1、2010 ASHRAE 447This paper is based on findings resulting from ASHRAE Research Project RP-1453.ABSTRACTCalculation of climatic design conditions and cooling andheating degree days using data from different decades for 1274stations worldwide reveals long-term trends. Over the lastthree decades, climat
2、ic design conditions have increased at anaverage rate of 0.76C/decade (1.37F/decade) for the 99.6%heating dry bulb temperature, 0.38C/decade (0.68F/decade)for the 0.4% cooling design temperature, and 0.28C/decade(0.50F/decade) for the 0.4% dehumidification dew pointtemperature. Annual heating degree
3、-days have decreased onaverage by 118C-day/decade (212F-day/decade) whileannual cooling degree days have increased by 68C-day/decade (122F-day/decade). These changes are indicative ofa warming of the climate experienced by the monitoringstations. However, the magnitude of that warming indicatesthat
4、it is probably less related to global warming than to theurban heat island effect; it is likely an indication of the builtenvironment encroaching on locations where meteorologicalstations are situated, particularly airports. The paper alsostudies the appropriate period of record to use for the calcu
5、-lation of climatic design conditions and degree-days. The useof a 30-year period is recommended; the use of shorter periodsof record encompassing only recent years, in order to bettercapture climatic trends, results in an added uncertainty that isgreater than the observed climate trends themselves.
6、INTRODUCTIONFor a number of years, the American Society of Heating,Refrigerating and Air-Conditioning Engineers (ASHRAE)has provided, in its Handbook Fundamentals, tables ofclimatic design conditions suitable for the proper sizing ofheating and cooling systems. These tables include values suchas dry
7、-bulb temperature, wet-bulb temperature, dew-pointtemperature, enthalpy, and wind speed at various frequenciesof occurrence over a long-term period, corresponding meancoincident values of some other parameters, and averages ofsome extremes (ASHRAE, 2005). These climatic designconditions are widely u
8、sed by the Heating, Ventilating andAir-Conditioning (HVAC) industry, and are referenced in anumber of building-performance standards, issued either byASHRAE (Standard 90.1; ASHRAE, 2004), or by otherprofessional organizations (Manual J and Manual N; ACCA,2006, 2008). The latest version of the Handbo
9、ok (ASHRAE,2005) includes climatic design conditions for over 4,400stations worldwide, including over 750 in the USA, over 350in Canada, and over 3,300 outside North America.Heating and cooling degree-days (DD) have long beenused by HVAC engineers to assess a climates severity.Degree-days can also b
10、e used to provide a simple estimate ofannual load of a building, as long as its indoor temperature andinternal gains are relatively constant (ASHRAE, 2005, ch. 32 Energy Estimating and Modeling Methods). Degree-daysare made available by ASHRAE through Standard 169(ASHRAE, 2006) and will also be incl
11、uded in the next edition(2009) of the Handbook of Fundamentals.Climatic design conditions and heating and coolingdegree-days are also available from other sources, such as theEngineering Weather Data Handbook (AFCCC, 2006).The availability of large sets of data in electronic format(Del Greco et al.,
12、 2006) now enables the calculation ofInfluence of Long-Term Trends and Period ofRecord Selection on the Calculation ofClimatic Design Conditions and Degree DaysDidier Thevenard, PhD, PEngMember ASHRAEDidier Thevenard is principal of Numerical Logics, Inc., Waterloo, ON, Canada.OR-10-048 (RP-1453) 20
13、10, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions 2010, Vol. 116, Part 1. For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted without ASHRAE
14、s prior written permission. 448 ASHRAE Transactionsclimatic design conditions and degree-days for an everincreasing number of stations, with data drawn from an everlarger number of years. As an example, for the 2001 Hand-book, 12 years of data (1982-1993) were used to calculateclimatic design condit
15、ions for 1460 locations worldwide. Forthe 2005 Handbook, 20 years of data (1982-2001) were usedfor 4,422 locations. For the 2009 Handbook under develop-ment, up to 30 years of data could be used, and over 5,400 loca-tions are expected. The use of a longer period is generallyviewed as beneficial, as
16、statistics generated from a larger poolof data are more accurate. Climate Normals, for example, aredefined by the World Meteorological Organization (WMO) as“period averages computed for a uniform and relatively longperiod comprising at least three consecutive 10-year periods”(World Meteorological Or
17、ganization, 1984). On the other hand, scientists are realizing that climate ischanging on a global scale because of anthropogenic factors,more rapidly than even experienced in the past (Trenberth etal., 2007). This has HVAC engineers starting to question thevalidity of using climatic design conditio
