Wind Classification Maitland, Yorke Peninsula and Surrounds

Introduction to Wind Speed Classification

In Australia we have two wind loading codes AS1170.2 and AS4055. The latter (AS4055) whilst it refers to housing, was created to provide a simplified classification system for use with manufactured structural products (MSP’s). It is neither practical nor economical to use AS1170.2 directly to design a structured product. The issue is that the loads which may be experienced are infinitely variable, and it is not economical to custom design a structural element to suit each and every project. Rather products are made in a range of sizes, and each item in the range also has a limited range of suitability for a given purpose.

For example bolts vary in size as follows: M10, M12, M16, M20. The bolts do not vary in size by 1mm increments, whilst it is possible to make a bolt of any specific size wanted, such bolts are not readily available and would typically have to be custom made. The size increments for a range typically follows a series of preferred numbers: with a common series being the Renard Series. A geometric sequence is typically chosen to define the sizes, as it provides small incremental changes at the lower end of the scale and larger increments at the upper end. The scale thus generally favours using the standard sizes at the lower end of the size ranges, and otherwise favours custom design and the larger sizes.

The wind classification system however is not based on sizes and resistances of structural elements, it is based on what appears to be a relatively arbitrary division of an assumed maximum design load into a series of classes. The classification system given in AS4055 is as follows:


The first thing to note is there is no 28 m/s, 33 m/s  or 41 m/s wind speeds in the  table. This is because such speeds became obsolete around 1989 when AS1170.2 first introduced limit state design. When AS4055 was first released in 1992, not all codes had been converted from permissible stress design to limit state design: consequently the two different loading methods were maintained. There has been no increase in the load the building is designed as a consequence of the change from permissible stress design to limit state design. Whilst the wind speed has increased from say 28 m/s for wind class N1 to 34 m/s, the change does not change the design load. The load derived from a speed of 28 m/s has to be multiplied by a design factor of 1.5 before it is used in design, whilst the load derived from the 34 m/s speed is multiplied by a design factor of 1. For limit state design the wind speed relates to the state of the building when it experiences such wind load. For permissible stress design, the wind speed chosen does not relate to the state of the building, as the building is designed for a magnified load: and therefore the design load does not relate to the loading experienced at the selected wind speed.

So whilst we have a Adelaide metropolitan wind speed map, with speeds of 28. 33, 41 m/s indicated, no council should be advising applicants for development approval that they should get the house of other structure designed for a wind speed of 41 m/s unless they identified the method of design (permissible stress or limit state). Rather they should be indicating a need to have designed for wind class N3, if  they are using permissible stress wind speeds. Alternatively they can indicate they don’t know and an individual assessment is required.

The Adelaide wind speed map is rough and ready, it is not very reliable. It is not reliable because cannot assess the site wind speed by driving around the neighbour and making a judgment call for large areas. The site wind speed has to be determined by checking the environment in the surrounds of the the site. On a large area map terrain category could be identified, but shielding and topography cannot. Even terrain category is questionable, since need to make an assessment for 8 cardinal compass directions. An extreme assessment for example would put a road at terrain category 2, whilst the rest of the surrounds are terrain category 3. A house at the end of a T-junction road would have a wind speed determined from wind flow along the road.

To put simply most wind class N1 sites on the Adelaide wind speed map are not N1, they lack the full shielding required in all 8 directions. For example flow parallel to the road, the roofs of houses form a near continuous roof scape. The end houses have no shielding but he walls of the others are shielded by the neighbours. Across the back gardens, depending on the spacing of the houses and other structures in the space, the shielding is likely only partial. Whilst across the front gardens and across the street, the spacing is likely to high too provide shielding.

In similar manner the terrain across the neighbourhood is likely terrain category 3, but the terrain along the street and along the back gardens, have terrain category closer to terrain category 2 or 2.5. The problem however is that these differences push the wind speed into the lower region of the range for wind class N2, and the design pressure for wind class N2 is significantly different than that for wind class N1. So people paying for the structure obviously prefer the initially lower cost structure resulting from N1 design loads.

By similar measures most sites identified as wind class N3 on the map are at the upper end of the range for wind class N2 or lower end of the range for wind class N3. If at the upper end of wind class N2, then using wind class N2 for design is economical. If at the lower end of N3, then using wind class N3 is likely uneconomical. In such situation it is better to design direct to AS1170.2 rather than use the wind classification system. If the design load is within 5% to 10% of the maximum design load for wind class N2, then selection of wind class N2 components may be a more economical option. Noting that we don’t actually know the magnitude of the wind load which will be experienced, we merely have an approximate statistical model of the likely wind speeds which will be experienced: that is we have a sophisticated guess,but nonetheless it’s still a guess. Other factors also come into consideration: windows for example do not need to be wind class N3, if they are otherwise shielded from the weather.

The point is that wind load needs to be assessed for the site, so local councils requiring residents of rural and coastal regions to have their houses designed for wind class N3 , or 41 m/s as they usually demand: is not acceptable.


