering in the light of recent advances in measurement and analysis techniques. For example there are recent indications (Jensen, 1999; Courtney and Troon, 1990; Richards et al., 1999) that the existence of a ‘‘spectral gap’’ in the van den Hoven spectrum (which is the basis of

ARTICLE IN PRESS C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

863

Fig. 13. Once in 50 year wind speeds in Ireland, UK, and France from code data (Miller, 2003).

Fig. 14. Once in 50 year wind speeds in Ireland, UK, and France from surface pressure data (Miller, 2003).

ARTICLE IN PRESS C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

864

Normalised spectral density

much wind engineering analysis) is not as clear as might be supposed (Fig. 15), and there may well be significant interactions between micro- and macro-meteorological wind conditions, which would have implications for the specification of design wind speeds. In a similar way recent work by colleagues of the author based on the conditional sampling of extreme gust velocities from full scale high response sonic anemometer measurements have indicated that the structure of atmospheric turbulence at full scale and that simulated in the wind tunnel are somewhat different (Fig. 16). Also it would be wise to give further consideration to the commonplace assumption that Reynolds number effects can be neglected provided a minimum value is achieved in wind tunnel tests—the large Reynolds number full-scale measurements of Hoxey et al. (1997) for low rise buildings, the measurements of Larose and D’Auteuil (2004) for large cylinders, and measurements made by the current author for wind loads on trains, and (Fig. 17) give pause for thought in this regard.

0.3 0.25 0.2 0.15 0.1 0.05 0 1.00E-08

1.00E-04

1.00E-06

1.00E-02

1.00E-00

Frequency (Hz) Denmark

UK

Fig. 15. The ‘‘van der Hoven’’ spectrum as measured by Courtney and Troon (1990) in Denmark and Richards et al. (1999) in the UK.

Conditional sampling around peak values (Velocity - mean) / standard deviation 3 2.5 2 1.5 1 0.5 0 -10

-8

-6

-4

-2

0

2

4

6

8

10

Time from peak (sec) Full scale

Wind tunnel geometric scale

Fig. 16. Comparison of turbulence structure at full scale and in wind tunnels.

ARTICLE IN PRESS C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870 1 Lift force coefficient

1 Side force coefficient

865

0.8 0.6 0.4 0.2 0

0.8 0.6 0.4 0.2 0

0

5

10

15

0

5

10

15

10m wind speed m/s

10m wind speed m/s

30

90

60 120

150

Fig. 17. ‘‘Reynolds number effects’’ on a full-scale train coach (figure shows variation of side and lift force coefficients for a variety of wind angles, with 901 normal to the vehicle). At 10 m/s the Reynolds number based on train height is approximately 2  107.

(f) There can be expected to be an ever-increasing use of CFD techniques, to predict both wind environment and, in the medium to long term, loads, and a corresponding fall in the use of boundary layer wind tunnels. The CFD techniques that will prove to be of most use will be those that will faithfully model the turbulence structure within the atmospheric boundary layer, e.g. LES or DES techniques. The use of RANS based techniques will decrease over time, although their relative simplicity and economy will ensure their continued use for many applications. Applications will become widespread in areas where wind velocities rather than surface pressures are required, such as the assessment of pedestrian comfort. These trends may well lead to the concentration of boundary layer wind tunnel testing for complex structures into a smaller number of institutions over the next few decades. (g) There will be an increasing realisation of the importance of non-stationary wind systems, such as thunderstorms, frontal systems, etc., particularly where wind loading is important for serviceability considerations. New types of physical and numerical modelling techniques will be developed to simulate a variety of such unsteady, nonstationary wind effects—for example thunderstorm and tornado generators. (h) The use of time series analysis will replace the use of statistical techniques to some extent—for example in the input of fluctuating wind loads directly into unsteady finite element codes to predict specific load effects. An example of these developments was presented at a UK Wind Engineering Society meeting in February 2005 (Cunningham, 2005), which described the calculation of frame force and moment time histories using pressure time histories from a large number of pressure tappings on a wind tunnel model, and the time domain application of a finite element model. Such techniques will allow a full consideration of the loads due to the non-stationary wind systems described above, as well as allowing the effects of structural non-linearity to be considered. A consequence of this will be a reduced need to carry out loading calculations using frequency domain methods.

