Printer Friendly

Updating the ASHRAE climatic data for design and standards.


ASHRAE provides tables of climatic conditions for many locations in the United States, Canada and around the world in the Climatic Design Information chapter of its Handbook -Fundamentals (HOF; ASHRAE, 2005a, 2009a). These tables include values such as dry-bulb temperature, wet-bulb temperature, dew-point temperature, enthalpy, and wind speed at various frequencies of occurrence over a long-term period, corresponding mean coincident values of some other parameters, and averages of some extremes. ASHRAE also sells a companion product, the Weather Data Viewer (ASHRAE, 2005b, 2009b), which can be used to display actual design values, coincident/joint frequency tables, or summary statistics for dry-bulb, dew-point, and wet-bulb temperature, as well as enthalpy and wind speed, for all the locations given in the Climatic Information chapter. Finally the climatic conditions, along with some other data, are the primary component of ASHRAE Standard 169, Weather Data for ASHRAE Building Design Standards (ASHRAE, 2006).

ASHRAE Research Project 1453-RP, Updating the ASHRAE Climatic Data for Design and Standards, aimed at updating the climatic design condition tables for inclusion in the 2009 HOF and Standard 169. This update was necessary for the following reasons:

* The Climatic Information chapter in the 2009 HOF will regroup additional climate information, such as clear sky solar irradiance, which is currently scattered throughout various chapters of the Handbook. This information had to be calculated and added to the climatic tables;

* Project 1363-RP, Generation of Hourly Design Day Weather Data, resulted in the calculation of new climatic design conditions which needed to be added to the HOF climatic tables;

* The data from which the climatic tables have been derived is constantly being updated, with more recent years added to the period of record. This enables the calculation of design conditions from a longer period of record, thus enhancing the quality and reliability of the data; the expansion of the tables to a greater number of stations, thus enhancing geographical coverage; and the capture of trends such as the variation of design conditions with climate change.

* This project enabled the recalculation of climate statistics currently in Standard 169 (such as heating and cooling degree days), using the same period of record as for the climatic design conditions in the HOF, thus providing consistency between the various calculations.

A similar project (1273-RP) completed for the previous edition of the HOF was summarized in a paper by Thevenard and Humphries (2005), which contains technical details about the algorithms used to calculate the climatic design conditions. The 2009 update reused much of the same techniques, and their description is not repeated here. Instead, this paper focuses on the main differences between the 2005 and 2009 tables, which are covered in the following sections:

* The selection of new percentiles of monthly dry bulb and wet bulb temperatures to represent 'less extreme' conditions than those in the 2005 Handbook;

* The addition of new elements resulting from research project 1363-RP: dry bulb and wet bulb temperature ranges coincident with the 5% monthly dry bulb or wet bulb design temperatures, which can be used to derive daily temperature profiles;

* The calculation of heating and cooling degree-days base 50[degrees]F (10[degrees]C) and 65[degrees]F (18.3[degrees]C), in support of various standards including ASHRAE Std. 90.1-2004, and the development and testing of a new method to evaluate degree-days to any other base;

* The development of a new clear sky solar irradiance model to overcome the known limitations of the existing ASHRAE clear sky model, and to extend its applicability to the whole world.

The last section of the paper summarizes the development of the tables themselves and provides statistics about the sites included in the 2009 HOF.


The Climatic Design Information chapter of the 2005 HOF lists:

* the 0.4%, 1% and 2% annual design dry-bulb temperatures, which are exceeded 35.1, 87.7 and 175.3 hours in an average year;

* the 0.4%, 1% and 2% monthly design dry-bulb temperatures, which are exceeded 2.9, 7.3 and 14.6 hours in an average month.

As the revision of the chapter was underway, the question was raised as to whether these monthly design temperatures were 'too extreme' and would lead to significant oversizing of equipment.

There is no easy way to answer that question, but one simple guideline could be that monthly design temperatures for the hottest month should be roughly equivalent to the yearly design temperatures. A comparison of dry-bulb temperatures corresponding to various frequencies of occurrence in the hottest month, calculated for all 4,422 locations contained in the 2005 HOF, to the average yearly design conditions, lead to establishing a rough correspondence between yearly and monthly design conditions, summarized in Table 1. A simple rule of thumb is that the monthly percentile to use is five times the yearly percentile. This is illustrated in Figure 1 which compares the 5% monthly design dry bulb temperature for the hottest month to the 1% yearly design dry bulb temperature for all stations in the 2005 HOF. The information is presented both as a scatter plot and as a histogram (similar graphs can be plotted for other frequencies of occurrence). The correlation between monthly and yearly values is of course not perfect: it depends on how temperatures are distributed throughout the year. Nevertheless, it is apparent that over 90% of monthly design conditions are within 1 [degrees]C of their yearly counterparts. This was a strong argument to use the 2%, 5% and 10% monthly design conditions in the 2009 HOF, instead of the 0.4%, 1% and 2% monthly design conditions that were used in the 2005 edition.
Table 1. Approximate Correspondence between Yearly Design Dry-Bulb
Temperatures and Monthly Design Dry-bulb Temperatures for the Hottest

Yearly Design Dry-Bulb Temperature  Monthly Design Dry-Bulb Temperature
                                           for the Hottest Month

               0.4%                                  2%
                1%                                   5%
                2%                                  10%

After some discussion, it was decided to nevertheless retain the 0.4% monthly conditions in addition to the 2%, 5% and 10% values; the rationale is that the 2% monthly value would not be extreme enough for some applications. Although the choice would typically be up to building owners and designers, the various monthly percentiles could be suggested depending on the type of design considered:

* 0.4%: critical process, hospital operating rooms, etc.

