Loading...
HomeMy WebLinkAboutAK Snow Loads 1973USACRREL August 1973 _j -, -, -CJ ' ~ _] _j "" _j _.. _j r- LO LO "¢ 00 00 0 0 0 LO LO I' ('I) ('I) -' ARLIS Alaska Resources Library & Information SerVices Ancl 3 ~ ~~ ask.t=J ALASKAN SNOW LOADS by Wayne Tobiasson and Robert Redfield Presented at the 24th Alaskan Science Conference University of Alaska August 1973 U. S. ARMY COLD REGIONS RESEARCH AND ENGINEERING LABORATORY HANOVER, NEW HAMPSHIRE INTRODUCTION Very little specific information on snow loads is available for Alaska even though for most of the state, snow loads are the maximum climatic loads induced on structures. The Uniform Building Codel7 is perhaps the most widely used building code in Alaska and it simply states that "snow loads shall be determined by the Building Official." A few large communities arm the Building Official with a specific snow load. The Building Code of the City of Fairbanks9 adopts the Uniform Building Code, replacing the q_uoted sentence above with "The snow load. is hereby determined to be 40 pounds per sq_uare foot 11 • While on the subject of Fairbanks it is interesting to note that current design snow loads there vary from 30psf to 65psf depending on the reference chosen.9,10,ll,l3 In most other Alaskan communities, lack of criteria, not variations depending on data source, is the problem. Few Building Officials have direct knowledge of appropriate snow loads and snow load q_uestions are freq_uently referred to engineers and architects who generally have never measured a snow load but through experience, have suggestions for design. Many refer to the map "Design Data for Military Construction in Alaska"10 prepared by the Alaska District, Corps of Engineers in l958 or the ioint Army-Air Force _Tech- nical Manual "Load Assumption for Buildings" 1 issued in 1966. Both documents are currently under revision with significant changes anti~ cipated. Until recently the 1955 American Standards Association publication "Minimum Design Loads in Buildings and Other Structures"2 was used to establish loads for the "lower 48" states. It did not contain loads for Alaskan areas. In 1972 that document was superseded by the American 1?/ y95 .71P'I ;913 National Standards Institute (ANSI) puhlication·, "Building Code Req_uire- ments ;for Minimum Design Loads in Buildings and Other structures"l but again snow load criteria-for Alaska were absent. National Weather Service records for Alaska were analysed for inclusion in the 1972 ANSI standard but they were not published sine~ many of the ground snow load measurements were of q_uestionable value.2 Most design snow loads currently in use in Alaska are essentially opinions based on experience. The vast·majority of Alaskan structures hold up well under snow loads and in that light there is a tendency to believe that since structures are not collapsing~ the proper loads are being used. Of course, wasteful overdesigns may also be occurring. When a sound bui·lding fails under snow loads, authorities generally react by increasing the design snow load for the region rather than introducing provisions to account for special loading conditions. Failures are seldom documented much beyond the point. of stating that "the building collapsed under a heavy snow load". Several snow load case histories and a. few well-documented building failures indicate that most structures are significantly overdesigned. but an occasional structure is underdesigned·because attention has not been paid to special situations such as drifting and sliding snow. To improve upon this situation new snow load criteria have been developed based on a statistical analysis of weather records and review of snow load case. histories. The elevation, the local site conditions and the geometric, thermal and aerodynamic features of facilities have been considered. INFORMATION SOURCES Except for· recent work by Isyumov18 who is developing a method of predicting snow loads by considering daily snowfall, air temperature, windspeed, wind direction and roof properties, it is generally accepted that knowledge of the snow load on the ground is the first step in·. developing snow load criteria for roofs. Much of the weather data needed for Isyumov' s approach is not available for Alaskan stations. There are only 18 National Weather Service stations in Alaska that measure 1~oth the dept·h and load (i.e. water eq_ui valent) of snow on the ground. Depths are measured in inches of ~ and gravimetric measure- ments of load are presented in inches of water. By multiplying inches of water by 5.2 the load can be converted to psf. Unfortunately the ground snow load information collected at ten of the eighteen stations is either an estimated value based on the assumption that 10 inches of snow has a water eq_uivalent of one inch (ie the specific gravity of the snow is 0.1) or is unreliable due 2 [ [ .[ . [J [ [ [ c c [ [ [ L [ . [ "[j [ [ c-> -, -. _j -, __j . , -' ' oJ ' :J ' ~ -, _. -, __; .... to lack of trained observers.22 The eight remaining stations report valuable ground snow load information but eight widely-separated stations do not provide much of a base on which to develop state-wide criteria. Fortunately a second source of snow load information is available. USACRREL initiated and has co-sponsored a program of detailed snow ob- servations at 17 stations in Alaska for periods from 7 to 17 years. The United States Air Force, Air Weather Service and the United States National Weather Service (formerly the U. S. Weather Bureau) participated in the data collection.