18、ns and degree-daysbased on the last 30 years of data. Will these conditions beappropriate for the next 20 to 30 years, which is typically thelife span during which HVAC systems designed today willoperate? Should one instead derive different design conditionsfrom shorter and more recent periods of re
19、cord, which will bemore indicative of the climate of the future rather than that ofthe past? Guttman (1989) discusses the use of 30 years of data toderive long-term averages, and warns against attaching apredictive value to normals. Colliver and Gates (2000) deter-mined as 12 the minimum number of y
20、ears needed to providereasonably accurate estimates of climatic design conditions.They did not find, however, a significant interdecade variabil-ity of the design conditions. They suggest to use up to 30 yearsof data when available, arguing that effects related to theevolution of climate are smaller
21、 than the additional variabilitythat would be introduced by using a shorter period-of-record.Hubbard et al. (2005) recommended the use of a minimum of10 years of data to derive climatic design conditions. They alsolooked specifically at the possible effect of long-term trends inclimate on design con
22、ditions, and found increases consistentwith known trends in urban warming for areas undergoingrapid urbanization; however within the limited set of stations(17) used in the study they found no evidence that globalclimate change leads to a variation in climatic design condi-tions. More recently Huang
23、 (2007) has noticed an increase incooling degree-days, and a decrease in heating degree-days,for stations in China over the 1973-2006 period.This paper will show that long-term changes in climaticdesign conditions can indeed be observed. Theses changes areclearly visible both on dry bulb and dew poi
24、nt design condi-tions, and on heating and cooling degree-days. The paper willprovide an estimate of the magnitude of this effect and showthat it is more likely correlated to urban island effect thanglobal warming. The paper also reassesses the errors intro-duced by using periods of records shorter t
25、han 30 years andprovides an estimate of their magnitude.BACKGROUND: CALCULATION OFCLIMATIC DESIGN CONDITIONS ANDDEGREE DAYSCalculation of Design ConditionsThe data tables in the ASHRAE Handbook of Fundamen-tals contain many climatic design conditions. For simplicityonly four were considered in this
26、study:the 99.6% annual dry bulb temperature;the 0.4% annual dry bulb temperature;the 0.4% annual dew point temperature;the mean daily dry bulb temperature range for the hot-test month.The x % annual design condition is the condition that isexceeded, on average, x % of the year. For example the 0.4%a
27、nnual design dry bulb temperature is the temperature that isexceeded on average 0.4% of the year, or 35 hours per year. Itis therefore representative of fairly hot conditions, and is oftenused for sizing cooling equipment. It is therefore referred to asthe 0.4% cooling design temperature. Similarly
28、the 99.6%annual dry bulb temperature is used for sizing heating equip-ment, and is referred to as the 99.6% heating design temper-ature. Finally the 0.4% annual dew point temperature is usedfor sizing dehumidification equipment and is referred to as the0.4% dehumidification dew point temperature.The
29、 daily dry bulb temperature range is the differencebetween the maximum and minimum temperature for the day.The mean daily dry bulb temperature range is simply calcu-lated as the average of the daily values over the days in thehottest month. This parameter is used for generating design-day temperatur
30、e profiles used in the Residential Cooling andHeating Load Calculations of the ASHRAE Handbook ofFundamentals (ASHRAE, 2005).For details about the procedure used to calculate climaticdesign conditions, please refer to Thevenard and Humphries(2005).For completeness, a fifth parameter was added to the
31、 listabove: the yearly average temperature, calculated as the sumof daily minima and maxima, divided by two. Calculation ofthis parameter enables one to relate changes in climatic designconditions to overall changes in the average climate.Calculation of Heating and Cooling Degree DaysHeating and coo
32、ling degree-days are often used to obtaina rough estimate the energy consumption of buildings. Heat-ing bills usually correlate well with heating degree-days base18.3C (65F), and cooling bills with cooling degree-daysbase 10C (50F). Monthly heating degree-days are calculatedwith the following formul
33、a: 2010, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. (www.ashrae.org). Published in ASHRAE Transactions 2010, Vol. 116, Part 1. For personal use only. Additional reproduction, distribution, or transmission in either print or digital form is not permitted without A