* 5%: corporate headquarter, upscale hotel, patient room, etc.

* 10%: competitive building, retail space, speculative office building.

The 2009 HOF therefore lists the monthly 0.4%, 2%, 5% and 10% design dry and wet bulb temperatures, as well as their coincident wet bulb and dry bulb temperatures.



Twenty-four-hour weather data sequences are required as input for many HVAC design procedures including residential and non-residential cooling load calculations (both Heat Balance and Radiant Time Series), fenestration heat gain calculation, and equipment performance analyses. Calculations are often done for multiple conditions (all months of the year, for example). It is impractical to publish multiple 24-hour sequences for the thousands of sites listed in the Fundamentals climatic data tables. Instead, sequences are derived from a few tabulated design conditions using models. A 24-hour dry bulb temperature profile is generated by application of a generic profile to the design temperature and the daily range. ASHRAE has long published a single generic dry-bulb profile for cooling, found most recently in Chapter 28, Table 2 of 2005 Fundamentals. In that same chapter, TC 4.2 has added a method for generating coincident hourly wet-bulb temperature based on an assumption of constant absolute humidity.
Table 2. Fraction of Daily Temperature Range

Time, h  f

   1       0.88
   2       0.92
   3       0.95
   4       0.98
   5       1.00
   6       0.98
   7       0.91
   8       0.74
   9       0.55
  10       0.38
  11       0.23
  12       0.13
  13       0.05
  14       0.00
  15       0.00
  16       0.06
  17       0.14
  18       0.24
  19       0.39
  20       0.50
  21       0.59
  22       0.68
  23       0.75
  24       0.82

The models for generation of 24 hour sequences in the 2005 Handbook had several shortcomings. In particular:

* The source and geographic applicability of the generic dry bulb profile are not known (no reference has been located for the method).

* The profile is generally applied to the average daily range, as opposed to the daily range coincident with design conditions, apparently due to prior limitations in available data.

* The profiles represent only the high dry-bulb ("hot day") condition; some design problems require analysis under conditions of high wet-bulb (enthalpy) or high dew-point (absolute humidity).

* Finally, the wet bulb method added in the 2005 Fundamentals is known to be approximate at best and was included only because some source for the data was required for latent cooling load calculations.

New Temperature Profile

In response to these shortcomings, ASHRAE sponsored Research Project 1363-RP, Generation of hourly design-day weather data (Gard Analytics, 2009). This project used weather data from 21 cities with a variety of climatic conditions to develop a daily profile which describes the variation of dry bulb and wet bulb temperature on design days, identified as days with high dry bulb or high wet bulb temperatures. The profile defines the hourly temperature as a fraction of the difference between the daily high and low, as described by the design temperature and the design daily range. A single profile was found to be applicable to all climates and times of the year for both dry bulb and wet bulb profiles. Dew point temperature variations were found to be highly variable, such that no single profile can provide reliable predictions.

Research project 1453-RP built upon the findings of 1363-RP and reformatted them in such a way that the desired daily dry bulb and wet bulb temperature ranges could be calculated from historical data at the same time as other design conditions. The daily ranges are 'coincident' with high dry bulb or high wet bulb conditions in the sense that they are averaged over days for which the maximum dry bulb or wet bulb temperatures exceeds the 5% monthly design values. There are therefore 4 design day profiles: dry bulb and wet bulb profiles on the dry bulb design day, and dry bulb and wet bulb profiles on the wet bulb design day. In practical terms, the calculation is performed via a {mean daily temperature range, max daily temperature} joint frequency matrix, where 'temperature' can be either the dry bulb or the wet bulb temperature. This enables to prepare the calculation of the mean daily temperature range without knowing the design temperature for which it will be calculated; the method is simply a variation of the one used for other coincident design conditions (see Thevenard and Humphries, 2005). The processing software simply calculates the mean daily temperature ranges for all bins of max daily temperature, then proceeds by linear interpolation to calculate the temperature range corresponding to the actual design temperature.

Based on the analysis of 21 cities mentioned above, the 1363-RP profile was found to represent adequately the variation of dry bulb temperature and wet bulb temperature during an average design day. This profile is obtained by subtracting the fraction f of the dry- or wet-bulb daily range from the dry-or wet-bulb design temperature. Fraction f is listed in Table 2; in that table, the time h should be understood as the apparent solar time or, if it is not practical to calculate it, the standard time of the closest meridian. Note that the temperature profile presented here is the one from 1363-RP, but shifted by one hour; that is, the maximum occurs between 14:00 and 15:00, not between 13:00 and 14:00 as found by 1363-RP. This shift was required to match the empirically derived daily profiles data. It is assumed that a processing error in 1363-RP led to this problem.

Figure 2 compares the average daily temperature profiles from all 21 test sites, to the one from Table 2. Daily profiles were obtained by averaging dry bulb temperature, on an hourly basis, for all days when the dry bulb temperature exceeded the 5% design condition. The daily profiles are normalized between their minimum and their maximum prior to averaging. Each grey curve represents the average of all 12 months for one station (stations were not color-differentiated). It is interesting to note that the time at which the maximum temperature occurs varies somewhat from station to station. The profile defined by Table 2 is therefore rather approximate; however it provides a convenient way to describe the daily variation of temperature from a condensed amount of information.


Similar figures (not shown here) can be plotted for the wet bulb temperature profile on days with high dry bulb temperature, and for dry bulb and wet bulb temperature profiles on days with high wet bulb temperature. The adequacy of the data from Table 2 to represent all these profiles varies from profile to profile and from station to station; however they are 'close enough' for engineering purposes and, for the sake of simplicity, a single profile was kept to represent all.

