Creating weather files

A collection of weather files is supplied with WUFI. Please note the sources given below from which you can obtain weather data for additional locations. Unless noted otherwise, WUFI can read the respective data formats; it is for the user to decide whether the data are suitable for the purpose at hand.

Depending on the aim of the investigation, appropriately selected weather data should be used. For example, if the typical long-term moisture balance of a component is to be determined, typical weather data which are representative for the location are appropriate. For design purposes, the performance or durability of a component exposed to more severe conditions is of interest, and weather data describing such a more demanding scenario should be used. If the cause of a specific damage case is to be investigated, weather data for the relevant time and location are needed.

Requirements Concerning Quality and Scope of the Weather Data

Required weather elements

WUFI can read a variety of weather data formats. Different weather data formats may comprise different sets of weather elements. All common weather data formats contain temperature and relative humidity; solar radiation is sometimes only given in terms of global radiation; rain is often missing. Therefore, the user must make sure that the weather file to be used contains at least all the weather elements which are needed for the planned investigation. The same applies to weather data which the user has obtained in order to create a weather file to be read by WUFI: it must be made sure that all weather elements are present which are needed for this simulation, and a file format must be chosen which can accomodate all these weather elements.

The weather file formats which WUFI can read are described at length in WUFI’s on-line help. For newly created files we recommend the WAC format which is flexible with respect to the number of contained weather elements. Thus, there is no need to obtain weather elements which are not needed for the simulation.
Furthermore, directional weather elements (i.e. radiation and rain) must be converted for the orientation and inclination of the component to be simulated, which is done automatically by WUFI, using simplified models. The WAC format allows to bypass these models by providing ready-made radiation and rain data specific to the orientation and inclination of the component in question. These specific data may have been produced by more refined conversion models or by on-site measurements.

• Temperature, Relative Humidity

Usually, for hygrothermal simulations at least temperature and (unless a purely thermal calculation is intended) relative humidity are needed.

• Solar Radiation

Unless the component is completely shaded or you intend a simulation “on the safe side” without allowing for the additional drying potential due to solar heating, solar radiation has to be taken into account.

Solar radiation is a directional quantity, affecting differently oriented and inclined surfaces to different degrees. In the ideal case radiation data measured on a surface with the required orientation and inclination are available. If the WAC format is used, these data can be declared as measured data, prompting WUFI to use them without modification. These measured data can describe the effect of solar radiation with the highest accuracy (including all environmental effects, such as shading, reflections etc.), but they are confined to use for this individual surface situated in this individual environment.

Usually, however, only data for the radiation incident on a horizontal surface are available. In particular, this is the case when the radiation data have been supplied by a normal meteorological station. Unless a flat roof is to be simulated, the available “horizontal” data must be converted for a surface with the required orientation and inclination. WUFI automatically performs this conversion during the simulation run. Since the diffuse and the direct radiation component have different transformation properties, the conversion needs separate data on the diffuse and the direct component. Therefore, the weather file has to contain either diffuse and direct or diffuse and global radiation (global = sum of diffuse and direct). Many meteorological stations only measure global radiation. If needed, the corresponding diffuse radiation may be estimated from the global radiation, using a statistical-meteorological model (e.g. “Perez model”). If only horizontal surfaces are to be simulated, global radiation data may be sufficient, because in that case no directional conversion is necessary.

Since the directional conversion involves determining the solar position in the sky, correct time stamps for the radiation measurements are indispensable. Did the measuring system have the correct time? Do the time stamps refer to zone time or true solar time? Do the time stamps refer to UTC, or the local time zone or some other time zone? Do the time stamps switch to daylight saving time and when? Are the radiation data single measurements acquired at the full hour, or are they average values over one hour (accordingly, the solar position representative for the measured data is either the position at the full or at the half hour; WUFI assumes that the half hour is representative for the measurements).

• Normal Rain, Driving Rain

If the effect of rain shall be investigated or at least has some non-negligible influence on the component behavior, data describing rain intensity are needed. Unless a flat roof is being simulated, the driving rain hitting the surface under investigation is the relevant quantity (whereas the rain hitting a horizontal surface is called “normal rain”).