* The program aud several relationships developed from the data are described by Bilello. ,5 By combining the 8 National Weather Service stations and the 17 sites studied by USACRREL, 25 loca- tions in Alaska with reliable ground snow load information were obtained . Perhaps it would be possible to generate state-wide criteria using only these 25 stations but a far better job can be accomplished by also considering snow depth measurements available at these stations and at 112 additional smaller weather stations throughout the state. Peri.ods of record range from 4 to 23 years depending on location. Where only snow depths are measured, the density of that snow must be estimated before snow loads can be computed. The climatological series of the maximum annual depth of snow on the ground at 137 locations has been analysed statistically. Then, using the combined depth and load information available at 25 of these locations, conversion densities were developed, regionalized and applied to all 137 locations to generate ground snow loads at each site. STATISTICAL ANALYSIS OF DEPTHS Extreme value statistical studies by Thom 2 3 indicate that climato- logical series of annual maximum snow depths on the ground closely follow log-normal distributions (ie when plotted on log-normal probability paper the distribution is linear). He uses the mean and standard deviation of the logarithms of the series to establish confidence intervals, then aids the reader in visualizing his complex statistics by presenting results using Blom plotting positons.6 Where a series consists of 10 or fewer values Blo~'s positions are preferred over the more common positions of Gumbell5 ,l . Both plotting positions are defined below: Probability (Blom) *We are indebted to Mr. Michael Bilello, Research Meteorologist, USACRREL, for permitting us to utilize the snow load information collected under his direction and to Mr. Roy Bates, Meteorological Technician, USACRREL, for guidance in the reduction and analysis of this information. 3 Probability (Gumbel) = (n: 1 ) 100. where: n = number of years of record m = rank of a particular annual extreme value. For the smallest value, m = 1; for the largest, m = n. Probability is in percent. All statistical results in this report are based on the use of Blom's plotting positions since a third of the locations have fewer than 10 years of snow depth records. Representative log-normal prob- ability plots of the annual climatic series of maximum snow depths on the ground are shown in Figure 1. Return periods can be used to relate probabilities to the "design life" of a facility. Return periods for annual series are defined as follows: 100 Return period = 100 _ Probability where probability is in percent and return period is in years. A return period scale is superimposed on the probability scale of Figure 1. The snow depth from a regression line at a specific return period is the depth that can be expected to be equalled or exceeded once during that period. For example in Figure 1 the snow depth for a 25-year return period at Cape Lisburne would equal 3.1 ft. The snow depth corresponding to 5, 10, return periods have been determined for 137 a digital computer. The Fortran II program is presented in Appendix A of this report. lations of both the input data and results Periodic updating of the data and revision tively easy tasks. 25, 30, 50 and 100-year locations in Alaska using developed for this task Samples of computer tabu- are presented in Appendix B. of results will be rela- RELATING DEPTHS AND LOADS An annual record of snow depth and load for two representative locations is presented in Figure 2. In the simplest case, such as at Cape Lisburne, both the depth and load maximize on the same date. 4 [ [ .n -n c [ [ c c [ [ [ [ [ .[ ·[ [ [ -' ;_.j •. -' -' --' -' =' Return Period, yrs 30 4.0 • I I I I I I I I I r 1P 2 r 1 s,o ~g.o I --"C 3.0 c: :1 0 ... (!) Q) ..c. -c: 0 3: 0 c: 2.0 (/) 1.0 -0 ..c. Q. Q) 0 CAPE LISBURNE 0 ·5 I ~ ; ~~ do 3 1o 4 1o ~o ~o io ~o So 9~ I ~8 Ss ~9.5 96.7 Probability, % Return Period, yrs 9.0 i I I I I I I I 1 1 7 1P 2 ?3 1° 5,0 1 ?~ 1 -- 8.0 7.0 6.0 "C 5.0 ; 4.0 0 .... (!) Q) 3.0 ..c. c: 0 3: 2.0 0 c: (/) -0 ..c. -~ ~ 1.0 UTOPIA CREEK 0.5 I . 2 5 10 20 30 40 50 60 70 80 Probability, % 90 96 1 9a 99 99.5 96.7 Figure 1. Log-normal probability plots for Cape Lisburne and Utopia Creek 5· However for many locations, such as Utopia Creek, the maximum load occurs at a later date than the maximum depth. When the load and depth maximize concurrently, the ratio of the maximum load, in psf, to the maximum depth, in feet of snow, is equal to the density of the snow in pcf on that date. When they maximize on different dates the ratio is not a physically-measurable density, simply a convenient conversion factor relating depth and load. This factor has been termed conversion density. If the depth and load maximize on different dates, the conversion density will always be less than the density measured at the time of maximum load but greater than that measured at the time of maximum depth. 6r 60 I CAPE LISBURNE -"' Q. -r~ ~4 g4o -....I Q. cu 'S 0 0 c: 31: (J') ~ 2 ~ 20 e C) ol 0 6r 60 I UTOPIA CREEK -"' c.. = t~ -4 0 40 .c 0 ii ....I cu 31: 0 0 c: 31: (J') 0 (f> 2 2! 20 ::> e C) ol 0 Figure 2. Record of snow depth and load for one season for Cape Lisburne and Utopia Creek. 6 [ [ .L . [J 0 [' [ [ c [ [ [ [ [ [ .[ :[ [ [ ' _j ~, _j ~ _i "" ;___j -' _j ,. DEVELOPING CONVERSION DENSITIES For the 25 locations where both snow depth and water equivalent are measured, each annual conversion density was determined. Perhaps the most useful grpahical presentation, of the many investigated, was the relationship between each annual conversion density and its associated maximum annual snow load for a particular location. Two such graphs are shown in Figure 3. The density trend varies but for each station it is possible to extrapolate to a constant density for heavy snow loads. Similar curves were produced for all locations where ground snow loads were measured. The results indicated that a single conversion density could be established for each location regardless of the length of the return period. The conversion densities thus developed are presented in Figure 4. 30~---r--- -8. ,:.. -f/1 20 c: Q) c c: 0 ·c;; .. Q) 10 > c: 0 (.) 0 30 -I u 0. ,:.. ~ 2or- c: Q) c IJ c: 0 f/1 .. Q) > c: I 0 (.) 0 CAPE LISBURNE • • -----·---r---. _......-.r • --- .r/ Trend • • 10 20 30 40 50 Ground Snow Load, psf UTOPIA CREEK • • Trend • __ :__. ___ ...,._. _____ ---;-. . • • 10 20 30 40 50 Ground Snow Load, psf 60 60 Figure 3. Graphs used to establish conversion densities for Cape Lisburne and Utopia Creek. 7 I I \ I I \ I I 'b 27.5\ '--v-J ~11.9 L~ '1( I\ I?" ..,.,. I ~ 0 0 • • Figure 4. Conversion densities for 25 locations in Alaska. Using information in "Snow Surveys for Alaska"14 , the U. S. Geological Survey physiographic map of Alaska2 5, and Bilello's map of average seasonal snow-cover densities5 as guides, conversion densities were regionalized. The conversion density in each region was further categorized according to site elevation and proximity to the coast. Regionalized densities are presented in Figure 5. These values were then applied to the design snow depths determined for 5, 10, 25, 30, 50 and 100-year return periods. ' GROUND SNOW LOADS The place name for each location, its latitude, longitude and elevation and the ground snow load in psf for 6 return periods are pre- sented in Table I. For temporary facilities the 5-year return period snow load should be selected. For permanent facilities the 25-year return period is appropriate. However, as hazards to life and property increase in the event of a failure, design loads should increase toward the 100-year value. The snow load values for 23 locations are suspect 8 [ r L fJ ·C [j [ [ c c [ [ [ [ [ -C ·[ [ l -" -, _j -' -' ~ .:oastal28.1~ coastal 25.0 ~ 0 0 Conversion Density ( ;>cfl Region Oto500' 500to ooove 1500' 1500' A B c ~ D 12.5 15.6 18.7 25.0 15.6 18.7 21.8 21.8 18.7 21.8 25.0 21.8 \ Note: See mop for ccostol v:lues ;--~..,--c ~', ' (;~'- 000"0116.~~ coastal 15.6 Figure 5. Regionalized conversion densities. and are followed by an asterisk in Table I. Sixteen of the suspect locations have only 4 to 6 years of record. Since most of these limited records cover the heavy snow year of 1970-71, excessive statis- tical values generally result. Suggested 25-year return period values, based on comparisons with other stations in the vicinity and considera- tion of the 1970-71 snow loads as generally representing more than 15-year values, are shown in-parentheses in Table I. Suggested values for other return periods can be obtained by multiplying the asterisked value for the desired return period by the ratio of the suggested 25-year value (in parentheses) to the asterisked 25-year value. The National Building Code of Canada3 '8 '20 uses a single conver- sion density of 12pcf (.19 gm/cm3) throughout Canada to convert 30-year return period snow depths to ground snow loads. A surcharge equivalent to the weight of a one-day spring rainfall is added. The rainfall surcharge varies from less than 5psf in the Canadian prairies and Arctic to 25psf on Vancouver Island.7 Such a surcharge is not needed for the Alaskan loads presented in this report since spring rains are already part of the maximum annual water equivalent observations. 9 Station name Adak Adult Conservation Camp Allakaket Alpine Inn Alyeska Anchorage Angoon Aniak Annette Annex Creek Attu Auke Bay Barrow Barter Island Beaver Falls Bethel Bettles Big Delta Birch Road Cape Decision Cape Hinchinbrook Cape Lisburne Cape Newenham 9ape Romanzof Cape Saint Elias Cape Sarichef Central Chena Hot Springs Chignik Chitina Circle Hot Springs Clear Clearwater Cold Bay College Magnetic Observatory Cooper Lake Project Cordova Crooked Creek Eagle Eielson AFB Table I. Ground snow loads. Elevation, Latitude Longitude ft 51°53• 61°42• 66°34• 61°43• 60°58' 6101Qo 57°30• 61°35• 55°02• 58°19• 52°50• 58°23• 71°18• 70°08• 55°23• 60°47• 66°54• 64°00• 61°08• 56°00• 60°14• 68°52• 58°38• 61°46• 59°48• 54°36' 65°34' 65°03• 56°18• 61°32• 65°29• 64°18• 64°03• 55°12' 64°52• 60°23' 61°58' 61°52• 64°46• 64°40• 176°38' 148°59• 152°40• 148°54• 149°08• 150°01• 134°35• 159°32• 131°34• 134°06• 173°11'E 134°38• 156°47' 143°38• 131°28' 161 °48• 151°31• 145°44' 149°46' 134°08• 146°39' 166°07• 162°04• 166°03• 144°36' 164°56• 144°49• 146°03• 158°24• 144°27• 144°38' 149°11' 145°31• 162°43• 147°50• 149°40• 145°19• 158°06• 141°12• 147°06• 15 825 600 455 251 114 15 81 110 24 70 42 31 39 35 125 666 1268 460 39 185 45 475 434 58 175 960 1195 30 575 935 580 1100 96 621 445 1000 125 821 558 10 Ground snow load, psf '5yr. lOyr 25yr 30yr 50yr lOOyr 23 28 30 41 61 80 19 22 98 122 52 64 50 64 50 63 20 30 133 158 51 66 61* 79* 43 50 66 82 53 72 36 37 57 60 111 113 27 28 154 159 80 82 83 86 79 82 43 45 191 197 86 89 103* (100) 107* 59 60 104 108 101 105 42 71 131 30 180 92 98 92 54 216 102 123* 65 121 124 36 46 61 92 111 137 63 73 140 156 86 102 43 59 82 62 66 72 72 75 15 20 28 29 34 57 82 121 62 74 90 60 78 104 73 116 148 66 89 121 127 155 92 101 108 125 182 222 126 148 15 37 - 59 62* 38* 30 52 43 20 59 34 65* 42 61 62 21 39 77 83* 60* 33 65 58 24 71 39 79* 49 74 80 29 30 42 42 101 105 113* (80) 118* 100* (80) 108* 36 36 81 83 79 82 29 29 87 90 46 47 98* (110) 101 * 58 60 90 93 105 110 35 44 120 138* 139* 38 93 96 32 100 50 112* 65 104 126 48 86 156 34 206 105 114 106 67 240 119 168* 72 139 150 88 176 124 78 42 193 114 148 232 177 43 46 141 165* 187* 41 106 115 35 112 55 127* 72 116 147 [ [ .h [ [ [ [ [ c [ [ [ [ [ [' d ·[ [ [ -, -, _j _j _j -' -, --, --' :::ii Station name E klutna Lake Elmendorf AFB Fairbanks Farewell Five Finger Light Station Flat Fort Yukon Galena Gilmore Creek Glacier Bay Glennallen Gulkana Gunsight Haines Terminal Holy Cross Homer Hughes Iliamna Intricate Bay Juneau Kasilof Kasitsna Bay Kenai Ketchikan King Salmon Kitoi Bay Kobuk Kodiak Kotzebue Lake Minchumina Linger Longer Little Port Walter Mankomen Lake Manley Hot Springs Matanuska Agric. Exp. Sta. McGrath· McKinley Park Moose Pass Moses Point Nenana Table I (Cont'd). E{evation, Latitude Longitude ft 61°24' 61 °15• 64°49' 62°32' 57°16• 62°29• 66°33• 64°44' 64°59' 58°27• 62°07' 62°09• 61°54• 59°16• 62°10' 59°38• 66°04• 59°44• 59°34' 58°22' 60°19• 59°28• 60°34• 55°21• 58°41• 58°11' 66°54• 57°45• 66°52• 63°53' 59°26• 56°23• 62°59• 65°00• 61°34• 62°58• 63°43' 60°28• 64°43• 64°33• 149°09• 149°48• 147°52• 153°54• 133°37• 158°05• 145°12• 156°56• 147°31 1 135°53 I 145°32 1 145°27• 147°18• 135°27• 159°45• 151 °30• 154°15• 154°57 1 154°28• 134°35• 151°15 1 151°33 1 151°15• 131 °39• 156°39• 152°21 I 156°52• 152°31• 162°38• 152°17 1 136°17• 134°39 1 144°29• 150°39 1 149°16 1 155°37• 148°58 1 149°23• 162°04• 149°05 1 882. 222 436 1499 30 309 443 120 959 50 1456 1570 2960 175 200 67 545 145 170 12 75 12 86 15 49 15 140 21 10 701 700 14 3025 300 150 344 2070 485 15 356 11 Ground snow load, J!.Sf 5 yr 10 yr 25 yr 30 yr 50 yr 100 yr 47 42 52 38 24 55 50 56 48 63 48 38 68 57 52 59 57* 78* 85* 107* 29 34 42 52 56* 77* 67 91 68 82 45 57 62 77 46 59 34 40 46 -60 61 72 82* 103* 55 68 27 37 67 56 78 60 64 85 65 69 57 80 62 69 87 66 69 70 109* (90) 114* 137* (100) 142* 41 43 66 68 108* (90) 113* 126 132 100 103 74 76 97 101 77 80 47 48 79 83 85 87 131* (90) 136* 84 86 53 56 76 62 89 70 89 97 71 76 135* 161* 47 77 135* 156 113 86 113 91 52 95 96 153* 96 67 22 26 31 31 34 43 60 86 87 106 132 91 109 136 152 31 38 49 51 58 56 65 77 79 86 52 60 70 72 77 119 137 102 134 103* 130* 50 59 19 24 51 60 71 87 78 98 105* 127* 45 58 160 163 179 186 167* (120) 173* 70 72 31 32 71 72 109 112 126 130 156* (125) 161 * 76 79 177 215 197* 78 37 79 125 148 179* 90 85 68 101 79 120 110 77 82 164* 187* 52 86 164* 199 127 100 129 107 57 112 106 176* 109 83 37 134 172 66 95 84 193 255 227* 86 43 87 142 171 202* 106 Station name ~ikiski Terminal ~orne Northeast Cape Northway Nunivak Oil Well Road Palmer Paxson Lake Petersburg Point Hope Point Lay Point Retreat Light Station Port Alexander Port Alsworth Port Heiden Puntilla Richardson Russian Mission Saint Marys Saint Paul Island Seldovia Seward Shemya Shishmaref Sitka Sitkinak Skagway Slana Snowshoe Lake Sparrevohn Summit Talkeetna Tanacross Tanana Tatalina Teller · Thompson Pass Tin City Tok Tonsina Lodge Table I (Cont'd). Elevation, _ Ground snow load, psf Latitude Longitude ft 5· yr 10 yr 25 yr 30 yr . 50 yr 100 yr 60°41• 64°30 1 63°17• 62°55• 60°23• 61°14' 61°36' 62°57• 56°49' 68°20• 69°45• 58°25• 56°15• 60°12• 56°57• 62°06• 64°17• 61°47' 62°04 1 57°09• 59°26• 60°07• 52°43• 66°14' 57°03• 56°33• 59°27' 62°43' 62°02• 61°06• 63°20• 62°18• 63°24• 65° 10• 62°54• 65°16• 61°08• 65°34• 63°21• 61°40• 151°23• 165°26• 168°41• 141°56• 166°12• 149°43' 149°06• 145°30• 132°57' 166°48' 163°03• 134°57' 134°39• 154°18• 158°37• 152°45' 146°22• 161°19• 163°11' 170°13• 151 °42' 149°27' 174°06•E 166°07' 135°20• 154°08• 135°19• 143°44' 146°40• 155°33• 149°09• 150°06• 143°19' 152°06• 155°58• 166°21• 145°45' 167°55' 143°02• 145°11' 110 13 38 1713 44 370 225 2750 50 15 10 50 18 260 92 1832 875 50 25 22 31 70 122 14 67 53 10 2200 2410 1580 2401 345 1549 232 964 10 2500 269 1620 1500 12 57* 75* 70 89 90 114 51 61 57 74 40 45 32 45 67 79 60 84 56* 101* 99* (80) 103* 118* 114 118 134 148 154 175 74 76 84 98 102 117 51 52 56 63 66 79 95 97 106 120 127 152 189* (90) 207* 284* 140* 155 204 94 137 60 97 118 186 410* 35 44 57 58 66 76 40 58 87 92 112 142 56 72 95 99 113 133 24* 29* 35* (75) 36* 39* 44* 15 21 30 31 38 47 79 90 102 104 112 121 57 73 95 99 113 132 52 70 95 99 115 137 19* 23* 27* (100) 28* 31 * 34* 37 48 63 65 75 88 70* 103* 50 63 16 20 38* 40* 34 44 25 37 11* 13* 42 60 43' 49 72 98 122 148 106 130 4'8 57 59 74 70 78 41 66 176 223 52 70 42* 46* 35 37 156* (90) 80 26 42* (90) 57 166* 204* 83 93 26 30 42* 43* 60 74 56 60 74 14* (100) 14* 15* 88 94 114 56 57 60 136 142 168 180 186 205 161 166 186 69 71 77 94 97 109 88 89 95 109 117 151 288 299 340 97 102 120 50* (70) 50* 53* 40 40 42 261* 107 34 44* 95 95 16* 142 65 204 230 210 86 126 101 203 394 144 55* 43 [ [ .[ ·[ [ [ [ r~ , L• c~ [ [ [ [ [ [' . ' ·[ L [ ' . ,, --' _j ~. : -' _j _j -' -' _, . ..J, ..., Table I (Cont'd) . Elevati.on, Ground snow load, LJ.Sf Station name Latitude Longitude [t 5 yr 10 yr 2s rr. 30 yr 50 yr 100 yr Trims Camp 63°26' 145°26 1 2408 226* 299* 404* (200) 421 * 491* 585* Umiat 69°22' 152°08• 337 49 57 68 70 76 84 Unalakleet 63°53 1 160°48' 15 62 82 110 115 134 159 University Experiment Station 64°51' 147°52• 475 46 56 68 70 78 87 Utopia Creek (Indian Mtn.) 65°59' 144°29' 1220 66 85 111 116 133 155 Valdez 61°08• 146°15• 49 136 159 18'7 191 207 228 Wainwright 70°40' 159°50• 315 26 29 33 34 36 39 Wales 65°37• 168G03• 9 49 64 85 88 102 121 Wasilla 61"37• 149°24• 500 51* 74* 108* (60) 115* 140* 176* West Fork 65°28• 148°40' 425 39* 45* 54*(80) 55* 60* 66* Whittier 60°47' 148°4l' 15 190 250 334 349 404 478 Wild Lake 67°33' 151°33' 1190 46* 49* 54*(120) 55* 57* 60* Willow Trading Post 61''47' 145°11' 1400 82* 99* 119* (90) 123* 135* 151* Wiset~mn 67"'26• 150"13' 1286 84 103 128 132 148 167 Wrangell 56°28• 132°23' 37 32 46 67 70 85 106 Y:1kata~~a 6o··or)· 1:12":10• 27 80 102 132 138 157 182 Yakutat G\l'':1 1 • l:l9"40• 28 122 149 18·1 190 212 2·10 The conversion densities developed for Alaska vary from a low of 12pcf (0.19 gm/cm3) in the Copper River lowland to a high of 28pcf (0.45 gm/cm3) along the west coast below the Seward Peninsula. In the interior, on the Alaska Peninsula, on the Aleutians and on the "Panhandle," densities vary between 16 and 25pcf (.26 and .40 gm/cm3) depending on location and elevation. For the wind-driven snow on the North Slope a density of 24pcf (.38 gm/cm3) is used • Maps of Alaska overlaid with ground snow load isolines have not been made since they tend to obscure local variations and may result in hasty generalizations. To determine ground snow loads for sites not listed in Table I the loads reported at several sites in the vicinity should be inspected with attention paid to elevation and other geographical features. Such an approach will not only produce more meaningful cri- teria but will also alert the user to the extent of local variations in the vicinity. In some areas a 50-mile change in location has little influence on loads but in other areas loads might triple in the same distance. 13 Local variations are particularly significant in southeastern Alaska. The significant variation in the depth of snow on the ground for four locations all within an 11-mile radius of Juneau is shown in Figure 6. \__ 7r------.-----.------~----~------~----~ ..... 6 -~ 'U c: ::l 5 0 ...... (!) Cl) -!: 4 c: 0 == g 3 (f) -0 :: 2 a. Cl) 0 ----Annex Creek --Auke Bay ........ Juneau -·-Pt. Retreat Light Sta. {'\... A /f\ .'\ I~ /: ... \. f:.\ . \ '-\ .. \ . . .. 1:\ ):~. . .. \, .. . . .. \ . . . . . . ·.. . ~ . . . . . I = ·.. . ~ . I : . ... : ~ \ . . . . .. ~. ·. . . . / .......... . . . ·-.. . 1\ . --..... . . . . ···~ _J, .. · ·......:...··-.. :/· . -.· ~······...,.(.•' ---.... . ....., 0 I ~··"· -.i. ¥ r· .. Jan I Feb I Ma~·-·r ... \.A~;"""I ---.1965 1966------- Figure 6. Depth of snow on the ground for four locations all within an 11-mile radius of Juneau. Most snow depths reported for mountainous regions were measured in the populated valley areas. Such depths are often significantly less than those at high elevations in the area. A few measurements of the depth or water equivalent of snow on the ground are of little direct value in establishing design loads. However, if such measurements are also obtained at one or more of the locations listed in Table I at the same time, the ratios can aid in the transfer of criteria from one location to another. Such measurements should be taken at intervals during the late winter and early spring when the snow load is near its seasonal high. 14 [ [< .[ f J [ l [ [ [ [ [ [ [ [ L -[ L [ __. •. ""! ' ~ J . , _J _j ROOF LOADS Ground-to-roof conversion Snow loads on roofs are affected by local winds and temperatures, the exposure of the structure, its thermal characteristics and geometry and its aerodynamic position. Factors have been developed to consider these variables. The factors are based on "Commentary No. 2 Snow Loads" in the Canadian Structural Design Manua1 20 with modifications based on snow load observations in Alaska by the U.S. Army Engineer District Alaska and USACRREL and in Canada by the National Research Council.l9,2l The following equation has been developed to relate the "basic roof snow load" to regional winds, exposure of the structure, roof thermal characteristics, and ground snow loads. pr = C C Ctp r e g where: " pr c r = basic roof snow load in psf = regional ground-to~roof conversion factor which considers local winds and temperatures C = exposure of the structure e ct = thermal characteristics of the roof pg = ground snow load in psf for the appropriate return period • Appropriate values for Cr' Ce, Ct and p are listed in Tables II, III, IV and I respectively. g The "basic roof' snow l.o~," p , requires further modification to account for nonuniform and 'unbalan~ed loads., roof slope, extra snow collected in valleys, sliding of snow onto lower·roofs and wind drifting of snow onto .roofs located in areas of aerodynamic shade. M:Lnimum loads All roofs should be designed to sustain a minimum uniform live load of 20psf except that a 15psf minimum load can be used for unobstructed metal roofs with a slope greater than 6 on 12 and fabric roofs with the vertical angle from the eave to the crown greater than 34°, since such roofs will shed snow by sliding. 15 Table II. Regional groun~-to-roof conversion factor, Cr. ~ "3 t!J?..P Region Arctic Slope Northwest Inland Coastal and mountainous Yukon Southwest Mountainous Other areas South Central Coastal Other areas Southeast Yukon l6 cr .4 .5 .4 .5 .5 .4 .5 .6 .5 \ \ [ [ c .L r-, ·u [ L [ [ [ [ c [ [ [ .[ ·[ L L _, .; -' _. -' ~. _j Table III. Exposure factor, C e. Siting of structure Windswept Suburbs with few trees Near some trees or other windbreaks In among trees Table IV. Roof thermal factor, Ct. Thermal condition Heated building with unventilated roof having conventional insulation (R < 15) As above. but ventilated Heated building with unventilated, well- insulated roof (R > 15) As above but ventilated . Building kept just above freezing Unheated building Full and zero loads ce 1.0 1.1 1.2 1.3 ct 1.0 1.1 1.1 1.2 1.3 1.4 For all roofs the effect of removing the load from any portion of the loaded area should be investigated. In some instances, unloading certain areas will induce heavier stresses in the roof than with the entire roof loaded. Cantilevered roof joists are a good example: removing the load from the cantilevered portion will increase the bending stress and deflection at center span. In other situations undesirable stress reversals may result. Roof slope All loads acting on sloping surfaces should be considered to act on the horizontal projection of that surface. 17 The basic roof snow load, p , should be used without modification for simple shed, gable and hip r~ofs having a slope of 3 on 12 (14°) or less and for domed or vaulted roofs where the centerline rise is less than one tenth the span. For unobstructed metal roofs where snow can slide off the eave, the flat roof snow load can be reduced by the roof slope factor, C , determined using the dashed line in Figure 7. s 4 on 12 3 6 8 12 on on on on 1.0 12 12 12 12 0.81 \ \ ~II Other Surfaces I 0.6. cs I \ I 0.4, ·\ Metal or Other\ 0.2t-Slippery Surface \ 0 40° 60° 80° Roof Slope Figure 7. Graph for determining the roof slope factor, C s 18 [ [ n [ [ [ [ c [ c [ [ [ c [. • • ·[ [ [ j _. ., -' _.1 -, _j ~ _; _; For other roofs that cannot be relied on to shed snow loads by sliding, the solid line in Figure 7 should be used to determine the slope factor, C . s For curved roofs with the rise greater than one tenth the span, the roof slope factor, Cs, should be determined from the appropriate curve in Figure 7 basing the slope on the vertical angle from the eave to the crown. The dashed line in Figure 7 should be used for fabric structures, including inflatables. Unbalanced loads For hip or gable roofs with a slope exceeding 4 on 12 (18°) the structure should also be designed to sustain an unbalanced uniform load on the lee side equal to 1.25 times the balanced load. In the unbalanced situation the windward side shall be considered clear of snow. For domed and arched roofs with a rise greater than one tenth the span the lee side unbalanced load should be a triangular distribu- tion increasing from zero at the crown to 2.0 times the balanced load at the eave. Lower roofs (aerodynamic shade) Snow drifts will accumulate on roofs in the wind shadow of high~r roofs. The affected roof may be influenced by a higher portion of the.same structure or by another structure nearby if the separation is 20 ft or less. When a new structure is built near existing structures drifting possibilities should be investigated for the existing struc- tures and the new structures. The drift load will 2 ~ecrease.as ~he spacing between structures increases. The factor -spa~Ong ln feet can be used to decrease drift loads for spacings to 20 ft. For spacings greater than 20 ft drift loadings need not be considered; The surcharge load due to drifting can be approximated by a triangular distribution assuming that the drift rises to the level of the higher roof at the edge of the lower roof and tapers to zero at a distance along the roof equal to 2 times the clear height between the two roofs. The clear height is measured from the top of the balanced snow load on the lower roof to the closest point on the upper roof. If the clear height exceeds 5 ft the values for a 5-ft clear height will be used to establish the extent and magnitude of the sur- charge load. To determine the thickness of the balanced snow load and 19 the weight of snow in the drift, a density of 20pcf can be assumed. The drift load should be div~ded by the exposure factor, Ce, since the potential for drifting will also be influenced by site factors. Multiple folded plate and barrel vault roofs Such roofs collect excess snow in the valleys by wind drifting and by sliding. To account for the excess no reduction in load is allowed for slope as for shed, gable, hip and curved roofs. The redistribution of load toward valleys also requires consideration of nonuuniform loading for all such roofs having a slope or equivalent slope exceeding 2 on 12 (10°). The nonuniform load should decrease uniformly from a value two times the flat roof load Pr' at the valley to zero at the ridge. The total weight of snow on the roof is the same for the uniform and tri- angular loadings. For sawtooth and similar roofs with one surface vertical or nearly so, uniform loads and nonuniform loads are developed in a similar manner. Sliding snow Snow may slide off metal, plastic or fabric surfaces but usually remains on wood, composition shingle or built-up surfaces unless the slope exceeds 45°. Situations which permit snow to slide onto lower roofs should be avoided. Where this is not possible the extra load added by sliding snow should be considered. The final resting place of the sliding snow will depend on the size, position, and orientation of each roof. Distribution of sliding loads might vary from a uniform load 5 feet wide if a significant vertical offset exists between the two roofs, to a 20-foot wide uniform load where a low slope upper roof slides its load out over a second roof only a few feet lower much like a flowing glacier. For conditions where a portion of the sliding load is expected to also slide clear of the lower roof an appropriate percentage of the upper roof load should be used in the calculation. The Canadian Code 20 suggests using 50% of the upper roof load for the general case. Where all the upper roof load can be expected to remain on the lower roof after sliding, the full load should be considered. If the upper roof surface is metal, the upper roof load for the sliding load calculation should be based on the solid line in Figure 7 not the dashed line. 20 [ [ .[ ·[ [ [ [ c c [ [ [ [ [ .[ ·[j [ [ -, --. -, ., ~ -' --. _j ..... -, .... -, _j Roof projections A continuous obstruction longer than 15 ft may produce a significant drift on a roof. The loads causeO. by such a drift can be considered triangularly distributed on either side of the obstruction with a peak intensity 16 times the clear height of the projection. This value is based on the assumption that the drift reaches a maximum height equal to 80% the clear· height and that the drift snow has a density of 20pcf. This load should be divided by the exposure factor, Ce, since the potential for drifting will also be influenced by site factors. The lateral extent of the drift from a rectangular obstruction can be assumed equal to four times the clear height. L-or U-shaped obstructions (plan view) will increase snow loads to distances of 6 to 8 times the clear height respectively. For a four-sided obstruc- tion such as a perimeter parapet the affected. distance should be taken equal to 10 times the clear height. Example· The following example is included to illustrate the snow load calculation technique. 21 KENAI EXAMPLE Example: Determine snow loads for a large. conventionally insu- lated flat roof warehouse in an open area at Kenai. Also determine loads on the covered loading dock. Kenai Solution Pr = C~ Ce Ctpg For warehouse Pr = (.5)(1.0)(1.0)(84) = 42 psf. For loading dock Pr = (.5)(1.0)(1.4)(84) = 59 psf. 59 psf Depth of snow on dock roof = --= 3 ft. 20 pcf Clear height = 6 ft - 3 ft = 3 ft. Drift .load at wall = (20)(clear height) 20(3) c = -1-= 60 psf. e Total load at wall= 59+ 60 = 119 psf. Lateral extep.t of drift = (2)(clear height) = 2(3) = 6 ft. Design Snow Loads Warehouse 42 psf · Loading dock j_ T Edge of 1 59 psi Warehouse " _j_ 22 [ [ .[ ·[ c [ [ c [ [ [ [ [ [ L . ·l·· A L [ ... . '! 'l 'l ... ""'! -" ., :;lj -, "" " _jj ., j ...1 _.. . REFERENCES l. American National Standards Institute Inc. (1972) Building Code Requirements •· for Minimum Design Loads in Buildings .and Other Structures ANSI A 58.1-1972 • 2. American Standards Association (1955) American Standard Building Code Requirements·f'or Minimum Design Loads in Buildings and Other Structures ASA A 58.1-1Sl55. 3. Associate Committee on the National Building Code (1971) National . Building Code of Canada, 1970. National Research Council of Canada. NRC 11246, Ottawa, Canada. 4. Bilello, Michael ( 1969) Surf' ace Measurements of Snow and Ice for Correlation with Aircraft and Satellite Observations. USACRREL SpecialReport 127. 5. Bilello, Michael (1969) Relationships Between Climate and Regional Variations inSnow-cover Density in North America. USACRREL Research Report 267. 6. Blom, Gunnar (1958) Statistical Estimates and Transformed Beta Variables, John Wiley and Sons, New York. 7. Boyd, D. W. (1961) Maximum Snow Depths .and Snow Loads on Roofs in Canada. Proceedings of' the Western Snow Conference April l96l. Also issued as DBR Research Paper 142 (NRC 6312). 8. Boyd, D. W. ( 1970) Climatic Information for Building Design in Canada; Supplement No. l to the National Building Code of Canada. 9. City of Fairbanks; Alaska (1971) City of Fairbanks Building Code Ordinance No. 2070 • 10. Corps of Engineers (1958) Design Data for Military Construction in. Alaska. Map prepared by the Office of the .District Engineer, Anchorage, Alaska. ll. Departments of the Army and Air Force (1966) Load. Assumption ·for Buildings TM-5-809-l. 12. Department of Commerce, Weather Bureau (ESSA) and National Oceanic · and Atmospheric Administration ·(NOAA) Climatological .Data, Alaska. Asheville, N. C. (monthly records during the past 23 years). 23 13. Department of' the Navy, Naval Facilities-Engineering Command (1967) Des·ign Manual, Cold Regions Engineering NAVFAC DM-9. 14. Grant , Kenneth E. ( 1971) Snow Surveys f'or Alaska. Soil Conser- vation Service, Washington, D. C. 15. Gumbel, E. J. (1954) Statistical Theory of Extreme Values and Some .Practical Applications. Nat-ional Bureau of' Standards, Applied Mathematics Series No. 33. 16. GumbeJ,., E. J. (1958) Statistics of Extremes.< Columbia University Press, New York. 17. International Conference of Building Off'icials (1970) Uniform Building Code, 1970 Edit!on Volume I. 18. Isyumov, Nicholas (1971) An Approach to the Prediction of Snow Loads. Ph.d. thesis submitted to the Faculty of Graduate Studies. The University of' Western Ontario, London, Canada. 19. Schriever, W. R. ; Faucher, Y. ;· and Lutes' D, A. (1967) Sno.w Accumulations in Canada, Case Histories: I. Division of' Building Research Technical Paper 237, National Research Council NRC 9287, Ottawa, Canada. 20. Schriever, W. R•; Lutes, D. A.; and Peter, B. G. W. (1970) "Snow Loads, Commentary No. 2 11 in Canadian Structural Desig:n Manual which is Supplement No. 4 to the National Building Code of Canada. 21. Schriever, W. R. and-Lutes, D. A~ (1971) Snow Accumulations in Canada, Case Histories: . II. National Research Council, NRC 11915, Ottawa, Canada. 22. Sear by, Harold W. (1972) Regional Climatologist, U. S. Dept. of Commerce, NOAA, Anchorage Alaska; Personal communications. 23. Thom, H. C .• -S. (1966) Distribution of Maximum.Annual Water Equivalent of Snow on the Ground. Monthly Weather Review, Vol. 94,-No. 4, April 1966. 24. Thom, H. C, S. ( 1971) Environmental Data Service , ESSA (now NOAA) , WasJtington, D. c. Personal communications. 25. Wahrhaftig, C. (1965) Physiographic Divisions of Alaska. u. S. Geological Survey Professional Paper 482. (Map referred to is-. plate I of this reference. ) 24 [ [ . fj ·[ [ [ [ c c c L [ [ [ .[ ·[ [ [ _. . -, ,, _j -" --, _.3 -, _j _. -, .J _j ...i APPENDIX A: FORTRAN II PROGRAM FOR ANALYSIS OF SNOW DEPTHS C ·LEAST SQUARES ANALYSIS OF LOG•NORHAL PROBABILITY PLOT C INCLUDES CALCULATION OF CORRELATION COEFFICIENT AND ~ COEFFICIENT OF DETERHINATION FOR THE RE~RESSION LINE C USING BLOM 1 S PLOTTING POSITIONS C READ IN LIST OF VARIATES USED IN LAYING OUT PROBABILITY C PAPER AS THE FIRST SET OF DATA. DIMENSION Y!lOOl~ P!10QI.H!100l.XSCALEI100l.V199l.YLOGI100) READ PAPE.R_ TAPE 101. V 101 FORMATIE20.0l C 20 ANUM CHARACTERS PROVIDED FOR DATA LOCATION 4 READ PAPER TAPE. 103. NAl• NA2. NAJ. NA4. NA5• IDUH J03 FORMAT f5A4• 1101 C 8 ANUM CHARACTERS PROVIDED FOR DATA SOURCE READ PAPER TAPE 104. lSOl. ISQ2. IDUM 104 FORMATC2A4•I101 C 8 ANUH CHARACTERS PROVIDED FOR DATA TYPE READ PAPER TAPE 104. ITYPE1• ITYPE2o~. IDUH N=O J N= N+ 1 t Y VALUES READ !N IN ORDER "FRO~ SMALLEST TO LARGEST. C LAST ENTRY MUST BE 9999. C IF THERE IS ANOTHE~ DATA SET• A NUMBER LESS THAN 99999999. C MUST BE ENTERED FOLLOUING "THE 9999. ENTRY DESIGNATING-THE C END OF THE PREVIOUS DATl SET~ IF NO MORE DATA SETS FOLLOU. C 99999999. OR GREATER MUST "BE ENTERED. READ PAPER TAPE 102. Y(N) .102 FORMATCE20.0) IF (Y(Nl•9999.) 1.2.2 2 N&N-1 DO 10 1 ,.1. N PCII=IfFlOATIJl·.375)/(FLOATIN)+.25))•100. MC I )=Pr 11+0.5 XSCALEIJI:VIMII)) 10 YLOGIIJ~FtOGtOFCYCJl) SUHX=O.O SUHY"O.O SUHXSO•O.O SUHYSQ•O.O SUHXY=o.o DO 11 1•1. N SUHX=SUMX+XSCALEII) SUMY=SUMY•YLOGIIJ SUMXSQ•SUHXSO+XSCALEIII•XSCALEIJ) SUMYSO•SUMYSO+YLOG!li•YLOGIJ) Al 0001 0002 0003 0004 coos 0006 0007 0008· 0009 0010 0 011 0012) 0013 0014 0015 0016 0017 0018 0019 0020 0021 0022 0023 0024 0025 0026 0027 0028 0029 0030 0031 0032 003J 0034 OOJS 0036 0037 0038 OOJ9 0040 0041 0042 004J 0044 ll U3 105 106 107 108 J 109 uo 111 112 50 APPENDIX A SUHXY=SUHXY+XSCALEIII•YLOGIJ) ZN•N AVEX:rSUHXJlN AVEY 11 SUHY JZN XYBIG=SUMXY-ZN•AVEX•AVEY XBIG 2 SUMXSO-ZN•AVEX••2. B•XYBIG/XBlG AaAVEY-B•AVEX COCOR=XYBJG/!!SUHXS0-2N•IAVEX••2.Jl•ISUHYSQ-2N•(AVEY••2.)JI 1••0.5 CO DEl =CDC QRu2. Y101•10.••IA+8•2.674J YS•lO.••!A+B•5.842J Y30=10.••<A+B•6.817l Y50•10.••(A+8•7.054l Y100•10.••fA+B•7.J26l Y10=10.••(A+B•6.282l Y25=10.••(A+B•6.750l PRINT 113 FORHATI1Ht,59HlEASl SQUARES ANALYSIS OF LOG NORMAL !PROBABILITY PLOT IBLOH!l PRINT 105, NAl~ NA2, NA3•NA4• NA5 FORMA111H0~9HLOCATION ,2X.5A4) PRINT 106. IS01• 1502 FORHAT11H0,19HlNFORMATION SOURCE ,2X.2A4) PRINT 107. JTYPEl•ITYPr2 FORMAT!lH0•9HDAlA USED//2X,2A4,JX,4H?LOTJ DO J I •1• N PRINT 108, Y!J), H(Jl FORHAT!F8.2.J&l CONTINUE PRINT 109 .. 8• A FORMAl(lH0·6HSLOPE•·FlO.S,2X•10HINTERCEPT•,F10.5) PRINT 110, ITVPE1•ITYP[2•Y10l.YS•Y10•Y25•Y30•YSO•V100 FORHAT!1H0•13HRETURN PERlOD,2X.2A4.2X.8HPLOTTING/4X, 15HYEARS.16X~8HPOSITION//7X.4H1.Q1.5X.F6.2~7X.1Hl/ 17X.tHS-8X.F6.2.6X.2H80/6X.2H10•8X.F6.2•6X.2H90/6X .. 2H25• 18X.F6.2,6X.2H96/6X.2H30.6X,f6.2,6X.4H96.7/6X.2H50,8X. 1F6.2.6X.2H98/5X.JHl00•8X•F6.2.6X.2H09). PRINT 111.COCQR.CODET . FORMAT11H0,24HCORRELAliON COEFFICIENT=•F10.5•5X~29HCOEFF !ICIENT OF DETERMINATION••FlO.Sl READ PAPER TAPE 112• ZoOM FORMATIE20.0l IFlZOON-99999999.1 4.So~50 STOP END A2 [ [ .[ ·U 0045 0046 f' 0047 __ i 0048 0049 0050 [ 0051 0052 0053 0054 c 0055 0056 0057 0058 c 0059 0060 0061 0062 [ 0063 0064 0065 0066 [ 0067 0068 0069 0070 c 0071 0072 0073 0074 [ 0075 0076 0077 0078 [ 0079 0080 0081 0082 [ 0083 0084 0085 0086 .[ 0087 0088 0089 0090 ·[ 0091 [ [ APPENDIX B: COMPUTER PRINT-OUTS FOR CAPE LISBURNE AND UTOPIA CREEK ·-~ . -, _,_, lEAST SQUARES ANALYSIS OF LOG NORHALPROBABILITY PLOT (BLOH) l LOCATION CAPE LISBURNE ALASKA INFORMATION SOURCE BILELLO DATA USED _. DEPTH PLOT l s.oo 6 12.00 16 ~ 15.00 26 15.00 3!) 17.00 45 d 19.00 55 19.00 65 •27.00 74 29.00 84 _.. 29.00 94 ....., SLOPE.= .t&3t6 INTERCEPT., .33194 RETURN PERIOD DEPTH PLOTTING YEARS POSITI:ON 1.0! 6.63 t 5 25.24 80 10 ~0.39 9o 25 37.02. 96 30 38.0& 96.7 -~ 50 42.0& 98 lOO 47.20 99 CORRELATION COEFFICIENT• .96621 __; COEFFICIENT OF DETERMINATION• ~93744 -= ,J --' -' Bl APPENDIX B LEAST SQUARES ANALYSIS OF LOG NORMALPROBABILITY PLOT (BLOMI LOCATION UTOPIA CREEK. ALASKA lNFORHATION SOURCE BILELLO DATA USED DEPTH PL·Ol _9.00 4 13.00 11 14.00 17 19.00 24 20-00 30 20:00 37 21-00 43 27-00 50 28-00 57 32-00 63 36.00 70 42.00 76 4s.oo 83 ss.oo 89 69.00 96 SLOPE= .25555 INTERCEPT., .13644 RETURN PERIOD DEPTH PLOTT INC YEARS POSITION 1.01 6.60 1 5 42.60 80 10 55.18 9o 25 72~68 96 JO 75.60 96.7 50 86.92 96 too 102.00 99 CORRELATION COEffiCIENT• .99498 COEFFICIENT OF DETERMINATION~ .98998 B2 [ --, ~" [ .[ ·C [ ' [ c D [ [ [ [ [ [ [ ·[ [ L