Figure 2 also compares the new dry bulb design-day profile to the one that was present in the 2005 ASHRAE Handbook - Fundamentals. The profiles are not significantly different. The novelty of the new profile, though, resides in discovering that it is also applicable to wet bulb temperatures. In the 2005 HOF, the wet bulb temperature profile was derived from the dry bulb temperature profile using psychrometric formulae and a constant dew point temperature; but the constant dew point temperature assumption was never properly justified. Work done by 1363-RP, and confirmed by 1453-RP, shows that the dew point temperature profile changes very significantly from station to station. The new method specifies deriving the dew point temperature profile from the dry bulb and wet bulb temperature profiles, defined above, using psychrometric formulae. The shape of the profile does not follow a general rule: it depends on the relative values of the design conditions, coincident conditions, and coincident daily ranges (Thevenard, 2009a).



Heating, cooling, and other degree-days, calculated to various bases, are used extensively in a variety of fields of engineering and science. For example heating degree-days base 18.3[degrees]C (65[degrees]F) are often used to estimate the energy required to heat buildings, via a proportionality factor specific to each building (ASHRAE, 2005a, Chapter 32: Energy Estimating and Modeling Methods). Similarly, cooling degree-days base 10[degrees]C (50[degrees]F) are often used to correlate cooling energy requirements to the local climate. Degree-days are also used in a number of standards, such as ASHRAE Standard 90.1 (ASHRAE, 2004), which relies mostly on heating degree-days base 18.3[degrees]C (65[degrees]F) and cooling degree-days base 10[degrees]C (50[degrees]F) to classify locations into climate zones. For these reasons, it was decided to include heating and cooling degree-days in the tables of the 2009 Handbook - Fundamentals. Because of space constraints only two base temperatures could be included; the most commonly used, 18.3[degrees]C (65[degrees]F) and 10[degrees]C (50[degrees]F), where chosen. Since degree-days to other bases may be required by users, it was decided to also provide in the Handbook a method suitable for the evaluation of degree-days to other bases. This section succinctly describes this method.

Degree-Days Calculated from Observations

Heating and cooling degree-days are defined as the sum of the differences between daily average temperatures and the base temperature. For example the number of heating degree-days in a month, [HDD.sub.b], is calculated as:

[HDD.sub.b] = [N.summation over (i=1)][([T.sub.b] - [[bar.T].sub.i]).sup.+] (1)

where N is the number of days in the month, [T.sub.b] is the base temperature to which the degree-days are calculated, and [[bar.T].sub.i] is the mean daily temperature. The "+" superscript indicates that only positive values of the bracketed quantity are taken into account in the sum. Similarly, monthly cooling degree-days [CDD.sub.b] are calculated as:

[CDD.sub.b] = [N.summation over (i=1)][([[bar.T].sub.i] - [T.sub.b]).sup.+] (2)

For historical reasons, the mean daily temperature [[bar.T].sub.i] is by definition calculated by adding the maximum and minimum temperatures for the day, then dividing by 2; this is also the method used in 1453-RP. Only days with a complete 24 hour record are used so as not to skew the calculation.

Yearly degree-days are simply the sum of monthly degree-days over the 12 months of the year.

Estimated Degree-Days

There may be circumstances where it is necessary to estimate degree-days to bases other than 18.3[degrees]C (65[degrees]F) and 10 [degrees]C(50[degrees]F). A solution would have been to tabulate degree-days to multiple bases; however this option was rejected as it would have occupied significant space in the tables, and drowned the users under a deluge of seldom used data. Fortunately, approximate methods can be used to estimate degree-days using a reduced set of parameters. Two well known methods are the Thom method (1954a; 1954b; 1966) and the Erbs, Klein and Beckman (1983) method, the latter being also included in Chapter 32 of the 2005 ASHRAE Handbook - Fundamentals. Both derive monthly degree-days from the average and the standard deviation of the monthly mean temperature. The Erbs et al. method goes one step further by relating the standard deviation of monthly mean temperature to the deviation of the monthly average ambient temperature from the annual average ambient temperature.

More recently Schoenau and Kehrig (1990) derived a simple method for calculating degree-days to any base. The method was extensively tested during 1453-RP, and those tests are documented in the project's final report (Thevenard, 2009a). This paper will therefore only present a brief summary of the method and its performance.

Schoenau and Kehrig assume that the daily mean temperatures are normally distributed around the monthly mean. Heating degree-days [HDD.sub.b] to base [T.sub.b] can be expressed as:

[HDD.sub.b] = N * [s.sub.d][[Z.sub.b] * F([Z.sub.b]) + f([Z.sub.b])] (3)

where [Z.sub.b] is the difference between base temperature [T.sub.b] and monthly average temperature [[bat.T].sub.m], normalized by the standard deviation of the daily average temperature [s.sub.d]:

[Z.sub.b] = [[[T.sub.b] - [[bar.T].sub.m]]/[s.sub.d]] (4)

Function f is the normal (Gaussian) probability density function with mean 0 and standard deviation 1, and function F is the equivalent cumulative normal probability function:

f(Z) = [1/2[pi]]exp([-[Z.sup.2]/2]) (5)

F(Z) = [Z.[integral] -[infinity]]f(z) * dz (6)

Cooling degree-days [CDD.sub.b] to base [T.sub.b] are calculated by the same formula:

[CDD.sub.b] = N * [s.sub.d][[Z.sub.b] * F([Z.sub.b]) + f([Z.sub.b])] (7)

except that [Z.sub.b] is now expressed as:

[Z.sub.b] = [[[[bar.T].sub.m] - [T.sub.b]]/[s.sub.d]] (8)