Driving rain is a directional quantity, affecting differently oriented and inclined surfaces to different degrees. Ideally, you can use rain data measured on a surface with the required orientation and inclination. If the WAC format is used, these data can be declared as measured data, prompting WUFI to use them without modification. These measured data can describe the effect of driving rain with the highest accuracy (including all environmental effects, such as local wind patterns, sheltering effects etc.), but they are confined to use for this individual surface situated in this individual environment.

Usually, however, only data for the rain incident on a horizontal surface are available. In particular, this is the case when the rain data have been supplied by a normal meteorological station. Unless a flat roof is to be simulated, the driving rain hitting the inclined surface has to be estimated from the data on normal rain and wind speed and direction. WUFI automatically performs this evaluation during the simulation run.

• Wind

Data on wind speed and direction are only needed if the driving rain has to be estimated from the normal rain, or if a wind-dependent heat transfer coefficient shall be used.

• Atmospheric and Terrestrial Long-wave Counterradiation

Any surface of a building component continuously exchanges long-wave (“thermal”) radiation with its environment. Due to its temperature, the building surface emits thermal radiation into the surroundings. On the other hand, it also receives thermal radiation emitted by the surroundings (so-called counterradiation). In particular, it receives terrestrial counterradiation, which is thermal radiation emitted by the ground and other terrestrial objects, and atmospheric counterradiation, which is thermal radiation emitted by the atmosphere.

In practical building physics, this radiative heat exchange is usually not accounted for separately, it is instead lumped together with convective heat transfer by adding some appropriate but fixed radiative contribution to the convective heat transfer coefficient. For most calculations in building physics this approximation is sufficient, and WUFI uses it when working in standard mode (i.e. with “explicit radiation balance” switched off). In this case, no data on atmospheric or terrestrial long-wave counterradiation are needed.

There are, however, applications where long-wave radiation exchange must explicitly and separately be taken into account. This is the case, for example, if the surface temperatures need to be computed with higher accuracy.
Furthermore, the conventional treatment by adding a radiative contribution to the convective heat transfer coefficient assumes that convective and radiative heat flow go in the same direction. This is often the case, but not always. In particular, when night-time radiative cooling causes the surface to overcool (i.e. to cool below ambient air temperature), the surface loses radiative heat towards the sky, while the convective heat flow is directed from the warmer air towards the cooler surface. Therefore, if night-time overcooling and its consequences for dew deposition or reduced drying potential (for flat roofs in particular) shall be simulated, WUFI must be switched to “explicit radiation balance” mode.

As a consequence, however, the user then must also provide data describing the long-wave emissions of the atmosphere and of the ground (if the latter is within the field of view of the component). Ideally, the incoming long-wave radiation has been measured on a surface with the required orientation and inclination. If the WAC weather file format is used, these data can be declared as measured data, prompting WUFI to use them without modification. These measured data can describe the effect of atmospheric and terrestrial counterradiation with the highest accuracy (including all environmental influences, taking into account all the individual atmospheric and terrestrial areas and objects in the component’s field of view, etc.), but they are also confined to use for this individual surface situated in this individual environment.

Usually, however, only data for the atmospheric counterradiation incident on a horizontal surface are available. In particular, this is the case when the radiation data have been supplied by a normal meteorological station. Unless a flat roof is to be simulated, the available atmospheric counterradiation and the measured or computed terrestrial counterradiation must be converted to give the total long-wave radiation incident on the inclined surface. WUFI automatically performs this conversion during the simulation run.

Since night-time overcooling is often relevant for the simulation but counterradiation data are rarely available, WUFI can estimate the atmospheric and terrestrial thermal emissions from other weather elements. For this purpose, it can use a variety of simplified models, depending on which weather elements are known. For example, the atmospheric counterradiation can be estimated from hourly data on the cloud cover, if these are available. If cloud cover data are missing, too, WUFI can instead use a constant cloud cover specified by the user (see the on-line documentation for the WAC weather data format for details). The accuracy of these simplified estimation methods decreases however, the more simplifications must be made. It is the responsibility of the user to provide sufficient data with an adequate accuracy level if the simulation results are required to reach a given accuracy level. Test simulations can help to determine the sensitivity of the intended investigation with respect to different quality levels of input data.