34、SHRAEs prior written permission. ASHRAE Transactions 449(1)where is the number of heating degree-days in themonth, N is the number of days in the month, is the refer-ence temperature to which the degree-days are calculated often 18.3C (65F); this is the value used in this study, and is the mean dail
35、y temperature calculated by adding the maxi-mum and minimum temperatures for the day, then dividing by2. The + superscript indicates that only positive values of thebracketed quantity are taken into account in the sum. Simi-larly, monthly cooling degree-days are calculated as:(2)where is equal to 10
36、C (50F). Min and max temperatures are calculated from hourlydata. To avoid biases introduced by missing data, the calcula-tion is performed only for days with no missing temperature;then, the number of degree-days is prorated to the total numberof days in the month. Annual degree-days, which will be
37、 usedin the rest of this study, are calculated by summing monthlydegree-days.SOURCES OF DATA ANDCOMPLETENESS CRITERIAClimatic design conditions for the 2009 Handbook ofFundamentals are calculated mostly from the IntegratedSurface Database (ISD) provided by the National ClimaticData Center (NCDC). Th
38、e ISD consists of global hourly andsynoptic observations compiled from numerous sources into asingle common ASCII format (Del Greco et al., 2006). Thedatabase comprises over 20,000 stations worldwide. Somestations have data as far back as the early 1900, although thebulk of the data is contained wit
39、hin the 1973-2006 period. ISDincludes numerous parameters such dry bulb and dew pointtemperatures, wind speed and direction, station pressure,cloud cover, present weather, and depending on the station,various other elements. Data originate from the AutomatedSurface Observing System (ASOS), the Globa
40、l Telecommu-nication System (GTS), and various other sources. For moststations, data is available either on a 1-hour or 3-hour basis.For some Canadian locations, the Canadian WeatherEnergy and Engineering Data Sets (CWEEDS) weather fileswere used (Environment Canada, 2007). This datasetcomprises dat
41、a that tends to be serially more complete thanthat in the ISD. It also extends the geographical coverage tosome additional locations not present in the ISD.The calculation of climatic design conditions is possibleonly for stations for which the period of record is completeenough. Gaps up to 6 hours
42、in the time series are filled bylinear interpolation. Then, whole months are accepted in theanalysis, or rejected from it, based on completeness criteriadescribed in Thevenard and Humphries (2005). The complete-ness criteria ensure that no seasonal or day/night bias is intro-duced in the analysis by
43、 poor quality data, and that thestatistics are derived from sufficient data.METHOD OF ANALYSISOverviewTo study the influence of period of record on designconditions and degree-days, we calculated these variables fora large number of stations for different periods. Then, for eachstation, we calculate
44、d the difference between design condi-tions for two periods. Because of the natural variability of theclimate, not all stations will experience the same change. Theresults are best plotted as a histogram that groups the stationsby magnitude of the change. The average magnitude of thechange can also
45、be calculated. The use of a large number ofstations enables to obtain a statistically meaningful picture ofthe evolution.Years Used for CalculationTwo competing effects are at play. Because of the normalvariability of the weather, the use of a shorter period of recordresults in more uncertainty in t
46、he determination of the climaticconditions or degree-days; but their values can also be affectedby long-term changes in the climate. These two effects have tobe studied separately.To study the effect of long-term climatic changes,climatic design conditions and degree-days were calculatedfor a number
47、 of successive or overlapping 10-year periods:1977-1986, 1982-1991, 1987-1996, 1992-2001, and 1997-20061. By contrast, to study the effect of shorter or longer peri-ods of record, climatic design conditions and degree-dayswere calculated for periods of decreasing length ending in2006: 1977-2006 (30
48、years), 1982-2006 (25), 1987-2006 (20),1992-2006 (15), 1997-2006 (10).Method for Calculating the Design ConditionsClimatic design conditions calculated using various peri-ods of record have been published in previous editions of theHandbook, which is updated every four years. However thesepreviously
49、 published values were not used. The reason is thatthere have been significant changes in the calculation methods(Thevenard and Humphries, 2005) or even in the completenessof the hourly data used, that would cast doubts on the compar-ison between data from recent Handbooks and data fromHDDbTbTi()+i 1=N=HDDbTbTiHDDbCDDbTiTb()+i 1=N=Tb1.The CWEEDS data set used for this study ends in 2005. For thatreason all dates were shifted by one year earlier for the 88 stationsconcerned. That is, where 1977-1986 was used for ISD stations,1976-1985 was used for CWEEDS stations