The method derived by Schoenau and Kehrig is elegant, compact, and requires knowledge of only two variables, [[bar.T].sub.m] and [s.sub.d], both of which can easily be calculated from hourly or daily temperature records. The performance of the method was assessed using a set of over 150,000 monthly degree-day values calculated for locations spread out over the whole world (Thevenard, 2009a). Model performance varies according to the value of [Z.sub.b] and was estimated in terms of mean bias error MBE and root mean square error RMSE:

MBE = [1/N][SIGMA]([[DD.sub.est] - [[DD.sub.true]) (9)

RMSE = [square root of [1/N][SIGMA][([DD.sub.est] - [DD.sub.true]).sup.2]] (10)

where [DD.sub.est] represents degree-days estimated with (3) or (7), [DD.sub.true] represents degree-days calculated from observations, and N is the number of points considered in the test. Values of [Z.sub.b] were binned and model performance was evaluated inside each bin; the resulting histogram is shown in Figure 3. Mean bias error is small at less than 1[degrees]C-day (1.8[degrees]F-day) for all bins; and root mean square error is also quite acceptable at less than 3[degrees]C-day (5.4[degrees]F-day) for all bins. Figure 3 also presents a similar figure derived with the Erbs et al. formula and shows clearly the superiority of the Schoenau and Kehrig method.

In view of this, the Schoenau and Kehrig method was added to the 2009 ASHRAE Handbook - Fundamentals; the parameters it requires (monthly average temperature and standard deviation of daily average temperature for all twelve months of the year) were added to the Tables of climatic design information for the 5,564 stations worldwide included in the Handbook.


The proper sizing of cooling and air-conditioning equipment requires knowledge of the cooling load to which internal spaces are subjected. A major contributor to cooling loads is solar radiation, which adds heat to the inside space either directly (through fenestration) o r indirectly (through the building envelope). Solar gains used for the sizing of cooling equipment are generally evaluated for clear sky conditions, since those are the conditions for which solar gains tend to be maximum. Starting with the 1967 edition of its Handbook - Fundamentals, ASHRAE has provided a clear sky solar radiation model which can be used for that purpose. The model was initially developed by Threlkeld (1962), with subsequent improvements by Stephenson (1965) and Machler and Iqbal (1985). It is described in Chapter 33, Solar Energy Use, of the 2007 ASHRAE Handbook - Applications (ASHRAE, 2007), and it is referenced in particular in Chapter 30, Nonresidential Cooling and Heating Load Calculations, and in Chapter 31, Fenestration, of the 2005 ASHRAE Handbook - Fundamentals. This latter chapter also contains a summary of the model, although with different coefficients than those found in chapter 33 of the ASHRAE Handbook - Applications.

The ASHRAE clear sky model calculates beam and diffuse irradiance under clear sky conditions. The model uses only one variable, the sun's zenith angle, Z, and three empirical coefficients (A, B, C), according to:

[] = A exp ([-B/cosZ]) (11)

[E.sub.d] = [] (12)

where [] is the beam normal clear sky irradiance, [E.sub.d] is the diffuse irradiance on a horizontal surface, and Z is the solar zenith angle. An optional clearness number, CN, was introduced to act as a multiplicative correction factor to [], to improve the predictions of the model. Twelve monthly values of coefficients A, B and C are tabulated in the Handbook; the clearness number CN is provided as a small map covering the continental United States, with an indication of the seasonal applicability (summer/winter) of the isolines.

The ASHRAE clear sky model of 1967 suffers from a number of well-known limitations:

The clearness number CN is not universally available; CN varies between 0.87 and 1.20 according to location and season in the United States. In other countries, users often take CN = 1 by lack of better data;

* Many studies have shown that the model lacks universal applicability; values of A, B and C need to be altered according to location;

* The application of the model to locations in the Southern hemisphere is awkward;

* There are inconsistencies in the description of the model in different chapters of the Handbook (Chapter 33 in the 2007 Handbook - HVAC Applications uses the original coefficients by Threlkeld (1962); Chapter 31 in the 2005 Handbook - Fundamentals uses revised coefficients by Machler and Iqbal (1985));

* Clear sky diffuse irradiance is proportional to beam irradiance, a fact that is runs contrary to intuition (hazier skies should lead to less direct but more diffuse);

* The model was derived from a very limited number of measurements and its applicability outside the USA has never been demonstrated.

To overcome these limitations, ASHRAE research projects 1363-RP initiated, and 1453-RP continued, a significant update of the model. The objectives of the update were:

* Accuracy, through the use of a more detailed and state-of-the art model;

* Universality, in order to extend the applicability of model to the whole world, and provide clear sky data for all 12 months of the year in support of monthly load calculation algorithms;

* Ease of use, with the aim of using condensed formulae suitable for a hand-held calculator or spreadsheet, and using a limited number of site dependent parameters that can be easily tabulated in Handbook.

Finally, the new clear sky model was to be moved to the climatic information chapter, so as to avoid duplicating the model description in different chapters, and so that the model could be found along the tabulated parameters that it requires.

The model was developed through a two-step procedure. First, a high-performance clear-sky radiation model called REST2 was run. This model requires detailed and sophisticated inputs about atmospheric aerosols, water vapor, ozone, etc. which are beyond the reach of most engineers and scientists. Then, a simplified model, suitable for inclusion in the Handbook, was derived by fitting a 2-parameter model to the full REST2 model results. An in-depth description of these two steps is beyond the scope of this paper and can be found in a separate publication (Gueymard and Thevenard, 2009). Only the outline of the procedure is

outlined here.