• Barometric Pressure

Barometric pressure is usually of minor importance for hygrothermal simulations. It slightly affects vapor transport (the diffusion coefficient in air is weakly pressure-dependent). If no barometric pressure is provided in the weather data, WUFI uses the altitude of the location to estimate a typical pressure. If the altitude is unknown, WUFI uses some standard pressure.

Summary

If a standard multi-purpose weather data set with a frequently used combination of data is to be created, the weather elements highlighted below usually lend themselves to inclusion in such a file. Depending on the availability of data or requirements for specific applications, other combinations of weather elements are possible, of course.

Data sets containing data which have been measured on a surface with the required orientation and inclination offer the most accurate description of the environmental influences, but they can only be used for this specific surface.
Data sets containing the usual weather elements as measured on horizontal surfaces are multi-purpose data sets which can be used for surfaces with arbitrary orientation and inclination. The necessary conversion of the directional weather elements for non-horizontal surfaces involves some loss of accuracy, which will however be acceptable for most hygrothermal simulations.

•	Temperature
	always needed (except for purely hygric simulations)
•	Relative humidity
	always needed (except for purely thermal simulations)
•	Solar radiation
	always needed (except for simulations of completely shaded components or calculations “on the safe side”, ignoring the drying potential due to solar heating)
	If a flat roof is to be simulated:	• global radiation on a horizontal surface.
	Otherwise, sorted by preference:	• global radiation, measured on a surface with the required orientation and inclination. Is rarely available and can only be used for this individual surface. Used by WUFI without further modification.
		• a combination of global + diffuse or direct + diffuse or direct normal (“beam”) + diffuse, measured on a horizontal surface. If a meteorological station measures more than just global radiation, it’s usually global + diffuse. WUFI uses a simplified model to convert the radiation for the required orientation and inclination.
		• global radiation measured on a horizontal surface, diffuse radiation can be estimated using the Perez or a similar model. Many meteorological stations only measure global radiation. WUFI uses a simplified model to convert the radiation for the required orientation and inclination.
•	Rain
	is needed if the effect of rain shall be investigated or at least has some non-negligible influence on the component behavior.
	If a flat roof is to be simulated:	• normal rain Used by WUFI without further modification.
	Otherwise, sorted by preference:	• driving rain, measured on a surface with the required orientation and inclination. Is rarely available and can only be used for this individual surface. Used by WUFI without further modification.
		• normal rain, wind speed and direction. Will be used by WUFI to estimate the driving rain via a simplified model.
•	wind speed and direction
	are only needed if the driving rain has to be estimated from the normal rain, or if a wind-dependent heat transfer coefficient shall be used.
•	Long-wave atmospheric and terrestrial counterradiation
	are only needed if surface temperatures have to be computed with higher accuracy and/or if night-time overcooling shall be taken into account.
	If a flat roof is to be simulated:	• atmospheric counterradiation as incident on a horizontal surface Used by WUFI without further modification.
	Otherwise, sorted by preference:	• long-wave irradiation (sum of atmospheric and terrestrial), measured on a surface with the required orientation and inclination. Is rarely available and can only be used for this individual surface. Used by WUFI without further modification.
		• atmospheric and terrestrial counterradiation on a horizontal surface. WUFI uses a simplified model to convert the long-wave radiation for the required orientation and inclination.
		• atmospheric counterradiation on a horizontal surface. WUFI estimates the terrestrial counterradiation by assuming that the terrestrial objects have air temperature and then uses a simplified model to convert atmospheric and terrestrial counterradiation for the required inclination.
		• cloud cover and terrestrial counterradiation on a horizontal surface. WUFI estimates the atmospheric counterradiation from air temperature, relative humidity and cloud cover; then it uses a simplified model to convert atmospheric and terrestrial counterradiation for the required inclination.
		• cloud cover. WUFI estimates the atmospheric and the terrestrial counterradiation from air temperature, relative humidity and cloud cover; then it uses a simplified model to convert both for the required inclination.
		• constant cloud cover. As in the above two items, but instead of hourly data on the cloud cover WUFI uses a constant value specified by the user.
•	barometric pressure
	is nice to have, but optional. If missing, WUFI erstimates a constant value from the altitude of the location. If the altitude is not known, WUFI uses a standard value. May refer to sea level or station height.