ARTICLE IN PRESS 866

C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

(i) Davenport (1999) wrote As Martin Jensen has reminded us the lack of full scale verification that has been tolerated is ‘‘embarrassing’’. It is not characteristic of other technologies such as shipping, transportation or aeronautics. The quote remains relevant and there is a continuing need for full-scale experimentation to validate loading data in codes, the use of wind tunnels and CFD. (j) It is likely that the above developments will lead to another round of code revision in two or three decade’s time, and radically different types of code can be expected that will allow time domain calculations of load using standard wind and loading datasets. (k) To enable the wind engineering community to contribute to improving the maintenance and sustainability of existing infrastructure, more involvement with other ‘‘stakeholders’’ is required—for example with emergency planners (to develop appropriate methods for storm prediction, and for reducing building vulnerability), city planners and engineers (to develop rational methods for maintenance of urban infrastructure to withstand wind storm), environmental scientists (to study the effect of the dispersion of pollutants in urban areas, which has a major effect upon quality of life), and transportation professionals (to reduce the vulnerability of road and rail links to wind storms), etc. If the above developments are to come about, the author is of the opinion that the education sector needs to play a major role in a number of areas. (a) In the author’s experience, which is of course very largely UK based, there is often a major vocabulary gap between wind engineers and others with whom they need to work, and very often much research is not applied as fully as it could be for that reason. There is a need for wind engineers to be able to communicate the fundamentals of their subject in a wide variety of ways—through the encouragement of undergraduate and postgraduate modules in this area, suitably constructed for, say, architectural students, urban planners, maintenance engineers, etc., through the provision of short continuing educational courses for a wide range of professional and managerial groups. (b) In addition there is probably a place for a small number of specialist Wind Engineering programmes at Masters cycle level in Europe to train specialists in the field—perhaps through the ERASMUS MUNDUS scheme. (c) Associated with such initiatives there is a need that is growing ever more urgent for the thorough archiving and cataloguing of much of the report literature of the last few decades, so that the information contained in such reports is available to all involved in the discipline. (d) Research funds, at both national and international level, are becoming more and more directed to multi-disciplinary rather than discipline specific projects. If wind engineering is to survive within universities, the development of multi-disciplinary research teams is becoming more and more vital in the effort to obtain research funding.

ARTICLE IN PRESS C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

867

As can be seen from the above, the discipline of wind engineering remains active and vibrant, with the need for much further research and development. But it is perhaps appropriate at this point to recall the fundamental rationale of the discipline–to help societies to cope with one of the most destructive forces of nature. This was perhaps summed up by McGonnogal over 100 years ago when he concluded his offering on the Tay Bridge disaster, with the last two lines summing up the whole discipline of wind engineering. Oh! ill-fated Bridge of the Silv’ry Tay, I must now conclude my lay By telling the world fearlessly without the least dismay, That your central girders would not have given way, At least many sensible men do say, Had they been supported on each side with buttresses, At least many sensible men confesses, For the stronger we our houses do build, The less chance we have of being killed

Appendix. The Tay Bridge Disaster by William McGonnegal Beautiful Railway Bridge of the Silv’ry Tay! Alas! I am very sorry to say That ninety lives have been taken away On the last Sabbath day of 1879, Which will be remember’d for a very long time. ‘Twas about seven o’clock at night, And the wind it blew with all its might, And the rain came pouring down, And the dark clouds seem’d to frown, And the Demon of the air seem’d to say‘‘I’ll blow down the Bridge of Tay.’’ When the train left Edinburgh The passengers’ hearts were light and felt no sorrow, But Boreas blew a terrific gale, Which made their hearts for to quail, And many of the passengers with fear did say‘‘I hope God will send us safe across the Bridge of Tay.’’ But when the train came near to Wormit Bay, Boreas he did loud and angry bray, And shook the central girders of the Bridge of Tay On the last Sabbath day of 1879, Which will be remember’d for a very long time.