It should also be noted that what constitutes a 'clear sky' is somewhat difficult to define but generally refers to the absence of clouds. It can range from pristine atmospheric conditions with high beam and low diffuse irradiances, to hazy and thick atmospheres with lower beam and higher diffuse irradiances. It should also be noted that for some locations the concept of 'clear sky' is actually more abstraction than reality; for example Yin and Wah (2004) report on the absence of cloudless days in Singapore. For such locations, what can be referred to as 'clear sky' simply represents a theoretical 'upper limit' in terms of solar radiation conditions to be used in load calculations. 'Average' clear sky conditions (in terms of turbidity, etc.) were modeled for the 2009 HOF.

REST2 Model

REST2 is a high-performance, two-band clear-sky radiation model, which allows the calculation of the direct and diffuse components of clear-sky broadband horizontal irradiance, illuminance, and photosynthetically-active radiation. The first spectral band, 290-700 nm, is dominated by scattering processes due to molecules (Rayleigh effect) or aerosols (Mie effect). Conversely, absorption processes (mainly by water vapor and carbon dioxide) are predominant in the second band (700-4000 nm). This two-band arrangement provides a simple and accurate solution to the problem of generalizing the Bouguer-Lambert-Beer extinction law (normally restricted to monochromatic radiation) to the case of broadband radiation. In each band, equations for the direct and diffuse transmittances are parameterized from a large dataset of simulated band irradiances obtained with the SMARTS spectral code (Gueymard, 2001, 2005, 2008a). Details on the derivation and validation of REST2 can be found in Gueymard (2008b).

The REST2 model requires detailed knowledge of a number of inputs about various atmospheric constituents, such as aerosols, water vapor, ozone, etc. In order to extend the applicability of the model to the whole world, a number of datasets, mainly derived from space observations, were used to obtain these inputs. Water vapor data were derived from the NVAP satellite/radiosonde assimilated dataset for 1988-1999 (Randel et al., 1996), corrected for elevation (Thevenard 2009a). Total ozone amount was derived from observations of the TOMS instrument aboard the Nimbus 7 satellite ( for 1988-1992. A fixed NO2 amount of 0.4 matm-cm was used throughout the world. Far-field ground albedo was obtained from the Surface and Atmospheric Radiation Budget (SARB) based on CERES data (Charlock et al., 2004) for 2000-2005.

Aerosol turbidity data received special attention since they are the primary inputs that condition the accuracy of the direct and diffuse irradiance predictions under clear skies. Aerosols are mostly characterized by the aerosol optical depth (AOD), which is related to Angstrom's turbidity coefficient [beta] and the wavelength exponent [alpha] through:

[AOD.sub.[lambda]] = B * [[lambda].sup.-[alpha]] (13)

where [AOD.sub.[lambda]] represent the aerosol optical depth at wavelength [lambda]. Three categories of aerosol data were used: (a) grid-based data using spaceborne measurements [MODIS] or (b) modeled data [MATCH], and (c) point-source data for ground truthing. MODIS (Moderate Resolution Imaging Spectroradi-ometer) data are space-borne sets recorded by instruments aboard NASA's Terra and Aqua satellites ( They include at least 5 years of data, but present the drawback that the instruments experience difficulties over areas with high reflectivity, such as snow-covered surfaces or deserts. MATCH (Model of Atmospheric Transport and Chemistry) data are derived from a chemical transport model (Rasch et al., 1997; Clarke et al., 2001); six years (2000-2005) of simulated monthly-average aerosol optical depth at 550 nm were prepared with MATCH by the Science Directorate/Climate Science Branch at NASA Langley Research Center, which also supplied aerosol single-scattering albedo estimates. Aerosol optical depth data from MATCH and MODIS were compared with a large number of ground-based sites, mostly from the AERONET network ( 196 sites used with at least 3 years of data for a given month were used. This comparison showed that AOD was better estimated by combining MODIS and MATCH data:

AOD = [a.sub.0] + [a.sub.1] * [AOD.sub.MATCH] + [a.sub.2] * [AOD.sub.MODIS]

where the coefficients [a.sub.0], [a.sub.1] and [a.sub.2] were determined by a linear fit using ground data.

Figure 4 shows, as an example, the combined AOD for the month of July on a planetary scale. The high-AOD areas, such as parts of Africa (because of biomass burning and dust storms) and Asia (because of anthropogenic aerosol loading), are clearly visible. Other details about REST2 and the preparation of its inputs can be found in Gueymard and Thevenard (2009) and in Thevenard (2009a).


Simplified Model

As noted earlier, REST2 requires many inputs and is too complicated to be incorporated as is into the Handbook. A condensed model was sought, which would require the tabulation of no more than two parameters per month, and enable the calculation of clear sky irradiance for any time of any day with a set of simple equations. To achieve this, the REST2 model was run for a very wide range of conditions:

Month: 1 to 12 (model was run for the 21st day of each month)

Local apparent time: 0 to 12

Latitude: 0, 15, 30, 45, 60 and 75[degrees]N

Atmospheric pressure: 1013.25 and 750 mb

Precipitable water: 0.5, 1, 2, and 6 cm

Angstrom's turbidity coefficient [beta]: 0.01, 0.02, 0.03, 0.05, 0.07, 0.1, 0.15, 0.2, 0.3, 0.5

Results from the REST2 model were fitted to a simple 2-parameter model. The functional form of the fit was suggested by the work of Mueller et al. (2004); it calculates the beam normal and diffuse horizontal components of clear sky solar irradiance as:

[] = [E.sub.o] * exp[-[[tau].sub.b] * [m.sup.ab]] (14)

[E.sub.d] = [E.sub.o] * exp[-[[tau].sub.d] * []] (15)


[] = beam normal irradiance (measured perpendicularly to the rays of the sun),

[E.sub.d] = diffuse horizontal irradiance (measured on a horizontal surface),

[E.sub.o] = extraterrestrial normal irradiance,

m = air mass,

[[tau].sub.b] and [[tau].sub.d] = beam and diffuse optical depths ([[tau].sub.b] and [[tau].sub.d] are more correctly termed "pseudo" optical depths because optical depth is usually employed when the air mass exponent is unity; "optical depth" will be used here for convenience), and

ab and ad = beam and diffuse air mass exponents.