Time range

The weather data sets supplied by weather services or other sources usually comprise one year. For a hygrothermal simulation this means that the spectrum of hygrothermal boundary conditions occurring during one year is acting on the simulated construction. For some investigations, shorter periods of time may be sufficient.

For investigations which extend over several years, it is usually sufficient to repeat the same year an appropriate number of times, if this year is sufficiently representative for the purpose of the investigation. When WUFI reaches the end of the weather file before reaching the end of the simulation period, it jumps back to the beginning of the file and repeats reading the file until the end of the simulation is reached. Since WUFI does not take leap days into account and the weather file may not contain any, no increasing mismatch between calendar count and weather file occurs.

WUFI’s calculations can proceed with any arbitrary time step width, but the calculation step width must be coordinated with the step width of the weather data (one weather data step must contain an integer number of calculation time steps). Weather data are usually available in one-hour steps. This also happens to be an appropriate time step for most hygrothermal simulations. Hygrothermal processes usually are so slow that a shorter time step would only mean much wasted computational effort. On the other hand, the duration of quickly changing boundary conditions, such as rain or solar radiation, must be represented in a sufficiently realistic way. Here too, a one-hour time grid turns out to be adequate in general.

Accuracy

In general, the accuracy of data supplied by professional weather stations should be easily sufficient for WUFI simulations. If a quantitative comparison between simulation results and measured data is intended, some attention should be given to adequate accuracy, but in general hygrothermal simulations will make less demands on the weather data than, say, meteorological or climatological analyses. In particular, it should be kept in mind that the influence of the weather elements on the construction depends not only on the intensity of the weather elements themselves but also on the corresponding surface transfer coefficients, such as radiation absorptivities, heat transfer coefficients etc., for which often only approximate or estimated values are available. An increased accuracy of the weather data would thus increase the accuracy of the weather influence only to a minor degree.

For example, solar radiation first has to be converted for the orientation and inclination of the component, using a conversion model. This model makes simplifying assumptions concerning the distribution of diffuse radiation across the sky. If some ground is within the field of view of the component, it is usually difficult to correctly estimate its reflectivity for solar radiation (the visible ground may comprise areas with different and partly direction-dependent reflectivities, it may be covered with variable vegetation or variable snow cover…). Shading of the component surface (or of the ground which contributes reflected radiation) is not taken into account. The fraction of solar radiation absorbed by the component is described by a constant absorptivity which does not take into account any moisture or direction dependence and which usually can only be roughly estimated anyway.

The models WUFI uses for estimating driving rain are strongly simplified and they should not be expected to yield strictly correct quantitative determinations of the amount of driving rain. Furthermore, the influence of the surroundings (exposed / sheltered location, water running down from higher regions of the facade) can only be described very crudely by parameters supplied by the user. Another parameter to be specified by the user describes the losses due to some of the incident rain water splashing back off the surface; this parameter can only be roughly estimated as well.
On the other hand, modern facade surfaces are not very absorbent in general, so that only a small fraction of the incident driving rain can be taken up by the component (the remaining water runs off without being considered further). In these cases it does not matter whether the incident amount of driving rain has been determined correctly or has been under- or overestimated, as long as it is above the absorption capacity of the facade: the facade absorbs what it can take up, the rest runs off and its precise amount doesn’t matter. Therefore the simple driving rain models used by WUFI are usually sufficient.
Surfaces which are strongly absorbent and strongly exposed, such as fair-faced brick masonry, may possibly require higher accuracy for the amount of driving rain. If the accuracy of WUFI’s driving rain models is not sufficient, the user must provide driving rain data which have been determined by on-site measurements or by sufficiently detailed models. Declaring these data as “measured” in a WAC file prompts WUFI to use them without further modification.