So here I am sat in the rural service town of Maitland on the Yorke Peninsula in South Australia, and documents I have seen thus far all indicate adoption of wind class N3. Wind class N3 is difficult to get using AS1170.2, easy if use the simplified symbolic tabular approach of AS4055. So if in a rural or coastal area do not use AS4055, it is preferable to get an assessment to AS1170.2.

Maitland is basically a square plan town, similar plan to Adelaide. the main township is approximately a 1 km square, surrounding this are parklands and miscellaneous buildings and additional housing: extending to the square to approximately 2km. For buildings less than 50m high, the terrain averaging distance is 1000 m (AS1170.2 Table 4.2(a)).

Reference Wind Speed

The site reference wind speed is calculated from :

Vzu = Vr.Md (Mz,cat. Ms. Mt)

This has to be calculated for 8 compass directions (N,E,S,W, NE, NW, SE, SW). This is then converted into two orthogonal wind speeds along the axes of the building. Wind directions are identified by the symbol theta. The main directions for design are: theta=0 and opposing theta=180, and theta=90 and opposing 270 degrees.  The two orthogonal wind speeds are typically reduced to a site maximum wind speed, and only one wind speed is used for design. Though if one direction does have a lower wind speed it can be beneficial for designing a more economical building.

The multipliers account for the variation of wind speed relative to the location that the wind speed was actually measured for a region. Typically wind speed is measured at airports and airfields, 10m above the ground, in terrain defined as terrain category 2. This gives a terrain category multiplier Mz,cat=1. This multiplier is used to account for the variation of wind speed with height above a surface. The further from the surface the less drag the surface has on the flow of the fluid mass (air). Different surfaces have different roughness, and therefore produce less or more drag than the reference terrain at the airport/airfield.

Md accounts for variation of the regional reference wind speed (Vr), with direction. The multipliers Mz,cat, Ms, and Mt, collectively account for the influence of the environment. The multipliers are collected together in brackets, because the international guidelines (ISO) allow a single factor to be used, and many other countries do only use a single factor. In Australia however we have separate multipliers for terrain category (Mz,cat), shielding (Ms) and topography (Mt). The symbol z, refers to the reference height, which is typically the average height of the building being considered.

If there is no shielding then Ms=1, and if no topography such as hills and mountains then Mt=1. Properly assessing shielding and topography can make the wind speed assessment significantly more complicated. For the following exercise I will take ms=1, and Mt=1.

Terrain Category

Terrain category is required to get the terrain category multiplier Mz,cat. Terrain category is itself dependent on surface roughness length (z0).  A chart of surface roughness length is provided in AS1170.2 commentary, as well as reference photos for terrain category. Only terrain categories 1,2 ,3 and 4 are directly covered by AS1170.2 tables, additional calculations are required for other terrain categories. The appendices to As1170.2:1989 provided formulae for converting surface roughness to terrain category.

The importance of this is that rural terrain is not terrain category 2: it seldom ever has the appearance of an airfield. The surface roughness length of an airfield is z0=0.02m whilst land covered in crops is z0=0.04m, and sparsely scattered trees or long grass is z0=0.06. Rural terrain changes from having crops to having rough clods of soil in a ploughed field. Rural terrain varies away from terrain category 2, but doesn’t reach the z0=0.2m for wooded country or suburban buildings. Rural terrain also varies throughout the year.  The equivalent terrain categories are:


Applying this to the terrain profile of Maitland get the following diagram. In the diagram I have allowed for 1km beyond the edge of the town, because of the required 1000m averaging distance for terrain category.

Maitland Wind Terrain Profile
Maitland Wind Terrain Profile

The wind speed for a site now depends on its location and which terrain has dominant influence on the fluid (air) speed.

If take a building at the centre of the township, then the 1000 m averaging distance extends to the edge of the town, and includes TC3 and TC2.5. The wind speed isn’t just influenced by the buildings of the township, it is influenced by the surrounding parklands and more rural housing blocks.

Since the terrain varies, can either make a judgement of the terrain, or use the more formal method of terrain category averaging given in AS1170.2. The following diagram illustrates the basic principle of terrain category averaging for three terrain categories.

Wind Terrain Category Averaging
Wind Terrain Category Averaging

The new terrain category doesn’t immediately influence the wind speed, there is a lag, with the first terrain category continuing to have an influence. It is therefore necessary to calculate these lag distances. For the case of a building located in the centre of Maitland get the following assessment:

Wind Speed Centre of Maitland township (SA)
Wind Speed Centre of Maitland township (SA)

So whilst it may be a township, it is not fully suburban and the wind class is not N1, it pushes into the lower end of wind class N2. If use As1170.2 for design then the pressure is 92% of that for wind class N2, so some saving may be available. But still need to adopt wind class N2 components where custom product not available. That is the tie-down and bracing system can be designed to AS1170.2 but would still need windows to suit wind class N2.

If the building is now moved to the edge of the town. There the terrain comprises of the parklands, transitioning into rural crop land.

Wind Speed Edge of Maitland township (SA)
Wind Speed Edge of Maitland township (SA)

On the edge of town there is not much difference between the AS1170.2 derived speed and wind class N2. So in this situation, can simply adopt wind class N2 throughout the design of the building.

Now if move away from the township and into rural farmland, then the terrain comprises of crops and can adopt terrain category 2.3. The assessment only involves terrain of one type, so no terrain category averaging is required.

Wind Speed in Areas Surrounding Township of Maitland(SA)
Wind Speed in Areas Surrounding Township of Maitland(SA)

For the surrounds the calculated wind speed is 40.05 m/s which exceeds the 40 m/s which defines the end of wind class N2.  However, whilst strict adherence to the rules would push us into wind class N3, doing so is not sensible, the calculated wind pressure matches that for wind class N2. If the calculated wind speed was rounded down then would get wind class N2. {The calculation shows wind class N3, because the computer spreadsheet calculation doesn’t have any tolerance.} 

So unless near the top of a hill, or down on the coast, the rural regions are typically wind class N2 for single storey dwellings. In the above calculations I adopted a maximum width building of 7.2m, with a typical roof pitch of 22 degrees., and calculated the maximum height: which was just a fraction less than 4.5 m. However, AS4055 is based on a maximum height of 8.5m, this height cannot be varied if using the simplified approach of AS4055, as all the Mz,cat multipliers used by AS4055 are based on a height of 8.5m. Using As1170.2 the height can be changed and the terrain category can also be refined.

So if choose a 2 storey house, then will likely push the wind class upto wind class N3.

Being on the coast with single storey dwelling doesn’t necessarily push wind class upto N3, as the water is not terrain category 1. Strong winds generate high waves which increases the surface roughness length z0, therefore for ultimate strength design AS1170.2 commentary suggests using terrain category 2 for strength design and terrain category 1 for serviceability design. Serviceability design typically checks deflections at wind speeds lower than ultimate strength design: at the lower wind speeds the water surface is likely calm.

If have terrain category 2, then the results to AS1170.2 are likely located in the lower range of wind class N3, and therefore design to As1170.2 is likely more economical than design to N3. Though some components may still need selecting based on wind class N3.

Suitability of Manufactured Structural Products

The suitability of a manufactured structural product (MSP), such as window, depends on its resistance to applied loads. whilst windows are considered non-structural, they nevertheless have to resist structural loads: they are only non-structural in the sense, that no load should be intentionally placed on the window glass or its frame. The purpose of the window is like the wall, and that is to keep the wind and rain out. It therefore has to have adequate resistance to such loading. Similarly carports and sheds need to have adequate resistance to  the wind loads they may experience.

Few if any manufactured structural products are designed directly to AS4055 wind classes. Rather the product is designed subject to other requirements, then it is structurally assessed and assigned a wind class. The wind class which can be assigned to the MSP is the class with the maximum wind speed it can resist.

So for example, assuming I designed a transportable building to AS1170.2, and used the wind speed 38 m/s as calculated above. Whilst 38 m/s is greater than the 34 m/s for wind class N1, it is otherwise less than the 40 m/s required for wind class N2. Therefore the wind class for the transportable house is wind class N1.

However we look at the site defined above and get wind class N2 for the site. Our conclusion therefore is that the available transportable house is not suitable. It is not suitable because its wind class is N1, and we need a wind class N2 house, to place on the site. Yet to AS1170.2 both site and house are compatible: because both site and house have a design wind speed of 38 m/s.

From a manufacturers viewpoint therefore it is not beneficial to design to AS1170.2 if wind classifications are to be used and the product is going to be excluded from a large number of projects for which it is actually suitable. Better to design to the N2 wind class, so that the building can be selected and put to use.

On the other hand, the wind classes are not a sensible division of the range of applicability.  The incremental leap from one class to the next is too large. Barring some historical discrepancies and rounding, the basic ratio between one class and its neighbour is 1.5, which seems too high and not very practical. Furthermore the wind classification system ignores the minimum permitted wind speed of 30 m/s {or 25 m/s if using the old permissible stress approach}.

For the purpose of including the minimum I suggest introducing a wind class of N0. Further additional symbols are required if the system is to provide a more refined approach. Alternatively manufacturers could ignore AS4055 and introduce their own system, optimised to best suit the product which they are manufacturing.

The regional reference wind speed is 45 m/s [162 km/h], this speed is typically reduced to account for the environment at the site. It is rare the wind speed is increased to anywhere near 50 m/s [180 km/h] here in South Australia.

However if designing transportable buildings, which need to be robust enough to be transported, and which can be sold on and moved from their original site, it makes sense to adopt wind class N3 as the minimum design wind speed. By adopting N3, the building is more robust for transport and there is more of South Australia which the building can be transported to without having to get an assessment made on the suitability of a second hand building for the site. On the other hand life would be a lot easier if the manufacturers placed a plate on the building with a serial number and wind class. Not that such helps with old buildings built before such idea implemented.


[(10/04/2018)] : Original [First Draft]