ARTICLE IN PRESS 868

C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

So the train sped on with all its might, And Bonnie Dundee soon hove in sight, And the passengers’ hearts felt light, Thinking they would enjoy themselves on the New Year, With their friends at home they lov’d most dear, And wish them all a happy New Year. So the train mov’d slowly along the Bridge of Tay, Until it was about midway, Then the central girders with a crash gave way, And down went the train and passengers into the Tay! The Storm Fiend did loudly bray, Because ninety lives had been taken away, On the last Sabbath day of 1879, Which will be remember’d for a very long time. As soon as the catastrophe came to be known The alarm from mouth to mouth was blown, And the cry rang out all o’er the town, Good Heavens! the Tay Bridge is blown down, And a passenger train from Edinburgh, Which fill’d all the peoples hearts with sorrow, And made them for to turn pale, Because none of the passengers were sav’d to tell the tale How the disaster happen’d on the last Sabbath day of 1879, Which will be remember’d for a very long time. It must have been an awful sight, To witness in the dusky moonlight, While the Storm Fiend did laugh, and angry did bray, Along the Railway Bridge of the Silv’ry Tay, Oh! ill-fated Bridge of the Silv’ry Tay, I must now conclude my lay By telling the world fearlessly without the least dismay, That your central girders would not have given way, At least many sensible men do say, Had they been supported on each side with buttresses, At least many sensible men confesses, For the stronger we our houses do build, The less chance we have of being killed. References ABI, 2003. The vulnerability of UK property to windstorm damage, Association of British Insurers Report, 47pp. Aynsley, R.M., Melbourne, W., Vickery, B.J., 1977. Architectural Aerodynamics. Applied Science Publishers Ltd., London.