In most cases, missing data can simply be replaced by plausibly interpolated data unless some critical event which is relevant for the investigation has occurred during the gap and has to be reconstructed with more sophisticated methods (e.g. a storm with considerable driving rain load).
Since the hygrothermal processes to be simulated are usually quite slow, the component will respond to short-term influences only to a minor degree. Its behavior will usually be determined mainly by longer-term temperature and humidity averages (an example for an exception: a single damage-inducing driving rain event). Therefore short (and sometimes moderately long) gaps can often be tolerated if they have been filled with typical or averaged data in such a way that after the gap the component has a similar temperature and humidity level as would have been the case after exposure to the real data.
In these cases, missing hours can be filled by interpolation between the remaining hourly values. Missing days can often be replaced by similar neighboring days. Sometimes, constant fill data corresponding to typical average values may be sufficient.

If there is doubt about how strongly uncertainties in the weather data will affect the simulation result and whether more precise data must be obtained, test calculations with test weather data which have been varied within their respective range of uncertainty can explore the sensitivity of the simulation results with respect to the uncertainties in the input data.

Typical Weather Data

Measured Weather Data

In order to create a typical weather data set, a popular procedure is to select a year deemed “typical” from a longer data series, following appropriate criteria. A “typical” year may also be spliced together from “typical” months or other time periods which may be taken from different years. The resulting weather year then has never existed in this form, but it is expected to represent “typical” conditions as well as possible.

The advantage of selecting sub-series from measured data series is that the data are real, so they represent possible weather situations and thus preserve some properties in a realistic way, such as:

Sequential order and frequency of typical weather patterns, for example
- Macro weather systems,
- autumnal fog-prone weather,
- thunderstorms with subsequent sunshine in summer (creating reverse diffusion),
- frost-dew cycles
- etc.
daily temperature, humidity and radiation profiles
short- and long-term variability ranges
correlations between the individual weather elements

The disadvantage of such a selection is that any selected real weather sequence is unlikely to be typical for all weather elements, that is, it is usually unable to contain typical sequences for temperature as well as humidity, rain, wind, etc. In most cases, a selection must therefore compromise between several conflicting requirements. Many of the existing “test reference years” have been created in this or a similar way for the purpose of energetic investigations and thus have been selected to mainly represent typical temperatures. Since the temperature conditions also dominate the results of many (but not all) hygrothermal investigations, thermally selected test reference years can often be considered as representative for typical conditions in general hygrothermal simulations. However, it is for the user to decide whether a particular weather file is indeed suitable for the intended purpose.

As an alternative to one single year that has been selected as being typical, a multi-year data set may be used, because nothing can be more representative for a time span of, say, ten years than these ten years themselves. Furthermore, a multi-year data set contains a wider spectrum of weather situations than a one-year file. This also includes more extreme but less frequent “stress” situations for the component, so that such a long-term data set may be used for design purposes (see below), even though it is at the same time “typical”.

Synthetic Weather Data

Alternatively, hourly weather data may be created synthetically. For example, there are various models which can estimate the amount of incoming solar radiation, based on the astronomical position of the sun and taking into account data describing the current cloud cover and the clearness of the atmosphere. Other models (e.g. the ‘Perez model’) can estimate the diffuse solar radiation from the measured global radiation, using empirically determined correlation functions.

Some test reference years contain a combination of measured weather data and weather elements produced by synthetic models. It is also possible to create a yearly data set in a completely synthetic way, for example by using appropriate models to add quasi-realistic hourly day-night variations to given monthly mean values. These models will usually contain some physical component which produces a realistic diurnal variation (taking into account the correlation with other weather elements, for example the relationship of the temperature variation with the incoming solar radiation, the relationship of the solar radiation with the current cloud cover, etc.) A statistically-driven component of the model may add some random variation within a given range.

In this way it can be assured that the mean values of the created weather elements coincide with the long-term averages for the location, that at the same time realistic hourly variations take place and that the correlations among the weather elements remain preserved. How close the results come to reality depends on the quality of the involved models and may be different for different aspects of the created data.

Weather Data for Design Purposes

Investigations for design purposes usually test the ability of a component to withstand more strainful conditions than typically encountered. This may be done by confirming adequate resilience under a defined exposure, or by explicitly finding the failure threshold. A typical weather year is usually not suitable for this; instead, a year should be selected which exerts an appropriate more pronounced load on the component. Depending on the damage scenario under investigation, different damage mechanism may have to be provoked, such as

excessive condensation within the component
(in cold or temperate climate: moisture from the interior air penetrating into a winterly cold wall; in tropical climate: moisture from the exterior air penetrating into a wall cooled by air conditioning),
excessive absorption of rain water
(in a region with large amounts of driving rain and for an unsheltered position of the component on the weather-exposed side of the building).
frost damage
(i.e. frequent frost-dew cycles during a moderately cold winter),
growth of algae or fungi, occurrence of wood rot
(caused by insufficient drying potential, by the vapor permeance of components being too high or too low, by undesigned moisture infiltration, etc.)