Values of [[tau].sub.b] and [[tau].sub.d] are location-specific, and vary during the year. They embody the dependence of clear sky solar radiation upon local conditions, such as elevation, precipitable water content, and aerosols. Their average values were determined by running the REST2 model for the specific locations tabulated in the tables of climatic design conditions. The air mass exponents ab and ad were correlated to [[tau].sub.b] and [[tau].sub.d] through the following empirical relationships:

ab = 1.219 - 0.043 * [[tau].sub.b] - 0.151 * [[tau].sub.d] - 0.204 * [[tau].sub.b] * [[tau].sub.d] (16)

ad = 0.202 + 0.852 * [[tau].sub.b] - 0.007 * [[tau].sub.d] - 0.357 * [[tau].sub.b] * [[tau].sub.d] (17)

Figure 5 illustrates the goodness of the fitting procedure (to reduce clutter, only those cases corresponding to a precipitable water of 1 cm, and an atmospheric pressure of 1013.25 mb, are shown; the various curves correspond to the values of [beta] listed above and illustrate the predominant effect of turbidity on [] and [E.sub.d]). The simplified model is generally quite accurate for zenith angles up to 75[degrees]; its performance degrades when the sun is near the horizon. This is generally not significant for cooling load calculations, however future improvements of the model should certainly focus on that area of the curves, to address cooling loads occurring in the middle of the afternoon during the shoulder season (e.g. in September for a site located at 45[degrees]N, the zenith angle becomes greater than 75[degrees] between 4 and 5 p.m.), and more generally for the calculation of solar gain for any vertical window having near normal sun angles (NW in summer, W in shoulder season, or SW in winter (for a mid latitude northern hemisphere site).


Comparison to the Previous ASHRAE Clear Sky Model

Figure 6 compares the beam normal and diffuse horizontal irradiances calculated by the new and the old (1) clear sky models, for June 21 for four North-American cities: El Paso, TX (latitude: 31.81[degrees]N), Atlanta, GA (33.64[degrees]N), Denver, CO (39.77[degrees]N), and Vancouver, BC (49.25[degrees]N). A general conclusion is that the new model tends to predict less direct but more diffuse irradiance than the old model. Preliminary comparison with measured clear sky solar radiation data from two sources also found that the new model tends to be somewhat "heavy on diffuse"; it is possible that this may be related to the 'average' values of aerosol optical depths that were used for the REST2 runs, which were representative of 'average' rather than 'very clear' skies. Future work will address some of these issues through the use of better space-borne aerosol data sets, and better fits for the condensed model. Although the datasets used as input to REST2 in this study represent the current state of the art, the field of remote sensing is in constant evolution and it is expected that, within the next few years, better datasets will become available. Improvements are expected in several areas: longer length of the period of observation, improved calibration of sensors, and enhanced algorithms to extract information relevant to REST2. In particular, better aerosol data will certainly become available, especially over areas with high ground albedo (e.g., snow-covered areas and deserts).


Data Sources

To calculate the tables of climatic design for the 2009 Handbook - Fundamentals, two sources of data were used. For US and international locations, the source of data was the Integrated Surface Dataset (ISD) from the National Climatic Data Center (NCDC; Lott et al., 2001; NCDC, 2003; Lott, 2004). The Integrated Surface Database (ISD) consists of global hourly and synoptic observations compiled from numerous sources, into a single common ASCII format and common data model. The database comprises over 20,000 stations worldwide, with records beginning in the 1948-1973 time-frame for most stations, although some have data as far back as 1901. Access to the data was graciously provided by NCDC. For Canadian locations, Environment Canada graciously provided access to their National Climate Data and Information Archive (

Choice of Period of Record

A study of the influence of long-term trends and period of record selection on the calculation of climatic design conditions and degree-days was conducted (Thevenard, 2009b). Climatic design conditions and cooling and heating degree days were calculated using data from different decades for 1274 stations worldwide. This revealed long-term trends. Over the last three decades, climatic design conditions have increased at an average rate of 0.76[degrees]C/decade for the 99.6% heating dry bulb temperature, 0.38[degrees]C/decade for the 0.4% cooling design temperature, and 0.28[degrees]C/decade for the 0.4% dehumidification dew point temperature. During the same period, the yearly average temperature increase on average by 0.41[degrees]C/decade. Annual heating degree-days have decreased on average by 118[degrees]C-day/decade while annual cooling degree days have increased by 68[degrees]C-day/decade.

These changes are indicative of a warming of the climate experienced by the monitoring stations. However, the magnitude of that warming indicates that it is probably less related to global warming than to the urban heat island effect; it is likely an indication of the built environment encroaching on locations where meteorological stations are situated, particularly airports.

The reference cited above also studied the appropriate period of record to use for the calculation of climatic design conditions and degree-days. The paper showed that the use of a shorter, more recent period of record enables to better capture warming trends, but results in an added uncertainty that is greater than the observed climate trend itself.