ARTICLE IN PRESS C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

869

Bailey, A., 1933. Wind pressures on buildings, ICE Selected Papers, 139, London. Bailey, A., Vincent, N.D.G., 1943. Wind pressures on buildings, including effects of adjacent buildings. J. Inst. Civil Eng. 20, 243–275. Baker, B., 1884. The forth bridge. Engineering 38. Baker, C.J., 2004. Wind effects in the urban environment—considerations for human health, comfort and safety, ISWE Tokyo. Blackmore, P., Delpeche, P., 2002. A comparison of wind damage to structures in France and the UK. COST C14 Conference, Nantes. British Standards Institution, 1944. Code of functional requirements of buildings. Loading Code of Practice CP4 (Chapter V). British Standards Institution, 1972. Code of basic data for the design of buildings. Chapter V loading. Part 2 Wind loads. Code of Practice CP3. Browning, K.A., 2003. The sting at the end of the tail: damaging winds, Forecasting Research Technical Report 388, The Met. Office, UK. Browning, K.A., Field, M., 2004. Evidence from Meteosat imagery of the interaction of sting jets with the boundary layer. Meteorol. Appl. 11 (4), 277–290. Buller, P.S.J., 1988. The October gale of 1987: damage to buildings and structures in the south east of England. Building Research Establishment Report 138. Cook, N.J., 1985. The Designers Guide to Wind Loading of Building Structures. BRE, Butterworths. Courtney, M.S., Troon, 1990. Wind speed spectrum for one year of continuous 8 Hz measurements. In: Proceedings of the 9th AMS Symposium on Turbulence and Diffusion, Boston, pp. 301–304. Cunningham, A., 2005. The direct use of wind tunnel time history pressure data in structural design. Presentation made at UK Wind Engineering Society, London 1/2/5. Davenport, A.G., 1961. The application of statistical concepts to the wind loading of structures. Proc. ICE 19, 449–471. Davenport, A.G., 1995. The role of wind engineering in the reduction of natural disasters. In: Proceedings of the 9th International Conference on Wind Engineering, vol. 5, Delhi, India, pp. xi–xviii. Davenport, A.G., 1999. The missing links, In: Proceedings of the 10th International Conference on Wind Engineering, Copenhagen, pp. 3–15. Defoe, D., 1704. The Storm. Eaton, K., Mayne, J.R., 1975. The measurement of wind pressures on two storey houses at Ayelsbury. J. Ind. Aerodyn. 1 (1), 67–109. Encyclopedia Britannica, 2004. /http://www.britannica.comS. ESDU, 2004. Engineering Sciences Data Unit Wind Engineering Series, vols. 1–4 (9 parts). Fisher, R.A., Tippett, L.H.C., 1928. Limiting forms of the frequency distribution of the largest or smallest member of a sample. Proc. Cambridge Philos. Soc. 24, 180–190. Flag Fen, 2004. Flag Fen—Britain’s Bronze Age Centre http://www.flagfen.com/www.flagfen.comS. Frazer, R.A., Scruton, C., 1952. A summarized account of the Severn Bridge aerodynamic investigation. Aerodynamics Division National Physical Laboratory NPL/Aero/222. Gurmley, K., Masters, F., Prevatt, D., Reinhold, T., 2005. Hurricane data collection: FCMP deployments during the 2004 Atlantic Hurricane Season. In: 10th Americas Conference on Wind Engineering, Baton Rouge, Louisiana. Hill, D., 1984. A History of Engineering in Classical and Medieval Times. Routledge, London. Hoxey, R.P., Richards, P., Richardson, G.M., Robertson, A.P., Short, J.L., 1995. The Silsoe structures building; the completed experiment, Part 2. In: Proceedings of the 9th International Conference on Wind Engineering, Delhi, pp. 1115–1126. Hoxey, R.P., Robertson, A.P., Richardson, G.M., Short, J.L., 1997. Correction of wind tunnel pressure coefficients for Reynolds number effect. J. Wind Eng. Ind. Aerodyn. 69–71, 547–555. Irminger, J.O.V., 1895. Nogle Forsog over Trykforholdene paa Planer og Legemer paavirkede of Luftstromninger. Engineering News and Engineers. Irminger, J.O.V., Nokkentved, C., 1936. Wind Pressures on Buildings; Experimental Researches. Danmarks Naturvidenskabelige, Samfund. Jensen, M., 1958. The model law for phenomenon in the natural wind. Ingenioren Int. Ed. 2. Jensen, N.O., 1999. Atmospheric boundary layers and turbulence. In: Proceedings of the 10th International Conference on Wind Engineering, Copenhagen, pp. 29–42.

ARTICLE IN PRESS 870

C.J. Baker / J. Wind Eng. Ind. Aerodyn. 95 (2007) 843–870

Kareem, A., 2003. A tribute to Jack E Cermak—wind effects on structures: a reflection on the past and outlook for the future. In: Proceedings of the 11th International Conference on Wind Engineering, Lubbock, pp. 1–28. Larose, G.L., D’Auteuil, A., 2004. Experiments on 2-D rectangular prisms at high Reynolds numbers in a pressurized wind tunnel. In: Proceedings of the 4th Conference on Bluff Body Aerodynamics and its Applications, Ottawa, pp. 177–180. Levitan, M.L., Mehta, J., 1991a. Texas Tech field experiments for wind loads. Part 1. Building and pressure measuring system. In: Proceedings of the 8th International Conference on Wind Engineering, London, Ontario, pp. 1565–1576. Levitan, M.L., Mehta, J., 1991b. Texas Tech field experiments for wind loads. Part 1. Meteorological Instrumentation and terrain parameters. In: Proceedings of the 8th International Conference on Wind Engineering, London, Ontario, pp. 1577–1588. Marshall, R.D., 1975. A study of wind pressures on a single family dwelling in model and full scale. J. Ind. Aerodyn. 1 (2), 177–199. McDonald, R., 2004. Modelling response of mid-latitude storms to climate change. In: Proceedings of the UK Wind Engineering Society Conference, Cranfield. Meroney, R.N., 1999. Perspectives on air pollution aerodynamics. In: Proceedings of the 8th International Conference on Wind Engineering, Copenhagen, pp. 79–90. Meteorological Office, 2005. http://www.met-office.gov.uk/education/historic/1987.html/www.met-office. gov.uk/education/historic/1987.htmlS. Milford, R.V., Waldeck, J.L., Goliger, A.M., 1991a. Jan Smuts experiment—details of the full scale experiment. In: Proceedings of the 8th International Conference on Wind Engineering, London, Ontario, pp. 1693–1704. Milford, R.V., Goliger, A.M., Waldeck, J.L., 1991b. Jan Smuts experiment—comparison of full scale and wind tunnel test results. In: Proceedings of the 8th International Conference on Wind Engineering, London, Ontario, pp. 1705–1716. Miller, C., 2003. A once in 50 year wind speed map for Europe derived from mean sea level pressure measurements. J. Wind Eng. Ind. Aerodyn. 91, 1813–1826. Munich Re, 2002. Winter Storms in Europe, Munich Re Publication. Murakami, S., Mochida, A., 1999. Past, present and future of CWE. The view from 1999. In: Proceedings of the 10th International Conference on Wind Engineering, Copenhagen, pp. 91–104. Pryor, F., 2003. Britain BC. Harper Collins, London. Rathbun, J.C., 1940. Wind forces on a tall building. ASCE Trans. 105, 2056. Richards, P.J., Hoxey, R.P., Short, J.L., 1999. Spectral models of the atmospheric surface layer. In: Proceedings of the 10th International Conference on Wind Engineering, Copenhagen, pp. 315–321. Richardson, G.M., Hoxey, R.P., Robertson, A.P., Short, J.L., 1995. The Silsoe structures building; the completed experiment. Part 1. In: Proceedings of the 9th International Conference on Wind Engineering, Delhi, pp. 1103–1114. RMS, 2003. 1703 windstorm—a 300 year retrospective. Risk Management Solutions Report, 6pp. Scruton, C., 1963. On the wind excited oscillation of stacks, towers and masts. In: International Conference on Wind Effects on Buildings and Structures, National Physical Laboratory, London. Shropshire Tourism, 2004. http://www.shropshiretourism.info/ironbridge/www.shropshiretourism.info/ ironbridgeS. Solari, G., 2005. The International Association for Wind Engineering (IAWE): birth, development and perspectives. Keynote paper. In: Proceedings of the 4th European and African Regional Conference on Wind Engineering, Prague. Standards Association of Australia, 1973. SAA loading code Part 2—Wind forces As 1170 Part 2, Sydney. Surrey, D., 1999. Wind loads on low rise buildings; past, present and future. In: Proceedings of the 10th International Conference on Wind Engineering, Copenhagen, pp. 105–116. van der Hoven, I., 1957. Power spectrum of horizontal wind speed in the frequency range from 0.0007 to 900 cycles/hour. J. Meteorol. 14, 160–164. Whitlock, D., 1961. The Anglo-Saxon Chronicle—a Revised Translation, Eyre and Spottiswoode. Wind and Structures, 2005. In: Holmes, J. (Ed.), Special Edition on the IAWE Codification Initiative. Wyatt, T.A., Walshe, D.E., 1992. Bridge aerodynamics 50 years after Tacoma Narrows—1. The Tacoma failure and after. J. Wind Eng. Ind. Aerodyn. 40, 317–326.