There exist a variety of methods and criteria for selecting or creating suitable weather data. For example, if the intention is merely to determine whether the component is able to withstand the weather which is typical for the most extreme location to be expected, it is sufficient to select representative weather data for that location.

If the intention is to test the durability under more extreme conditions to have some safety margin, strainful situations can simply be created by using measured multi-year data series for a given location or for the most extreme location to be expected. These longer data series automatically contain the entire spectrum of normal as well as strainful scenarios with the respective frequencies. Standard DIN EN 15026 recommends this as the most suitable data source and proposes to use ten or (if possible) more measured years. The advantage is that measured real weather sequences are being used for the simulation. In general, a ten-year period will contain extreme situations which are to be expected once in ten years. (A statistical analysis may be used to confirm that this expectation is actually realized in the used data sequence.) Besides the occasionally occurring strain situations, such a data set also contains realistically distributed recovery periods which may be desirable or undesirable for design purposes.

Alternatively, and requiring less computational effort, a single suitably selected or created “design reference year” may be used which contains within a single year the strongest strain on the component which is to be expected in, say, ten years. Depending on the damage mechanism to be tested, this may be a particularly warm, cold, moist or dry year. For this purpose, a single year meeting the required criteria may be selected from a multi-year measured data series. It is also possible to use a synthetically created year whose hourly weather elements have been generated based on the desired extreme values rather than the usual long-term average values.

A hybrid type between measured and synthetical weather data are measured data which have been made more extreme by suitable modification of the originally average or typical weather elements. For example, the measured rain rates or wind speeds may be increased by some factor. Since many damage mechanisms involve particularly high or low temperatures, it is often sufficient to shift the measured temperatures. For these purposes, standard DIN EN 15026 proposes temperature shifts of +-2K, arguing that the results represent years which are likely to occur once in ten years. The relative humidity remains unmodified in the process.

Standard DIN EN 15026 states that in “most moisture applications a once in ten years failure rate is usually considered to be acceptable”, so that the strength of the design test load can be derived from load scenarios which occur approximately once in ten years. “However, in particularly sensitive applications, such as computer centres, art galleries or hospitals a lower failure rate might be required” and more severe load scenarios may be needed for design tests.

Occasionally strain scenarios may be created while using typical weather data. Modifying the absorptivity of the component surface can simulate reduced or increased solar irradiation, creating a reduced drying potential or a higher vapor drive towards the component’s interior, respectively. Modifying the driving rain coefficients or the adhering fraction of rain can change the rain load. Adjusting these model parameters does not change the frequency of the strain situations but their intensity.

Weather Data for Individual Investigations

Weather data that have been measured for an individual location and time period must be used if, for example,

the cause for a given damage case shall be investigated (‘forensic’ simulations),
measurements performed on or in a building are being accompanied by simulations for analysis or deeper insight
measurements and calculations shall be compared in order to validate a simulation model.

In such a case it may be sufficient to use data which a nearby weather station has collected for the relevant period of time. Often, however, you will have to perform your own on-site measurements. Data taken at the location are also needed if the radiation or rain load on the surface has to be determined more accurately than is possible with directional conversion models.

Sources for Weather Data

German Test Reference Years (2011)

The German Test Reference Years in the versions of 1986 and 2004 are included in the WUFI installation (with kind permission of the German Meteorological Service). WUFI can also read the new 2011 test reference years.

These new test reference years (TRYs) for the 15 German TRY regions have been derived from weather data series comprising the years 1988 through 2007 and thus include the effect of global warming up to that time. In addition to mean years, extreme years were created, with the extreme warm years containing an extremely warm summer period and the extreme cold years containing an extremely cold winter period.