The Project Monitoring Subcommittee adopted the most recent 25 years (1982-2006 at the time of processing) as a balance between accounting for long term trends and reducing the stochastic uncertainty associated with shorter time periods. While this does not automatically guarantee that the conditions will be representative of the climate for the next one or two decades, it provides a snapshot of current changes in the global climate and local conditions.

Tables of Climatic Design Conditions

Processing of the data led to the inclusion in 5,564 stations in the tables of climatic design conditions. The data is included in electronic format on the CD-ROM that accompanies the 2009 Handbook - Fundamentals; a mapping utility enables to locate the stations on Google Maps. A subset of the elements and of the stations is also available in the printed edition of the Handbook. The printed elements include the most frequently used yearly design conditions, and the printed stations include, for the USA and Canada, any station that is within 30 km of an urban centre with a population of at least 30,000, and for the rest of the world, any station that is within 30 km of an urban centre with a population of at least 300,000. The geographical repartition of the stations (both in electronic and print format) is summarized in Table 3.
Table 3. Geographical Repartition of Stations in the 2009

Region             Number of Stations on CD-ROM  Number of Stations
                                                      in Print

USA                            1085                      545
Canada                          480                      102
Rest of the World              3999                      807
Total                          5564                     1454

Weather Data Viewer

The Weather Data Viewer (WDView) was also updated, with a new version (4.0) now available (ASHRAE, 2009b). As before the Weather Data Viewer gives access to climatic design conditions, frequency distributions, joint frequency tables, and summary statistics for dry-bulb, wet-bulb, and dew-point temperatures plus wind speed and direction for all 5,564 locations listed in Handbook. The viewer can display the tables in numeric form or plot the frequency distributions, the cumulative distribution functions, and mean coincident functions. It also provides additional information such as frequency matrices of dry-bulb temperature and time of day (also known as temperature bin data), the time zone and daylight-saving time of the stations, and the months and years used in calculating the design conditions.


Research Project 1453-RP updated the tables of climatic design conditions in the 2009 ASHRAE Handbook - Fundamentals. The update included both revisions to elements already listed in the 2005 Handbook, and the addition of new elements. New percentiles (2, 5 and 10%) of monthly dry bulb and wet bulb temperatures were selected to represent 'less extreme' conditions than those in the 2005 Handbook. Results from research project 1363-RP, Generation of Hourly Design Day Weather Data, were validated and led to the calculation of new dry bulb and wet bulb temperature ranges coincident with the 5% monthly dry bulb or wet bulb design temperatures; these values can be used to derive daily temperature profiles. Finally, Heating and cooling degree-days base 50[degrees]F (10[degrees]C) and 65[degrees]F (18.3[degrees]C) were also added to the tables, as well as parameters suitable to estimate degree-days to any base.

The development of a new clear sky solar irradiance model represented a major undertaking of this project. The purpose was to overcome the known limitations of the existing ASHRAE clear sky model, and to extend its applicability to the whole world. A two-step approach was used: a sophisticated broadband clear-sky radiation model was run, and then fitted to a simple 2-parameter model. The fits enable a concise formulation requiring the tabulation, on a monthly basis, of only two parameters per station, referred to as the clear sky beam and diffuse optical depths. These two parameters are tabulated for all stations in the tables of climatic design conditions.

Processing led to the calculation of climatic design conditions for a total of 5,564 locations: 1,085 US stations, 480 Canadian stations, and 3,999 international stations. The data is available as electronic files located on the Handbook CD. A subset of the data is also available in printed format. A companion product, the Weather Data Viewer, version 4.0, gives access to climatic design conditions, frequency distributions, joint frequency tables, and summary statistics for all 5,564 locations listed in the Handbook.

The work summarized in this paper resulted in a significant update of the climatic data listed in the Handbook. Future work may include additional validation of and possible enhancements to the clear sky model; the addition of new elements such as ground temperature (for use in basement heat loss calculation or ground source heat pump systems), precipitation or snow cover; and the redevelopment of so-called transposition models used to calculate clear sky solar radiation incident on surfaces of arbitrary orientation.


This work was supported by the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) through Research Project 1453-RP, Updating the ASHRAE Climatic Data for Design and Standards. The author is indebted to members of the Project Monitoring Subcommittee for useful discussions and encouragements in pursuing this work.


ASHRAE, 2004. ANSI/ASHRAE/IESNA Standard 90.1-2004, Energy Standard for Buildings Except Low-Rise Residential Buildings. American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Atlanta, GA, USA.

ASHRAE, 2005a. Handbook - Fundamentals. American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Atlanta, GA, USA.

ASHRAE, 2005b. Weather Data Viewer, version 3.0. American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Atlanta, GA, USA.

ASHRAE, 2006. ASHRAE Standard 169 - Weather data for ASHRAE Building Design Standards. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. Atlanta, GA, USA.

ASHRAE, 2007. Handbook - Applications. American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Atlanta, GA, USA.

ASHRAE, 2009a. Handbook - Fundamentals. American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Atlanta, GA, USA.

ASHRAE, 2009b. Weather Data Viewer, version 4.0. American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Atlanta, GA, USA.

Charlock T.P. et al., 2004. Global retrievals of the surface and atmosphere radiation budget and direct aerosol forcing. Proc. 13th Conf. on Satellite Meteorology and Oceanography, American Meteorological Society, Norfolk, VA .

Clarke A.D. et al., 2001. Dust and pollution transport on global scales: Aerosol measurements and model predictions. J. Geophys. Res. 106D, 32555-32569.

Erbs D.G., Klein S.A., and Beckman W.A., 1983. Estimation of degree-days and ambient temperature bin data from monthly temperature averages. ASHRAE J., 25 (6) 60-65.

Gard Analytics, 2009. Generation of Hourly Design-Day Weather Data - Final Report, ASHRAE 1363-RP. American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Atlanta, GA, USA.