Using a tool provided with these files, the user can modify the data to take additional scenarios into account:

An urban heat island effect can be imposed on the original data, representing cities of user-defined sizes.
If the heights of the considered location and of the reference station for this climate region are very different, appropriately corrected test reference years can be created, taking into account the influence of the height difference on air temperature and water vapor content.
In order to take future global warming into account, new files can be created which represent the temperature levels expected for the years 2021-2050 (again, in typical or extreme variants).

These test reference years contain no quantitative rain data, so they are not adequate for those hygrothermal simulations which need to take the influence of rain into account.

The 2011 test reference years can be downloaded free of charge from:

“Aktualisierte und erweiterte Testreferenzjahre (TRY) von Deutschland für mittlere und extreme Witterungsverhältnisse”
www.bbsr.bund.de

Austrian Test Reference Years:

In addition to the three weather files included with WUFI, files for many other Austrian towns are available.

The distributor for these data is

	QUADRUPLE-M Elsässer GmbH Technisches Büro für Bauphysik GF Dr. Manfred Elsässer Erzherzog-Eugen-Straße 14/1 A-6020 Innsbruck
	Tel. und Fax	+43 512 251401
	Fax	+43 512 378550
	Mobil	+43 664 4324814
	E-mail	office@quad-m.at or manfred.elsaesser@uibk.ac.at

Korean Weather Data

The Korean Test Reference Years (TRY) were created based on the international standard ISO 15927-4:2005 for assessing the annual energy use. The TRYs are based on measured data from the years 2005 till 2014 sourced by Korea Meteorological Administration. The preparation of the TRY was performed by Passive House Institute Korea.
This research was supported by a grant (15CTAP-C098308-01) from Infrastructure and transportation technology promotion research program funded by Ministry of Land, Infrastructure and Transport of Korean government.

The files can be downloaded here for free.

Japanese Weather Data

Weather data for 842 Japanese locations are available (Expanded AMeDAS Weather Data).

The distributor for these data is

(http://f-ei.jp/archives/wufi_pro)

These locations have already been predefined in WUFI’s climate selection map (see there for a list of locations). Just copy the respective files into WUFI’s Climate folder; on the next start WUFI will recognize the files. They can the easily be selected via the climate selection map.

METEONORM:

The Swiss METEOTEST company has developed METEONORM, a program for creating synthetic hourly weather data for any location world-wide, based on measured long-term mean values:

“Usually, measurement data can only be used in the vicinity of a weather station. Elsewhere, the data has to be interpolated between different stations. The sophisticated interpolation models inside meteonorm allow a reliable calculation of solar radiation, temperature and additional parameters at any site in the world.”

“From the monthly values (station data, interpolated data or imported data), meteonorm calculates hourly values of all parameters using a stochastic model. The resulting time series correspond to ‘typical years’ used for system design.”

The program mainly addresses “simulation software in solar energy applications and building design”. Meteonorm can directly create weather files in WUFI’s WAC format. Files for single locations can be ordered via the Meteonorm website without purchasing the program (specify file format “WUFI/WAC”).

The files created in this way are a convenient, homogeneous and inexpensive source of weather data for arbitrary locations. Experience so far indicates that, in general, they are well suitable for hygrothermal simulations. However, if the quality of the driving rain data is important for the simulation, the adequacy of these data for the case at hand should be checked beforehand, since the wind and rain models are only to a limited extent designed to realistically reproduce the intensity and directional distributions of driving rain events, and they do not allow for local topography.

EPW Files

The providers of the building energy simulation software EnergyPlus offer an extensive global collection of free weather data for use with their program.

WUFI can read these files. However, since they contain no rain data, they are not appropriate for investigations which involve rain as a crucial factor.

Other Data Sources

There exist a variety of other sources for weather data, including international data repositories, national weather services and private weather station owners. Sometimes they can provide data for the location and time in question, sometimes not. Sometimes the data are free of charge, sometimes not. The available data vary widely in terms of quality, scope and completeness. It is then the responsibility of the user to check the data quality, to fix errors, to fill data gaps, to convert the data to a suitable file format, etc. Hopefully, the remarks further up on this page are useful for this.

Data repositories which may or may not be useful include
NOAA National Climatic Data Center: www.ncdc.noaa.gov/
…
and others…

Last Update: December 2, 2024 at 17:57