Gueymard C.A., 2001. Parameterized Transmittance Model for Direct Beam and Circumsolar Spectral Irradiance. Solar Energy 71, 325-346.

Gueymard C.A., 2005. SMARTS code, version 2.9.5 User's Manual. Solar Consulting Services. Online PDF document, available from

Gueymard C.A., 2008a. Prediction and validation of cloudless shortwave solar spectra incident on horizontal, tilted, or tracking surfaces. Solar Energy 82, 260-271.

Gueymard C.A., 2008b. REST2: High performance solar radiation model for cloudless-sky irradiance, illuminance and photosynthetically active radiation--Validation with a benchmark dataset. Solar Energy 82, 272-285.

Gueymard C.A. and Thevenard D., 2009. Monthly average clear-sky broadband irradiance database for worldwide solar heat gain and building cooling load calculations. Solar Energy 83(11) 1998-2018.

Lott, J.N., 2004. The quality control of the integrated surface hourly database. 84th American Meteorological Society Annual Meeting. Seattle, WA. (Available at

Lott, J.N., Baldwin R., and Jones P., 2001. NCDC technical report 2001-01, The FCC Integrated Surface Hourly Database, A New Resource of Global Climate Data. National Climatic Data Center, Asheville, NC. (Available at

Machler M.A. and Iqbal M., 1985. A modification of the ASHRAE clear sky irradiation model. Trans. ASHRAE 91A, 106-115.

Mueller R.W. et al., 2004. Rethinking satellite-based solar irradiance modeling: the SOLIS clear-sky module. Rem. Sens. Environ. 91, 160-174.

NCDC. 2003. Data Documentation for Data Set 3505 (DSI-3505) Integrated Surface Hourly (ISH) Data. National Climatic Data Center, Asheville, NC.

Randel D.L. et al., 1996. A new global water vapor dataset. Bull. Amer. Met. Soc. 77, 1233-1246.

Rasch P.J. et al., 1997. Representations of transport, convection, and the hydrologic cycle in chemical transport models: Implications for the modeling of short-lived and soluble species. J. Geophys. Res. 102D, 28127-28138.

Schoenau, G.J. and Kehrig R.A., 1990. A method for calculating degree-days to any base temperature. Energy and Buildings 14: 299-302.

Stephenson D.G., 1965. Equations for solar heat gain through windows. Solar Energy 9, 81-86.

Thevenard D. and Humphries R., 2005. The Calculation of Climatic Design Conditions in the 2005 ASHRAE Handbook--Fundamentals. ASHRAE Transactions 111(1):457-466.

Thevenard D., 2009a. ASHRAE Research Project 1453-RP - Updating the ASHRAE Climatic Data for Design and Standards, Final Report. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta, GA.

Thevenard D., 2009b. Influence of Long-term Trends and Period of Record Selection on the Calculation of Climatic Design Conditions and Degree-Days. Accepted in ASHRAE Trans., ASHRAE winter meeting, Orlando, FL, Jan. 2010.

Thom H.C.S., 1954a. The rational relationship between heating degree-days and temperature. Monthly Weather Review, 82 (1) 1-6.

Thom H.C.S., 1954b. Normal degree-days below any base. Monthly Weather Review, 82 (5), 111-115.

Thom H.C.S., 1966. Normal degree-days above any base by the universal truncation coefficient. Monthly Weather Review, 94 (7) 461-465.

Threlkeld J.L., 1962. Solar irradiation of surfaces on clear days. ASHRAE J. 4, 43-54.

Yin B.T. and Wah W.Y., 2004. Clearness number and diffuse factor of the meteorological condition in Singapore. J. Instit. Eng. Singapore 1, 58-68.


David Okad, Associate Member ASHRAE, Associate, Stantec, San Francisco, CA: What plans are there to develop guidelines for forward-looking design temperatures, 20, 30, 50, 100 years out?

Didier Thevenard: A study was conducted as part of RP-1453 to look at the long-term evolution of climatic design conditions (Thevenard, D., 2010, Influence of long-term trends and period of record selection on the calculation of climatic design conditions and degree days, ASHRAE Transactions 116(1): 447-61). The study showed that over the last three decades, cooling climatic design temperatures have increased at a rate comparable to that of average temperatures, and heating design temperatures have increased slightly faster. These changes are indicative of a warming of the climate experienced by the monitoring stations. However, the magnitude of that warming indicates that it is probably less related to global warming than to the urban heat island effect; it is likely an indication of the built environment encroaching on locations where meteorological stations are situated, particularly airports. These conclusions have been included in the text of Chapter 14 of the 2009 ASHRAE Handbook--Fundamentals and can be used to adjust design conditions for the next 20 or 30 years--the typical life of HVAC systems (adjusting to 50 or 100 years may be more subjective).

(1) From the 2005 Handbook - Fundamentals, chapter 31. This model uses revised coefficients as described in Machler and Iqbal (1985).

Didier Thevenard, PhD, PEng


Christian A. Gueymard, PhD

This paper is based on findings resulting from ASHRAE Research Project RP-1453.

Didier Thevenard is principal of Numerical Logics, Inc., Waterloo, ON, Canada. Christian A. Gueymard is president of Solar Consulting Services, Colebrook, NH.
COPYRIGHT 2010 American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2010 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Thevenard, Didier; Gueymard, Christian A.
Publication:ASHRAE Transactions
Article Type:Report
Geographic Code:1USA
Date:Jul 1, 2010
Previous Article:Tight humidity control for flexible applications.
Next Article:A comparative analysis and validation of two DX cooling coil modeling methods.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |