Special Report 89-10 M

US Army Corps

April 1989 of EngineersCold Regions Research &Engineering Laboratory

Compacted-snow runwaysGuidelines for their design and construction in Antarctica

David S. Russell-Head and William F. Budd

0

00

ITI

ELE IIE

JUN 131989D

Prepared forDIVISION OF POLAR PROGRAMSNATIONAL SCIENCE FOUNDATION

Approved for public release; distribution is unlimited.

UNCLASSIFIEDSECURITY CLASSIFICATION OF THIS PAGE

Form ApprovedREPORT DOCUMENTATION PAGE OMB NO. 0704-0188

I Ex. Date: Jun 30, 1986la. REPORT SECURITY CLASSIFICATION 1B. RESTRICTIVE MARKINGS

Unclassified2a. SECURITY CLASSIFICATION AUTHORITY 3. DISTRIBUTION/AVAILABILITY OF REPORT

2b. DECLASSIFICATION/DOWNGRADING SCHEDULE Approved for public release; distribution is unlimited.

4. PERFORMING ORGANIZATION REPORT NUMBER(S) 5. MONITORING ORGANIZATION REPORT NUMBER(S)

Special Report 89-10

60. NAME OF PERFORMING ORGANIZATION 6b. OFFICE SYMBOL 7a. NAME OF MONITORING ORGANIZATION(if applicable) U.S. Army Cold Regions Research

University of Melbourne and Engineering Laboratorytc ADDRESS (City, State, and ZIP Code) 7b. ADDRESS (City. State, and ZIP Code)

72 Lyme RoadMelbourne, Australia Hanover, N.H. 03755-1290

8a. NAME OF FUNDING/SPONSORING 8b. OFFICE SYMBOL 9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER,Qfr[zA iZATIQI (if applicable)Dvision o 'olar Programs 86000 70704National Science Foundation-

8c ADDRESS (City. State. and ZIP Code) 10. SOURCE OF FUNDING NUMBERS

PROGRAM PROJECT TASK WORK UNIT1800 G Street NW ELEMENT NO. NO. NO. ACCESSION NO.Washington, D.C. 20550

11 TITLE (Include Security Classification)

Compacted-Snow Runways: Guidelines for their Design and Construction in Antarctica12. PERSONAL AUTHOR(S)

Russell-Head, David S. and Budd, William F.13a. TYPE OF REPORT 13b. TIME COVERED 14. DATE OF REPORT (Year. Month Day) j5. PAGE COUNT

FROM _ TO April 1989 7016. SUPPLEMENTARY NOTATION

17 COSATI CODES 18. SUBJECT TERMS (Continue on reverse if necessary and identify by block number)

FIELD GROUP SUB-GROUP Antarctica Runway technologyCompacted-snow runways Snow

I Runways19 ABSTRACT (Continue on reverse if necessary and identify by block number)

Only small areas near the margins of the ice cap in Antarctica are ice-free, and only a few of these exposed sites are suitablefor the construction of conventional runways. Wheeled aircraft have operated successfully on hard sea ice and exposedglacial ice, and skis have been fitted to a wide range of aircraft for use on snow. There has been a resurgence of interestin making snow runways suitable for use by conventional wheeled aircraft. Laboratory and field work has confirmed thatlow-density surface snow can be compacted in several ways to yield a strong, uniform, load-bearing pavement that cansupport heavy wheeled aircraft. The Soviets have constructed several full-scale runways in Antarctica. This reportprovides some of the technical background for the design and construction of compacted-snow runways in Antarctica.The technology is not particularly difficult, and it'is likely that the next few decades will see substantial changes toAntarctic air transportation as more snow runways are constructed throughout the continent.

20 . e'.. C,AA;rAB;LT," ., ABSIA i 21. ABSTRACI SECUI, YCLASSIFICATION

63 UNCLASSIFIED/UNLIMITED Q SAME AS RPT. -- DTIC USERS Unclassified220. NAME OF RESPONSIBLE INDIVIDUAL 22b. TELEPHONE (Include Area Code) 22c. OFFICE SYMBOL

Dr. Malcolm Mellor 603-646-4100 CECRL-EEDD FORM 1473, 84 MAR 83 APR edition may be usecl unil exhausted. SECURITY CLASSIFICATION OF THIS PAGE

All other editions ore obsolete.

UNCLASSIFIED

PREFACE

This report was prepared by David S. Russell-Head and William F. Budd of the Universityof Melbourne (Australia). It draws on a range of internal reports and other documents thatwere generated as a result of runway studies by the Australian Antarctic Division at CaseyStation, Antarctica. The work was carried out for CRREL as part of the 1987-88 program ofAntarctic Engineering Services provided to the Division of Polar Programs, National Sci-ence Foundation. Antarctic Engineering Services are provided under a memorandum ofagreement between NSF and CRREL. Funds for this work were covered by the designation8 6000 70 704. The CRREL work is under the direction of Dr. Malcolm Mellor, ExperimentalEngineering Division.

The contents of this report are not to be used for advertising or commercial purposes.Citation of brand names does not constitute an official endorsement or approval of the useof such commercial products.

Aecession For

NTIS GRA&IDTIC ?B UUnaawounoed 0Just iftT tiou

DistributtoD/

Avallablli.y Coldes

Avatl and/ot1Dist Spoolal

LU

CONTENTSPage

P reface .................................................................................................................................. iIntrod uction ........................................................................................................................ 1History of compacted-snow runways in Antarctica ..................................................... 1Mechanical testing of snow ............................................................................................ 2

Wheel loads on pavements ...................................................................................... 2Bearing tests ............................................................................................................... 2Unconfined compression tests ................................................................................. 2C om paction tests ........................................................................................................ 3Rammsonde tests ..................................................................................................... 3C legg im pact tests ..................................................................................................... 3

Mechanical properties of snow ...................................................................................... 3Compacted-snow runway technology .......................................................................... 5

Recent developments ................................................................................................. 5Design, construction and testing of compacted-snow runways ......................... 5

Literature cited .................................................................................................................. 9Appendix A. Casey snow runway tests 1983-84 ...................................................... 11Appendix B. A design and testing manual for the construction of

compressed-snow runway pavements ................................................................... 51Appendix C. A construction manual for compacted-snow runways ..................... 63

ILLUSTRATIONS

Figure1. Map of Antarctica showing distances between stations ................................... 12. Contours of vertical compressive stress beneath a uniformly loaded

circular area on a linear elastic half-space ................................................... 23. Laboratory CBR test, with a sample of the material constrained in a mold.. 24. Unconfined compression tests ............................................................................. 25. Rammsonde penetrometer in use in Antarctica .............................................. 36. Rammsonde hardness and CBR values of field snow .................................... 47. Relationship between CBR and CIV for soils .................................................... 48. Variation of CIV value with snow density if the relationship between

CBR and CIV is the same for snow as it is for soils .................................... 49. Density vs depth at several sites in Antarctica .................................................. 410. CBR strength vs snow density .............................................................................. 511. Self-propelled pneumatic-tired rollers used in the Casey trials .................... 612. Results of laboratory tests that confirm that snow is easier to compact

at higher temperatures .................................................................................... 613. Large compaction roller specially designed for compacting snow runways 614. Sequence of increase in roller loads and tire pressures during construc-

tion ....................................................................................................................... 715. Design curves showing the importance of subgrade CBR as well as pave-

ment CBR on expected settlement ................................................................ 716. Proof roller that mimics the aircraft load and landing gear for monitoring

the adequacy of the pavement ........................................................................ 817. Settlement of a fully-laden full-scale C-130 test plate on the Casey test

pavement, which increased with time but at a diminishing rate .............. 818. Test plate and pavement after 2.5 hours ........................................................... 819. Cross section of the pavement through the test plate zone ........................... 9

iii

Compacted-Snow RunwaysGuidelines for their Design and Construction in Antarctica

DAVID S. RUSSELL-HEAD AND WILLIAM F. BUDD

INTRODUCTION wheeled C-130 aircraft resulted in a trial construc-tion program near Casey in 1983-84. Experiments

A majorimpediment towidespread useof heavy with a snow mill, a motor grader and pneumatic-aircraft in Antarctica has been the lack of suitable tired rollers were used to construct a pavement.landing strips for wheeled aircraft. The annual sea Tests showed that the snow could be compacted toice landing field at McMurdo has been used by the required bearing capacity.wheeled C-130 Hercules and C-141 Starlifter air- This report summarizes the technology involvedcraft. For inland operations, C-130 aircraft equipped in producing compressed snow pavements for usewith large skis have been used with considerable by heavy wheeled aircraft in Antarctica. The proc-success. esses involved are relatively simple, but good

In recent times, Soviet construction crews have knowledge of the local conditions and an under-built runways of compacted snowat Molodezhnaya standing of the snow strengthening process areand Novolazarevskaya for use by heavy wheeled necessary before the technology can be success-aircraft (Fig. 1). Initially, Ilyushin 18D aircraft were fully applied.used, and recently the much heavier Ilyushin 76Thas been successfully introduced into the SovietAntarctic transport system. HISTORY OF COMPACTED-SNOW

Australian interest in constructing a runway for RUNWAYS IN ANTARCTICA

S7-- _ EThe U.S. Navy constructed the firstCape Town recorded compacted-snow runway in

'60W aWEI Antarctica in 1947 on the Ross Ice6WShelf. Subsequent U.S. experience dur-Uhuamw orombo dezhnay ing the 1960s showed that it was pos-

8 . E5/ 7 asible to densify snow by disaggrega-ar on tion with pulvimixers or snow mill-

0 er oh ers, and that higher densities resultfrom using rollers on the disaggre-

-wAmu SC It 90*E- gated snow. Pavements were pro-6Pth duced that supported C-130 aircraft

k ey '(Moser and Sherwood 1967, Abele etOy al. 1968). The successful introduction

of ski-equipped C-130 aircraft toIAntarctica curtailed further U.S. workon compressed-snow runways for

LeorwgrdskE. heavy wheeled aircraft.UC Hoborl 1,

Cly tchurch At Molodezhnaya, Soviet experi-2060 ments with conventional multi-tired

180 1ThE rollers were started in the mid-1960s.During the 1970s, a trial runway pave-

Figure 1. Map of Antarctica ;howing distances between stations (in ment of high-density snow was pro-nautical miles). (After SCAR Bulletin No. 56, May 1977.) duced, and testing showed it capable

of supporting heavy wheeled aircraft (Aver'ya-nov et al. 1975). A runway was completed in U1980 and has been used for regular summerflights by lyushin 18D aircraft. In 1986, anIlyushin 76T aircraft operated from com-pressed-snow runways at Molodezhnaya and -Novolazarevskaya.

In the 1983-84 summer season, a trial con- 3 •---struction of a compressed-snow pavement near."Casey by an Australian team showed that a 5 -. \strong pavement could be made by usingheavy multi-tired rollers to compress snow 3 2 ,disaggregated by a snow miller (Appendix R-di distance ,nradii)A). A full-scale plate bearing test showed that Fg r2 Co uso___the trial pavement would adequately support Figure 2. Contours of verti- o =oMMthe wheel loads of C-130H at maximum mass. cal compressive stress be-

neath a uniformly loaded Figure 3. Laboratory CBRcirculararea on a linearelas- test, with a sample of the ma-

MECHANICAL TESTING OF SNOW tic half-space. This theoreti- terial constrained in a mold.cal stress distribution is a A circular piston of 3 square

Wheel loads on pavements good approximation of ac- inches (19.4 square cm) isWhen a pneumatic tire is loaded onto a tual pavement stresses. driven into the test material.

pavement, the stress distribution produced inthe pavement looks somewhat like a layeredbulb (Fig. 2). The contact area is nearly circular and of small-scale plate test (Fig. 3). The CBR value is ais approximately equal to the wheel load divided measure of the resistance to penetration by theby the tire pressure. piston.

The maximum vertical stress occurs directly Forsnow testing, full-scale plate bearing tests areunder the tire contact area and is approximately a field test only and are expensive to perform. Theequal to the tire pressure. Below a depth of about CBR test is less expensive (because of its small scale)one diameter of the tire contact area, the pressure and is usually done in the laboratory. Field CBRfalls away quite quickly. tests on snow are much more difficult to do than

laboratory tests.Bearing tests

Snow mechanics shares some concepts and test- Unconfineding methods with soil mechanics. Two types of compression testsbearing tests in common use in soil testing have In an unconfined compres-been used for testingsnow. These areplatebearing sion test, a cylindrical sampletests and California Bearing Ratio tests. Both tests of snow is placed between par-load a stiff circular area onto the soil and simulate allel platens of a testing ma-the stress condition shown in Figure 2. chine, and the platens are then

Plate bearing tests are carried out in the field moved together, usually at asituation where stiff circular plates of areas com- constant speed (Fig. 4). Theparable to actual tire contact areas (i.e. 300-600 mode of deformation variesmm in diameter) are loaded to simulate the tire accordingtothesamplegeom-action on the soil. The resulting soil deflections are etry, platen speed and snowmeasured directly. age, temperature and density.

The California Bearing Ratio (CBR) test is astandardized soil test where a flat-faced piston is Figure 4. Unconfined compres-driven into a test sample at a constant speed. For sion tests, which are generallylaboratory testing, a sample of the material is con- performed in a testing machine.strained within a mold, but for field testing, the The platens are usually drivenpiston is driven directly into the surface material. together at constant speed, and aThe load on the piston is recorded as the test load cell indicates the force beingprogresses. The CBR test can be viewed as a type exerted on the snow sample.

2

In some cases, particularly for aged, cold, high- hammer blows) per unit mass of the snow to pro-density snow loaded at high platen speed, the duce a certain density or density increase.snow sample breaks in a brittle fashion. Conversely, Unconfined compression tests, if executed atfor fresh, warm, low-density snow loaded at low slow platen speeds, can result in the test sampleplaten speed, the sample compresses and barrels being compacted rather than broken, and somebut does not break. useful compactibility information may be gained

Unconfined compression tests are not very from the test. A confined type of compression testappropriate for snow pavement design and testing may produce a better simulation of the action of abecause: compaction roller wheel than the unconfined test,

1. The test does not simulate the loading condi- and it also may be more realistic than the hammertion imposed by wheels. impact type of compaction test.

2. The test has not been standardized with re-spect to sample size and platen speed. Rammsonde tests

3. It is not possible to assign an unconfined The Swiss rammsonde is a portable impactcompressive strength value to a snow sample that penetrometer that has had wide use in glaciologi-does not break but merely consolidates. cal studies. It was originally used to gauge the

density of snow at depth and has often been used inCompaction tests stratigraphic studies of the surface snow and firn in

There are no standardized tests for measuring Antarctica (Fig. 5).the compactibility of snow. However, soil mechan- The instrument consists of a 600 conical tip of 40-ics methods can be used to some extent in snow mm maximum diameter mounted on a 19-mm-compaction studies. diameter stainless steel shaft 1 m long. A hollow

Standard compaction hammers, developed for cylindrical hammer slides down a guide to strikethe preparation of soil test samples, can be used to the top of the shaft. Hammer masses of 1, 2 or 3 kggauge the compactibility of snow. A measure of the are normally used. The rammsonde number is acompactibility is the required energy (from the measure of the force required to penetrate the

snow.

Clegg impact testsDuring the 1970s an impact testing device was

developed to provide a rapid test of soil bearingproperties, as an alternative to the CBR test. Acorrelation of Clegg Impact Value (CIV) and CBRhas been given for soils (Clegg 1980).

The Clegg device comprises a 4.5-kg flat-facedimpact hammer, a guide tube and an electronicsystem to record the impact of the hammer onto thetest surface. The simplicity of the device and itsoperation are in contrast to the complexity of thefield CBR apparatus and procedure.

Recently CIV measurements have been made inthe McMurdo area (Lee et al. 1986). As yet, no directcorrelation between CIV and CBR for snow hasbeen determined.

MECHANICAL PROPERTIES OF SNOW

Laboratory and field tests on snow have shownthe overriding importance of snow density on themechanical properties of compressed snow. Figure6 summarizes the relationship between rammsondevalue, CBR and snow density. The effect of tem-

Figure 5. Rammsonde penetrometer in perature is not easily quantifiable. In general theuse in Antarctica. colder the snow, the stronger it is. However, for

3

1000 0

- - - - -- - - 6R-F-

10

0. 05 06 07 as Snow density lg m3

Snow density, fm-3

Figure 8. Variation of CIV value withsnow density if the relationship betweenCBR and CIV is the same for snow as itis for soils.

Density Mg M-3- 3 04. as 06 07 06 069 10

30-

/ 20

-11wdenIsity, i&Vm1 s~o - s.~

Fiue-Rammsonde hardness and CBR 60Ivalues of field snow, which are strongly 70 -j -dependent on snow density.I

P,,. 0917.

Figure 9. Density vs depth at several sites

loo in Antarctica. Typically thedensityof snowincreases with depth at net accumulation

60 -- -- - -sites.

30R -- --

typical of net accumulation sites, where the surface 1. An FOequately strong pavement can be con-snow of low density (approx 0.35 Mg m-3) is gradu- structed at Lanyon Junction on the Law Dome nearally buried and densified. The maximum possible Casey.density is 0.917 Mg m-3, which is the density of 2. Snow compaction is best achieved by pneu-pure, air-free ice. matic-tired rollers.

The load bearing capacity of snow of about 0.6 3. Snow compaction is most efficient when theMg mn-3 density is adequate for an infrequently snow is just moist.used airstrip. It is only the near-surface snow that 4. A very large towed roller should be built forhas low density, and the construction of a pave- constructing and maintaining a full-scale runway.ment for wheeled aircraft essentially involves thestrengthening by densification of the surface layer Design, construction and testingof snow. of compacted-snow runways

The experience gained in the Casey field trialshas been reviewed in the light of pavement design

COMPRESSED-SNOW in general and laboratory work in snow mechanics.RUNWAY TECHNOLOGY Two reports, included here as Appendices B and C,

summarize the outlook at the present stage forRecent developments designing, constructing and testing compacted-

The Soviets have had the most experience in snow pavements.constructing and operating from compressed snow Those two reports, presented in the form ofrunways in Antarctica. Unfortunately very little manuals, should be seen as early steps on the waypublished information has become available, but to developing a comprehensive understanding ofthe current state of their technology seems to be: compressed-snow runway technology. One can be

1. Compressed-snow runways have been con- quite confident that compressed-snow runwaysstructedat Molodezhnaya and Novolazarevskaya. are realistic for many sites in Antarctica and that

2. IL-18D and IL-76T aircraft have operated their construction will become more straightfor-successfully from these runways. ward with the inevitable advances in the technol-

3. A compressed-snow runway is being con- ogy. The major points of the two reports are sum-structed at Vostok. marized below.

4. The preferred construction technique is to use The basic purpose of the construction equip-towed road-working pneumatic-tired rollers to ment is to compact snow to a sufficiently highcompact the snow. density so that it forms a pavement that can with-

5. The most efficient compaction occurs when stand aircraft wheel loads. Figure 10 shows thethe snow is warmest. strong influence of snow density on CBR. Some of

Australian experience is quite limited by com- the construction equipment used in the 1983-84parison to the Soviets'. A trial construction pro- Casey trials is shown in Figure 11.gram was undertaken on the Law Dome near Caseyin the 1983-84 summer. The equipment used in- 100 ,

cluded a D7 Caterpillar tractor, a 14G Caterpillarmotor grader, a Schmidt snow miller, 20-tonne and . /38-tonne self-propelled pneumatic-tired rollers, anda small disk plow. 30 - I

The equipment was tried out in a number ofmodes in an effort to develop a suitable construc- CB. %tion method. A section of test pavement was con- ,0structed, and a full-scale plate load test showedthat the pavement was adequately strong for op- 6eration by fully-laden C-130 aircraft. Details of the .snow testing program have been published as _ I IIseveral Melbourne University Programme in Ant- 04 0 0 07arctic Studies (MUPAS) Reports, one of which is SVW dwt. V 3

included as Appendix A. The general outcome of Figure 10. CBR strength vs snow den-the trials can be summarized: sity.

5

0.8

= .1a5

0.3 -

Piston pressure, t6Pa

Figure11i. Self-pro pelled pneumatic-tired rollers used in Figure 12. Results of laboratory tests that confirmthe Casey trials. The motor grader was fitted with laser thatsnowiseasiertocompactathighertemperatures.leveling sensors.

Figure 13. Large compaction roller specialy designed for compacting snow runways.

The efficiency of the snow compaction depends (Fig. 12). The CBR of snow of this density is abouton tire pressure and snow temperature. The maxi- 35 (Fig. 10).mum pressure for compaction tires is around 1000 Figure 13 shows a suggested form of a speciallykPa (1 MPa). Laboratory tests suggest that a density built compaction roller. The concept is for a largeof about 0.6 Mg n -3 is achievable for tire pressures towed roller that can be dismantled into parts forof 1000 kPa and snow temperatures above -5°C air freighting by C-130 aircraft. The stackable steel

6

'. S,4

Figure 14. Sequence of in-creaseinrollerloadsandtire

- pressures during construc-rte Pressftuf kPa tion.

S&a9dW C8R. VO Arcrff C- H

500 I "

S~oo :

200 0r

,o 2 5 7 10 0 so

7 10Pavement CB. % a. Subgrade CBR = 10.

Subqvo'e CAIQ=, Amrwaft. C-130H$00 a-I 1 I J i fl X. [ I I X I I I I

--!~ l X I I I N % I I IIII7D 1 1 I X

- I ! I\ , \ IR I R tTSSoo

200 ~ r

1 2 5 7 to 20 so "t0 1007avent 10 % b. Subgrade CBR =4.

Figure 15. Design curves showing the importance of subgrade CBR as well aspavement CBR on expected settlement.

dead weights can be handled by a tracked loader Figure 17 shows the results of a full-scale dead-fitted with forks. Figure 14 shows the suggested weight plate test on the Casey test pavement madesequence of increase in load and tire pressure dur- in the 1983--84 summer. The settlement is similar toing the construction program. a creep curve where there is an initial settlement

Figure 15 is from the design manual and shows followed by further settlement at a rate that de-the strong influence of subgrade CBR on required creases with time. The steel test plate had beenpavement thickness. Figure 15a is for a subgrade placed on a grooved surface (Fig. 18), and the initialCBR of 10, which corresponds to the Casey site; settlements shown in Figure 17 are higher thanFigure 15b is for a subgrade CBR of 4, which is an what would be expected if the surface had beenestimate of the situation at the South Pole. Proof smooth. When the aircraft is moving, the loadingrolling is an important aspect of runway construc- time is less than a second and one would expect thetion, testing and maintenance. The design of a C- wheel ruts in the Casey test pavement to be only a130 proof roller shown in Figure 16 uses C-130 rims few millimeters. The test pavement was milledand tires fitted to a specially built chassis that ac- through to see what had happened in the plate testcommodates the steel dead weights of the compac- zone (Fig. 19). No cracking could be seen; the snowtion roller. The wheel base and track of the proof pavement had been simply further compacted.roller is the same as the that for the C-130 aircraft.

7

metws 0 1 2 3

Figure 16. Proof roller that mimics the aircraft load and landing gear for monitoring the adequacy of the pave-,,zent.

20

66 ,

1 owm 0, o01 10

'i. hrs W. hrs

a. Linear plot. b. Logarithmic plot.

Figure 17. Settlement of a fully-laden full-scale C-130 Iltest plate on the Casey test pavement, which increasedwith time but at a diminishing rate. The short-term 7 -settlement is a few millimeters. The settlement in thelong term (more than an hour) is several tens of milli-meters. Figure 18. Test plate and pavement after 2.5 hours.

8

LITERATURE CITED

Abele, G., R. 0. Ramseier and A. F. Wuori (1968)Design criteria for snow runways. USA Cold Re-gions Research and Engineering Laboratory, Tech-nical Report 212.Aver'yanov, V. G., K. A. Bezvinonnyy and V. D.Klokov (1975) Experiment in building a snow air-field for wheeled aircraft. Soviet Antarctic Expedi-tion Bulletin No. 90.Clegg, B. (1980) An impact soil test as alternative toCalifornia Bearing Ratio. Proceedings of the ThirdAustralia-New Zealand Conference on Geomechanics,Wellington, vol. 1, pp. 225-230.Lee, S. M., W. M. Haas and A. F. Wuori (1986)Development of methodology for design of snowroads and airstrips. Institute of Snow Research,Keweenaw Research Center, Michigan Technologi-cal University.Moser, E. H. and G. E. Sherwood (1967) Loadcarrying capacity of depth-processed snow on deepsnowfields. Proceedings, International Conference onLow Temperature Science, Sapporo, Japan, August14-19, 1966, vol.1, part 2, pp. 993-1005.

Figure 19. Cross section of the pavement through the testplate zone.

9

APPENDIX A. CASEY SNOW RUNWAY TESTS, 1983-84*

SUMMARY use in profiling dense snow. The performance of a200-mm-thick test pavement of 0.65 Mg m 3 density

This appendixdescribesthemethods and equip- was assessed by using a steel plate 600 mm inment used to collect data at the proposed site of the diameter loaded to 16.2 Mg, equivalent to a pres-Casey compressed-snow runway during the sure of 562 kPa (82 psi). The average settlement1983-84 summer season, and includes some analy- after a 2.5-hour test period was less than 10 mm.sis of the data obtained. The data used in this The results of the 1983-84 testing programappendix are tabulated in a companion report demonstrate that a pavement strong enough to(Russell-Head et al. 1984). support C-130 aircraft loads can be constructed at

The main purpose of the snow testing work was the Lanyon Junction site with the types of snow-to assess the condition and strength of the in situ processing and road-working equipment used forsnow and the snow processed to form a pavement, the trials.The stratigraphy, densityand particle size distribu-tions for the in situ material,new and aged drift snow,compacted natural snow,processed snow, and com-pacted processed snowwere obtained. Rammsondeand Scala penetrometer

C -- a 'i-

tests on the same types ofsnow were also obtained.Snow strength was assessed ' -- "by California Bearing Ratio - - - -,,(CBR) tests. Seventeen CBR ,*,., -,. \tests were made on snow . ',-, jr~ - - - -" - - - - .xowith densities in the 0.4-0.7 ..... :. - . ' , ,- I OO

M7CAE o ..Mg m -3 range. The average .",,- - ° hCBR of the in situ material " - / \ .. . / \was about 10%. The CBR of ypr I - -~r~ - - -' --- --.the pavement material after k2 .. ' - . Ia few day's hardening was ;, ,f, \ ,l " "-, , - Iabout 40%. f \ I , .

Relationships between .'... ,-- I, .__ /... . ,I% ' I I ' e~ ...CBR, rammsondehardness . \ I I " 0o.. //Eand density have been - .\ -- ,, - / .found for the in situ mate- ,, ,, i \I I I / . Ii. .. \ I II I \ 'rial. They follow the gen- II -"eral pattern of results from K, --7 ... \,...other sites and laboratory 1 - -I K\\,,tests. A newly developed __I. '"small-diameter rammsonde \ ,' \ /needs tobestandardized for ... - I .. ,, ) /

1___________-_____'*VI- - -.. . I I - -

I | Il .... 1o A sT"•Modified slightly from a report ji IIi I ,published by the Melbourne Uni- 0.61,_I ,II I ;, .,, "versify Programme in Antarctic 0 I-- i l__f -'-"Studies (Report No. 64, May 1984) __written by DS. Russell-Head, W.F.Budd and P.J. Moore. Figure Al. Map of Casey and Law Dome area.

11

g Rurnwy site

N1..

a. Lanyon function Station and Lanyon Junction (7 b. Location of runway site (4 February 1984).February 1984).

T

Lanyom Station

c. Test strips and trial pavement (4 February 1984). d. Lan yon Station and runway site (7 February 1984).

Figure A2. Aerial views of runway site.

12

BACKGROUND serve Casey (Russell-Head et al. 1982). The mainreasons for its selection were

The testing program described in this appendix , Lanyon Junction is about the closest site towas part of the Casey runway trials activities at Casey that offers a positive net annual accu-Lanyon Junction during the 1983-84 summer. The mulation of snow;overall objective of the trials program was to assess * The Lanyon Junction site is free of extensivethe practicability of cons tructing a compressed snow melting; andpavement that would support C-130 wheel loads. * The average direction of strong winds isThe main construction equipment included a D7 from the east and Lanyon Junction offers anCaterpillar tractor, a Schmidt snow miller, a 14G east-west runway siting with minimal sideCaterpillar grader, two pneumatic-tired rollers (a slope.Rollpac 21 and the equivalent of a Rollpac 38) and Russell-Head et al. (1982) give estimates of me-a disk plow. teorological and glaciological parameters for Lan-

yon Junction. The mean elevation of the runwaySnow testing program site is about 500 m, and the average daily tempera-

The major aim of the snow testing program was ture in December is estimated to be about -3°C,to define the design parameters of a compressed rising to about -10 C in January. The average dailysnow runway at the Lanyon Junction site. To that temperature variation is about 7*C. The annualend, measurements were to be made of snow den- accumulation averages 0.1 Mg m -2. The surfacesity, rammsonde hardness, California Bearing Ratio slope is about 1.5%, and the horizontal surface(CBR), plate bearing load-deflection behavior, and movement is less than 10 m per year.constant-load creep behavior. Figure Al shows the general Casey and Law

The results from these tests were to be analyzed Dome area. Figure A2 shows the relationship be-for relationships between the snow test parameters tween Lanyon Junction, the station, the test stripsand pavement performance. A correlation between and the runway site. Figure A3 shows the layout ofpavement bearing capacity and some easy method the Lanyon Junction test site in 1983-84.of in situ measurement was of particularinterest. The efficacy of the constructionequipment and the results of proof rolling east DO NOT SCALE FROM THIS SKETCHwere also to be recorded and analyzed, in end

association with the Airfield Project Engi-neer.

As a consequence of the 1983-84 field test-ing program, data collected at the site are 0 500 mtabulated in Russell-Head et al. (1984), andthe methods for snow runway testing aredocumented in Appendix B. This appendixdescribes the methods and equipment usedto collect the data, presents the data in gra- runwayphical form, analyzes the main features of the sitedata, and provides the relationships betweenvarious parameters.

Agencies responsible for other activities atLanyon Junction will have separate reports test stripsthat will be useful additional references. TheAustralian Survey Office provided the sur-vey control, the Airfield Project Engineer was testfrom the Department of Housing and Con- pavement trial base coursestruction, and the Bureau of Meteorologymounted a test flight forecasting program,which included the collection of meteoro- westlogical data at Lanyon Junction. endSaon

o snow pole GW01

Casey runway site 0 drum beaconLanyon Junction had been nominated as

the site for a compressed snow runway to Figure A3. Map of Lanyon Junction area.

13

TESTING ANDRECORDING EQUIPMENT

This section describes the equipment used toobtain the data in the field. The data are discussedin the next section.

Density measurementSamples for density measurement were gath-

.red in a number of ways:* Core cutting by impact hammer;" ico hand corer;* Egon Wherle's electro-mechanical drill;" Hand saw; and* CBR corer.

Two types of balances were used to measure themass of the samples:

* Digital Mettler balance (4.2 kg capacity, 0.1g sensitivity) and

* Ohaus triple beam balance (2.6 kg capacity,I g sensitivity).

The impact hammer for the core cutting had adrop mass of 4.5 kg and a drop height of 87 cm. Thecore cutters were cylindrical steel tubes of 99.8 mminternal diameter, 121.5 mm long with an internalvolume of 950 cm 3 .This system worked reasonably Figure A4. Cold laboratory in freezer van showingwell on medium-density snow but for higher- sieve shaker (left), CBR molds and digital balance.density snow (greater than about 0.6 Mg m-3) thesamples were badly fractured. The Schmidt processed snow was sieved imme-

The other sampling methods worked well, and diately after production. Samples of in situ snowa hand saw was then used to trim the snow into were disaggregated by crushingwith gloved handsvolumes for mass measurement. The mass-sample and rubbing the firn snow against itself to dislodgedimensions were measured with a vernier caliper the individual snow grains.(read to 0.1 mm). The mechanical balance was used The equipment performed quite well, exceptaway from the station, and the digital balance was that the shaker (Endercotts Model EVS1) did notused mostly in the cold laboratory. Both balances work on its intermittent cycle.performed well in the cold conditions.

Errors in the density determination arise mostly Penetrometersfrom the volume measurement (up to 3%). The Two types of penetrometer were used. First, theerror in the mass measurement is less than 0.25%. rammsonde of Swiss origin has a long history ofThe sample itself may not be a true average repre- use in glaciology. The drop masses of thesentation of the material sampled, and this sam- rammsonde are generally 1 kg and 3 kg. The maxi-pling error is probably larger than the error of mum drop height is 50 cm. The conical tip has andensity determination, included angle of 600 and a maximum diameter of

40 mm.Particle size distribution The other penetrometer is of soils testing origin

The particle size distribution of disaggregated and is available commercially as the "Scala pen-snow was obtained by sieving a sample through a etrometer." This instrument is more robust thanset of 10 wire-mesh sieves with openings ranging the light-weightrammsonde. Thedrop mass is9 kgfrom 0.075 to 6.7 mm. Snow samples were placed in and the maximum drop height is 50.8 cm. Differentthe top of the sieve set in a mechanical shaker in the tips can be fitted to the threaded end of the first rod.cold laboratory (Fig. A4) and shaken for 10-15 min- Three tips were specially made from hard stainlessutes. The masses retained by each sieve were meas- steel with the same conical geometry as theured with the digital Mettler balance (0.1 g sensitiv- rammsonde. The tip diameters were 20 mm, 40 mmity). (same a, rammsonde) and 80 mm. The purpose of

14

these tips was to test hard snow (with the 20-mmtip) and low-density snow (with the 40-mm tip).Drop penetrometers are very simple instruments,and there were no problems in their use in the field.

California Bearing Ratio testingBoth field and laboratory tests were performed.

The field test system comprised a motorized jackfitted to a cantilever beam attached to the rear of aD5 Caterpillar tractor. A load cell was placed be-tween the CBR piston and the jack, which allowedthe electronic recording of the piston load withtime.

This field CBR test system worked well me-chanically and electronically, but the test was in-validated by the slow melting of the piston into thetest snow surface. The problem could not be easilyovercome, and only two field tests were performed.The remaining 15 tests were performed on cored insitu samples which were placed in a CBR test Figure A6. CBR load cell instrumentation for loggingmachine (ELE Model EL29-001) in the cold labora- piston load with time. This equipment was housed intory (Fig. A5). the site office.

The load recording system consisted of a 50-kNload cell (Interface Model 1210-BF), a program-mable digital voltmeter (Hewlett Packard Model recorder (HP 82161A). Only the load cell was in the3468B), a calculator-controller (HP 41CV), a cold laboratory; the rest of the data acquisitionprinter-plotter (HP 82143A) and a digital cassette system was in the site office (Fig. A6).

The piston speed was 1 mm per minute, and theload was measured every 6 seconds, i.e. every 0.1mm of piston travel. The load was displayed ingraphical form by the printer-plotter during the15-minute test period, and the actual load values ateach 0.1-mm interval were recorded at the comple-tion of the test as digital values by the cassetterecorder.

The gearbox in the test machine needed to bewarmed before the drive system would run. Apartfrom that problem, the laboratory CBR system

-: worked very well.

Plate bearing testA circular plate, 600 mm in diameter, was made

from 10-mm-thick mild steel reinforced with sixradial ribs, 10 x 50 mm. The plate was loaded withthe smaller roller (Rollpac 21 weighing 16.2 Mg) viaa hydraulic jack (Fig. A7). The plate displacementwas measured with a resistance displacementtransducer (Pye-Ether PD20). This sensor wasmounted on a 4-m-long steel beam, weighed downat each end on blocks of polystyrene resting on thepavement. Because the plate was under the rollerand always in the shade, there was no melting of

Figure A5. California bearing ratio testing machine the pavement snow by the plate during the test.in the cold laboratory. The piston load is measure with The data recording system was the same as thatthe load cell. used for the CBR tests. The plate movement was

15

Figure A8. Autonatic data acquisition and recordingequipment in the station site office.

tions, but this problem was solved by groundingappropriate sensor terminals. Some data were alsolost due to 240-V power loss. Apart from these twoproblems ancillary to the system, the data record-ing equipment functioned without fault.

The positions of the various sensors are indi-Figure A7. Unladen Rollpac 21 roller balanced as a dead cated in Figure A9. The data acquisition systemzeight on a hydraulic jack on the plate. The settlement shown in Figure A8 was housed in the site office;logging equipmnent is in front of the roller, the layout of the snow and meteorological sensors

is shown in Figure A10.displayed and recorded every 15 seconds. The onlydifficulty in setting up the system was the place- Air temperaturement of the hydraulic jack to balance the roller. The air temperature was monitored in a small

screen clamped to a snow pole so that the sensorMeteorological and was 0.5 m above the snow surface. The position ofglaciological recording the screen was changed during the summer period.

An automatic data acquisition system was used From 19 to 26 January the screen was positioned toto record air temperature, snow temperatures, the south of the site office, and from 26 January toincident radiation, wind speed and wind direction 8 February the screen was to the east of the buildingat 15-minute intervals. The various sensors were line (Fig. A9 and A10).connected tu a Hewlett-Packard 3421A data acqui- The sensor was a platinum element resistorsition and control unit, and this was in turn con- (DegussaTypeW60/1).Itsresistanceina freezing-trolled byan HP85B computer via the HP-IB (IEEE- point bath was measured before installation in the488) bus. A specially written program for the HP screen (Fig. All), and the computer program per-85B initiated a sensor scan every 15 minutes, con- formed the conversion from resistance to tempera-verted the sensor values to physical units, and ture by solving the following quadratic equation:formatted and recorded the data on cassette tape inthe HP 85B. The data were also transmitted every Rt = R0 (At

2 + Bt + 1) (Al)15 minutes via the GP-IB to an Epson MX-80 printer(Fig. A8). where R, = resistance at t°C (ohms)

The data acquisition and recording system R° = resistance at 0°C (ohms)worked well after software debugging. There were t = sensor temperature (0C)some difficulties with the 3421A due to static charge A = -0.57841 x 10-build-upon the sensor lines duringblizzard condi- B = 0.39078 x 102

16

air and snow upwind of the station after it became clear0 temperature sensors

(26 Jan - 8 Feb) that the snow around the trench had beencontaminated with ash and soot from a fire(used to bum packing timber). Excessive

wind vane anemometer melting had occurred in the snow due to the-I !-increased absorption of radiationby the dark0 5 m soot. The meltwater had probably dissolved

some of the ash, and this would have de-me s s seeping pressed its refreezing temperature. Thesnow

temperature data need to be interpretedwith these effects in mind.

After 26 January the sensors were attachedkitchen sleeping to a cane rod with white insulation tape (Fig.

A12). The assembly was fitted into a holebored with a Pico hand corer and then back-filled. The sensors were placed at 0.1, 0.2,0.5

ablutions sleeping and 1.0 m below the snow surface. After aday or so the sensors should have reachedequilibrium with the ambient snow, and thetemperatures recorded in the new arrange-

store site air and snow ment should accurately reflect the generaltemperature temperature regime in the surface snow.sensors(19-26 Jan)

radiometer Solar radiationstore The incident solar radiation was measuredF_ with a Kipp and Zohnen type of radiometergf (Fig. A13). The sensor was mounted on thegenerator Thesensrfreezouter vanh

outer comer of the station store (Fig. A9).The voltage output was measured with the

high-impedance (1010 ohms) voltmeterwithin the HP 3421A. The calibration of this

Figure A9. Plan of the Lanyon Station area showing the positionsof meteorological and snow temperature sensors.

As the screen is not a perfect radiation reflector,its internal temperature was probably higher thanthe ambient air temperature during calm days.Also on near-calm days when the screen wasdownwind of the station, higher than true ambienttemperatures were experienced within the screen.

This effect was particularly noticeable when thescreen . is to the east of the site office. On warm,calm days, the screen could be bathed in warm airfrom the heated station structure. When there wasa slight west-southwest breeze, the generator ex-haust (which was discharged to the rear and below Ithe buildings) could drift over the screen. These °factors should be borne in mind when attempting I-pto interpret the air temperature record.

Snow temperature - -..

Platinum element resistors (Degussa Type W /60/1) were used to monitor in situ snow tempera-tures. The sensors were placed in the wall of abackfilled trench to the south of the site office from Figure A10. Layout of air and snow temperature sen-19 to 26 January. These were moved to a new site sors, cup anemometer and wind vane.

17

AAI.

Figure All. Platinum-resistance temperature sensor Figure A12. Snow temperature sensors taped to a canehoused in the small sc reent together with a Inercun/-in- rod before insertion in a hole drilled by a hand corer (28glass thermometer. January 1984).

sensor is not known. The voltage output is a linearfunction of radiation absorbed by the radiometer.A horizontal screen is normally fitted to this type ofinstrument, and some errors may occur with theinstallation as shown in Figure A8, especially whenthe sun is in a subhorizontal position.

The purpose of including this instrument in the.,sensor set was to provide some record of the sun-

niness and cloudiness of each day, rather than to- make accurate quantitative records of the net inci-

dent radiation. The millivolt readings should beS..viewed then as an indicative record of the general

A .,-cloud state during the period.An absolute calibration scale can be estimated

from previous detailed radiation studies in Antarc-tic latitudes (e.g. Weller 1967). Peak (noon) clear-day global radiation during January can be ex-pected to be about 0.84 kW m -2 in the Casey-LawDome region (1 cal cm- 2 min -' = 697.8 W m-2).

Wind speed

A cup anemometer (Rimco Type R/CGA) wasused to measure wind speed. It was mounted on apole at the leading edge of the building line (Fig.

Figure A13. Incident radiation sensor mounted on top A10). The height above the snow surface was aboutof station store module. 6 m.

18

The cup shaft was coupled to a small DC genera- resistance of one section of the potentiometer wastor, and its open circuit output voltage was sensed. measured with the HP 3421A and converted to aThe conversion to a wind speed value was incorpo- true bearing with the formularated in the computer program:

D = R/1.26 (A3)S = 137.5 V + 0.8 (A2)

where D is the true bearing (0) and R is the resis-where S is the wind speed (knots) and V is the tance (ohms)output voltage (V). There is a segment at the ends of the circular

Equation A2 is derived from the manufacturer's potentiometer where the wiper is out of contact.data and implies that winds below 0.8 knots will not The output for this section is variable and is usuallyrotate the cup shaft. Therefore, wind speeds below denoted in the record by a period or a negativethis stall speed will still be recorded as 0.8 knots. value. The orientation of the vane for these outputsThe instrument appeared to have more rotational is between 355 and 0' true.friction and noisier bearings after the recording The wind speed and direction values are essen-period than at the start, so the stall speed may have tially spot values, not averages. The wind speedsprogressively increased during its period of use. recorded are not necessarily maxima.

Wind direction TEST DATAA wind vane was mounted on a pole at the

leading edge of the building line (Fig. A10), and its This section presents the data collected at theheight above the snow surface was about 6 m. The Lanyon Station site, mostly in graphical form. Thevane shaft was connected to a rotating potentiome- original data are tabulated in Russell-Head et al.ter whose zero position was oriented due north. The (1984).

V r

a. Three-meter-deep trench dug by the Schmidt snow b. Sideof trench showing ice lenses alternating with firnmiller. snow. The ice lenses are typically a couple of centimeters

thick with 5-10 cn of snow between lenses.

Figure A14. Snow test pit near Lanyon Station.

19

Overview the station (Fig. A15). Figure A16 shows the iceThe data are presented in the sequence of snow layering in the cores and also the densities of a

structure, snow strength and pavement proof test. number of sections from the cores. Clear bubbly iceThe records of meteorological and snow tempera- alternated 'with firn snow throughout the cores,ture conditions are also discussed. and the transition between the two forms of ice was

sharp.In situ conditions The density of the ice was about 0.8 Mg m-3 and

The structure and strength of the in situ snow the density of the snow between the lenses aver-has an important influence on the thickness of aged about 0.43 Mg m-3.The total thickness of thepavement required to support a given wheel load. ice lensing in the first core of 1500 mm was aboutMost of the data collected on site have arisen from 235 mm, and in the second core of 2500 mm, 293the investigations of the in situ snow. mm was ice. The average ice lens thickness was

An unexpected snow structure was revealed about 16 mm.when trenches were dug into the snow. Earlier in- For an assumed dichotomy of snow and ice, thevestigations (Cameron et al. 1959) were either at average density of the material can be calculatedhigher altitudes (where there was a single annual from the proportions of ice to snow. The averagelayer of refrozen meltwater) or at lower altitudes proportion (by volume) of ice is about 16%, which(where there was essentially blue ice). At the Lan- gives an average density of about 0.48 Mg m-3.Theyon Junction site, there were multiple annual melt average annual accumulation at the siteis about0.1zones alternating with firn snow (Fig. A14). This Mg m-2' and this corresponds to an annual burialstructure of the subgrade snow has a much higher rate of 0.21 m.strength than the typical 0.35-Mg m -3 snow thathad been conservatively expected. The snow-ice Particle size distributionsstructure provides an excellent source of primary Samples of snow from core 2 were disaggre-material for processing by the snow miller. gated and sieved. The cumulative particle size

Snow processing and pavement production DO NOT SCALE FROM THIS SKETCHThe strength of snow depends strongly on its

density. The main purpose of working the in situmaterial was to increase the density. Methods weredeveloped with the construction equipment to 500mproduce a densified snow pavement for snow thatwas either cold or moist.

For moist snow the simplest way of starting thepavement production was to drag a heavily ladenplate over the moist snow. In cold snow the dragmethod of compaction did not work, and the bestmethod appeared to be track rolling, i.e. repeatedpasses of laden tracked vehicles. The track-rolledsnow needed some time to age-harden. The basecourse produced by either method allowed thecompacting and grading equipment to operate with rolled

processeda significantly reduced chance of bogging. snow

The pavement snow is produced by disaggre-nedrfnew drift/ track rolled

gating the in situ snow and ice with the snow snow cold snowmiller. The processed snow has an extended range processed rolled moist snowof particle sizes and is denser than the surface snow drag compactedsnow. The grader levels the processed snow, which mois t snowis then compressed to a higher density by the •pneumatic-tired rollers.

Cores 1 & 2

Snow stratigraphyTwo cores for stratigraphy analysis were ob- Figure A15. Map of test area showing positions of

tained with Egon Wherle's electromechanical drill samples taken for snow stratigraphy, density andat a site some 15 m from the northwestern corner of particle size distribution.

20

ITV

7tTT 0.5~Li~i .0

- Depth, mm

1.50.4 0.6 0.8

Density, Mq m- 3 a. Core 1.

I r 1 0 ..- -

. .. I

D, t 1.

---_ .. ._... ... . .... !..

4 . F

:ho

I_

1 -' " ' - Core2.

Figure A16. Ice layering and snow densities at Lan yon Station site.

21

100

80 __

% passed~~___

by mass

20 -

0

0.1 1.0 10

Sieve opening, mm

a. 720 mm below the surface.

100 - -

80

passed - ---. 1Iby mass

40 __

20 " .

00.1 1.0 10

Sieve opening, mm.

b. 800 mm below the surface.

Figure Al 7. Particle size distributions of disaggregated snow.

22

100 ____ - - --

80- - - -- -

60 -i -%passedby mass40 -

20 --.-

0 -0.1 1.0 10

Sieve opening, mm

c. 1100 nim below the surface.

IGO- - -

80 -- -

60 -

%passedby mass

40 - _ _ -

20

0 --0.1 1.0 10

Sieve opening, mm

d. 1600 mm below the surface.

Figure Al 7 (con t'd).

23

100

80---

60

Spassed

by mass 40

40

20 -- __ _

0.1 1.0 10

Sieve opening, mm

e. 1800 mm below the surface.

100 - - ---

80

60

%passedby mass

40-

20- -

0 - -0.11 1.0 10

Sieve opening, mm

2200 mm below the surface.

Figure Al 7 (cont'd). Particle size distributions of disaggregated snow.

24

100 Table Al. Typical densities ofvarious types of snow at Lanyon

Tye. of strou De,,sity (Mg m,')

% passed New drift snow 0.42by mass Snow between ice lenses 0.45

Ice lenses 0.8

40 I I Average in situ snow and ice 0.48

20 Track-rolled cold snow 0.6Drag-compacted moist snow 0.6Rolled moist snow 0.7Fresh processed snow (29 Jan) 0.63Fresh processed snow (7 Feb) 0.59

0 - Rolled processed snow 0.650.1 1.0 10

Sieve opening, mm

g. 2500 nun below the surface. distributions are shown in FigureA17. Distributions for all thesamples

Figure Al 7 (cont'd). Particle size distributions of disaggregated snow. are shown on the plot in Figure A18.The average particle size appears tocycle with depth, within the 0.7- to1.4-mm range.

I I The particle size distribution ofnew drift snow shows a much

/ smaller mean grain size of aboutimi1H{ i 0.47 mm (Fig. A19). The range of

grainsizesisquitenarrow.Thesnow

0.7 m I i , processed by the Schmidt snow4-1. -- 0.8 m' I! millerexhibitedawiderangeof grain60 1 .1 m _--T ---1.6 m//16sizes from 0.3 to greater than 10 mm

1.passd I 1 8 m (Fig. A20). The mean grain size wasby ra s 2.2m

2.5m I about 1.5 mm.4 ' '" 2.5 m40 I iFI Snow densities

I 1/ i Table Al lists typical densities of20i / I in situ and processed snow at the

20 Lanyon Junction site.

* H Penetrometer dataI it The sites of the penetrometer tests

0 are shown in Figure A15. The pro-0.1 1.0 10 files are shown in Figure A21. The

Sieve opening, mm penetrometer numbers in the fig-ures are from the tables in Russell-

Figure A18. Particle size distribution of all samples of disaggregated snow Head (1984). They had been calcu-from Core 2. lated with the formula

25

100 - - ,1

80 __LL- ~Ii

60

%passed I- -by mass 40~

40 .

I I

20__

I -

0,, IiLL i0.1 1.0 10

Sieve opening, mm

Figure A19. Particle size distribution of new drift snow.

100

8 0 . .. . ..

passed

I

by mass I

20

0 /0.1 1.0 10

Sieve opening, mm

a. 29 January 1984.

Figure A20. Particle size distribution of Schmidt processed snow.

26

80

60 '

%passedby mass

40 __ -- - [f.20

b. 7 February 1984.

0 -2 .Figure A20 (cont'd). Particle size0.1 1 .0 10 distribution of Schmidt processedSieve opening, mm snow.

40 mm tip diameter 3 kg drop mass 50 cm drop height0

0.25 -- -- -

0. 5

Depth, m j II

0.75 ___

1.0 -- --

a. Western end of runway site; 400-mm tip diameter; 3-kg drop mass;50-cm drop height; 25 January

1.251 11984.10 100 1000

Rammsonde number, kg Figure A21. Rammsonde profiles.

27

2 I L _

ri r'

SItso ________ --. --.-. _,,___ 4

I lL

I I I

a a)

28

C) It

o.a.

0

rl H t

E L

E---- 4 w - C-- C1; C; (NC

C)Its

__ _ _ _ _

IJIM

o __________29

-,L

ul0

00O1l __I_____I

L

IL ~ Ch

I- _ 30

(u z DM w E 01-. ) E a)..

V , 1 0 0 .4J E a -- E 00o. 0 CL 0 o

K1 4j CL 0 2 Q 4, d f 00'AiM- -- I -14---42......

AC_ 0__ w_ _ -2 ' w 0~ w E -M

41 , w' 0 J M 0J0 2oWo

E 0 .U 0. Ea' o x

. .......-- -- - - . - -2- .J'4 ,.O1V J'4

rz 0 Q 0t

0 .r34 2

w~ (2 0. 1 0 0 OU) 04J

-) 04 00 LO14 aE V0 ar

4__ 4J 2j0r CL 0 CL 0CL- '1 m

W C42 4 -WwMIO o

0'V

2.54

o o 0

4,4

a.1

200 has been a threefold increase within 24hours of placing the disaggregated snow.

CBR resultsThe CBR load-penetration curves for

the 17 tests are shown in Figure A24150. (Figure A25 shows the locations). The

S/divisions on the load axis are 1 kN for allthe graphs. Apart from CBR tests 1 and

Ramsonde 2 (which were invalidated due to melt-number,kg j I' ing of the test snow by the piston), the

maximum piston loads during the 15-100 minute test period ranged from about

100 _L 3.2 kN (Test 11 on in situ snow) to about11.2 kN (Test 15 on icy in situ material).

Table A2 lists the CBR values of eachtest. The density of the CBR samples,especially of in situ samples, was notuniform. There was a marked difference

50' between the CBR of the icy end of aIsample and the snowy end. Table A3shows the average of the CBRs for bothends of the sample together with theaverage density of the sample. The val-ues in Table A3 are plotted on a log scale

T 6 1in Figure A26.0 6 12 18 24

Time, hrs Plate bearing and creepThe plate test was performed on a

Figure A23. Increase with time in rammsonde hardness of processed section of test pavement 200 mm thicksnow. (see Figure A25 for location). Table A4

R = Wh (n/l) + P (A4) Table A2. CBRvalues of in situand pavement snow of differ-

where R = rammsonde hardness (kg) ent densities.W = hammer mass (kg) Test CBR M Density (Mg r- 3)h = hammer drop (cm)I = penetration (cm) for n blows 1 0.7

P = total penetrometer mass (kg). 2 0.73 16.5 0.5

The correlation between penetrometers is not 4 15.3 0.55* 29.9 0.63good using eq A4 (Fig. A22). It is improved if allow- 6 23.9 0.56

ances are made for an energy loss during the im- 7* 49.9 0.70pact of the hammer. The loss depends on the rela- 8 15.9 0.50tive masses of the hammer and the penetrometer. 9 11.1 0.50Another penetrometer number equation is devel- 10 29.5 >0.58

11 12.6

WE W Wc144~

caE

020

p1 00~

-,

*%I-.

0 N-

33-

.14 In

U) 0 C

0 0.

141Um

WVW

C4-

0E 0~

. X ........ g .- ~U*31 a 0 )0 zl

C

1 42z

00

34

n C )

o0,n m 41

u)

0 "

e 0

01

C .4 0

____ _ __ __ _______ _ __ __

a.I 0~

4 I j35

11E 0Occ". V 0.

0 ,) VO .000

S... . " . 0c

I 4

00

.1 I 41 .

%..

0 0

0.0

-41

0

U) 00

VE 4.4 41%00-

00

to "

361

0.,': :'.. g3 .co r,-0. .

2 *; 4 0~ 0

-:.,. 0

• - o

2.n

a, 0

E 0. E

'I 0 tU)tVC .~ l.JJ0

• , 3..

.. % I

0

0

37

mv J0W0 1

.... ,

" .

*-r - ----- L

91. 0.

0E Oro

Li0 m 4.E1

enn

0z ,-..

4

38

c'4 41 0~ m

~ 4JO 4J

00W

0W c E4~

AI 041 J

0 Ai

C;d .3.0

tn Aj ~ a

0 -

41

0. 0.

391 0~

f fa

00W

cooto04 (A

UU .14

In

C-)

14 t%

-41

$4-

41U

I-I-AtA ~ .

tI ' EU E0.4

'o4 .WVL.

I i ~. -r4-- Ii j - coto

I0i-I-

- ~ ~~40 M0 L

E

C-CD

E- C>4

rnA

0 GN)4

Lz

U) 000

I I 0 m D

S.4 0

..... .. O 0'

I I 4-o

I -- r

41-

100 --__--_.-. Table A4. Comparison of plate and C-

130 tire footprint.

-. --- Plate C?-130

Load 16.2 Mg 17.6 Mg (one wheel)Pressure 562 kPa 586 kPa (85 psi)Diameter 600 nun 612 mm

shows the compansonbetween theplateand the aircraft tire parameters.

CBR, The plate settlement over the three-hour test period is shown in Figure

10 A27. The same data are plotted on log-10 - -- log scales in Figure A28. Extrapolation__............_ [to very short and very long times is

... easier on the log-log plot. The validity-- I- - of the extrapolations needs to be tested

... ...... - -- I--.-by proof-roller loading.CBR = 574 95.4 The total settlement after 2.5 hours

was about 13.8 mm. The surface on- which the plate was bearing was not

flat. The surface undulations were about10 mm, and therefore the settlement re-corded is probably greater than if the

* pavement surface had been flat. The settlement ofo natural the plate after a time equal to the load-

material ing time of a C-130 taxiing at 0.5 kph(0.001 h) is on the order of 2 mm. The

0.4 0.Snow dens.ty,6 0. 0.8 settlement after standing for 10 hoursis

about 20 mm. These estimated valuesFigure A26. Logarithmic plot of CBR of natural snow vs density. The are for single passes of a fully ladenequation is a least-squares geometric fit. Also shown is the CBR of the aircraft.test pavement, which had not fully age-hardened. Although further testing needs to be

performed on full-scale pavement (par-ticularly by proof rolling), the pave-ment as tested here shows adequate

strength for a runway pavement. It also appearsTable A3. Averaged CBR values possible to construct a pavement sufficiently strongfor each sample of in situ snow to serve as a hard-standing area, without recourseand pavement snow. The den- to reinforcement.sity is the average sample den-sity. Meteorological and

glaciological recordsTests CBR (%) Density (Mg m-3) The data record for air temperature, snow tem-

perature, solar radiation, wind speed and wind3,4 15.9 0.697 direction is tabulated in Russell-Head et al. (1984).5,8" 39.9 0.51 The records are in computer files. Graphical rec-8,9 13.5 0.50

10,11 21.1 0.577 ords for the period need to be produced from the12,13 40.1 o.597 filed data. Samples of a daily set of graphs are14,15 41.8 0.621 shown in Figure A29.16,17 23.8 0.548 Once the graphical presentations of the com-

pavement plete data set are available, an assessment of theappropriate type of analysis can then be made (e.g.

42

15 Comparison of penetrometersThe scatter in Figure A22 shows

the problem of determining the aver-age resistance of the snow under test,

- %when penetrometers of differing con-___ ___figuration are used. The variations lie

_ -- ___ ....... in the differing drop and penetrome-ter masses (for the 40-mm tip size)

I and, more obviously, the penetrome-ter tip size.

10 .. .. .

___ _ .... _ ! .... __ - _ Drop mass and penetrometer massThe normally used rammsonde

Plat I equation (eq A4) assumes that all themovement, / potential energy of the falling ham-

mm mer is converted to plastic deforma-tion in the test snow. A number of

:____ _!_ workers have criticized the limita-/ I tions of eq A4 and have offered rela-

___ - tionships that fit their data better(Niedringhaus 1965, Waterhouse1966). The authors of this report sug-gest that the Hiley pile-driving for-

._.... _ . . mula as offered by Waterhouse (1966)is the best available, but perhaps it

___ .... - i_.* also should be slightly modified.i The Hiley (1925) formula allows

for an energy dissipation during the1impact of the drop hammer on the0_ penetrometer. The argument follows

0 1 Time, hrs 2 3 a set of Newtonian ideas about im-Time, hrs asto etna da bu mpacting bodies, where the loss at

Figure A27. Settlement into the test pavement of a 600-mm-diameter impact is accounted for by a coeffi-plate loaded to 16.2 Mg. The plate was unloaded after about 2.5 hours. cient of restitution, e. For perfectly

elastic collisions, e is 1; where there isnumber of days of significant melt, number of days no rebound after impact, e is 0.with wind speed above 15 knots). The analysis will The Hiley formula as applied to the penetrome-be of use to flight forecasters as well as construction ter isplanners.

R = (Wh/s) (W + eP) / (W + P) (AS)

ANALYSIS OF DATA where W = hammer mass (kg)h = hammer drop (cm)

This section brings together the results of, the s = "set" (cm per blow)various tests performed on the in situ and proc- e = coefficient of restitutionessed snow. Relationships between the penetrome- P = penetrometer mass (excludes ham-ter values, CBR and snow density are of prime mer).importance. A set of three equations coupling thethree parameters is offered. In the original rammsonde equation (A4), there

The results of the full-scale plate test are matched is a term covering the effect of the gravity-inducedagainst values predicted by the settlement analysis force of the penetrometer acting at the tip, in addi-of Russell-Head et al. (1982). The requirements of a tion to the impact force. This term is significant forrunway and a hard-standing area are then summa- low-density snow and should be included in equa-rized. tion (AS):

43

100 giving the same information as the,Jif . .. ,-.-. rammsonde (Fig. A21d and e).

The penetrometer number, as... given by p netrometers with tips of

.. 7 .the same geometry but different di-. -- _ ameters, appears to be a function of

the cross-sectional area of the tip.......... Equation A6 can be modified again

I by scaling according to the ratio of, :the tip area to the standard

S.rammsonde tip area:Plate .

movement, II , "h., ",",, R = (d/40 [(Wh/s)(W+eP)/

(W+P)+(W+P)] (A7)10 "

. .where d is the diameter of the non-.I,. . -standard tip(mm).

The return angle from the maxi-. .mum diameter of the tip to the shaft

may also be important, and unfortu-, ,natelytheangleofthethreetipstested

on site were different. The half-anglei ,of the standard rammsonde is about

9.4*; for the small tip it was 2.5* andI' for the large tip it was 200.

The higher-than-expected (by eqI , , ,A7) values of the small tip may be

ii~il[Ijjjlhiiijjjj due to the small reardcearance, and1 Llikewise the lower-than-expected0.001 0.01 0.1 1.0 10

Time, hrs values of the large tip may be due tothe generous rear clearance. More

Figure A28. Plate bearing data plotted on logarithmic scales. work should be done to clarify the tipsize effect by using tips with the stan-dard rammsonde clearance angle.

R = (Whs) (W+ e2 P) / (W+ P) + (W+ P) (A6)Relationship between penetrometer

As the value fore is about0.5 for steel/steel (and number and snow densityis reduced if the impacting surfaces are not clean) Table AS lists standard rammsonde values withthe rammsonde value as given by eq A6 may be as snow densities selected from Figure A21. There arelow as 60% of that by eq A4 when the hammer and not many situations where the snow density ispenetrometer masses are equal. The spread of R in sufficiently uniform for a rammsonde number toFigure A23 for penetrometers of different masses be ascribed to a density.and hammer weights with the standard 40-mm- The lowest R value isdiameter tip is reduced when R is calculated with for the in situ case where Table AS. Values ofeq A6. the rammsonde is with- standard rammsonde

in ice lenses. There are hardness for snow ofPenetrometer tip size two intermediate den- uniform density.

Two tips, 20 mm and 80 mm in diameter, were sity situations that cantested against the standard 40-mm rammsonde tip. be used for the correla- R (k) r (M')The 80-mm tip was tested for its usefulness in low- tion. One is the R of thedensity snow and was found to have no advantage aged drift snow, which 5025 0.7over the standard rammsonde (Fig. A21a and b). had a fairly uniform 88 0.5

77.5 0.48The 20-mm tip was tested in high-density (pave- density. The other is to 19.3 0.42ment) snow, and it offers easier penetration while average a complete pro-

44

00~

-1*

(D LnfC

U, -- + 0 -+ U U + U-I 0 ~I0 ~I0

a CL 30 C- ci (I L0. 0OE 02 OE 0; O 2 0

W - fE- ( CE4 -) E

ID''0D

In-

Lco

('4

N 0)

45

10000 -where r is the snow density (Mg m-3)._-_ The inverse of eq A8 is

r = 0.327 R 09° (A9)

-CBR of in situ materialThe CBR values in Table A3 are

-plotted on a log scale against densityin Figure A26. The points for the in

_000 Isitu material fall about a straight line,00C implying a power-law relationship.

_ _ A geometric regression gives theequation

R, kg CBR = 574 rW4 (A10)

_ _ where CBR is in%.

R 2 .46 x 1 Equation AlO has the inverse:

100 __ __ r = 0.308 CBR'-" .(All)

The ice lensing within the in situ_____ material gives rise to large CBR val-

ues, but the ice content averages about16% of the in situ material. The aver-

age density is about 0.48 Mg m-3, and_ _ _this corresponds to an average CBR

of 10.9%.,7_ The lowest density of the in situ10 snow is about 0.42 Mg m -3, and ac-0.4 0.5 0.6 r .7 0.8 cording to eq A10 has a CBR of aboutSnow density, Mg m-3 5.3%. The layered structure is suffi-

Figure A30. Relationship between rammnsonde hardness and snow den- ciently consistent to allow the use ofFityfragued so. Ren the higher (average) CBR of 10.9% insity for aged snow. estimating the pavement perform-

ance, provided that the covering pave-file through in situ snow where the average density ment sufficiently protects the structure.is known. The profile at the western end of the stripsite was averaged and matched with the average CBR of processed snowdensity from the cores near the station. Only one sample of pavement snow was able to

The high-density case was provided by the be tested. The density of the sample was not uni-pavement. The average maximum value, which form, and there was a significant difference in thecoincided with the top of the base course, was used. CBR values obtained from each end. The CBR of theThis snow was aged longest and therefore fits best denser end was 49.9%, and for the less dense endwith the fully aged in situ snow. The penetrometer (corresponding to the pavement surface) it wastip had also been sufficiently deep to avoid surface 29.9%. Theaverage of theCBRs from both ends waseffects, which significantly reduce the hardness 39.9%.values (Neidringhaus 1965). Plotted on Figure A26, the point for the average

The values in Table A5 are plotted against log- CBR and average density (0.691 Mg m- ') is belowlog scales in Figure A30. The straight line is a the line applicable to the in situ CBRs. This is mostgeometric regression fit and has the equation likely due to the short period of age-hardening of

the pavement snow. The CBR for 0.691 Mg m-3 byR = 2.46 x 105 r11" (A8) eq A10 is 78%, which is twice the measured value.

46

It seems reasonable to expect that the pavement important that the runway pavement is sufficientlyCBR would substantially increase with further thick to accept some surface weakening and re-aging. main operationally safe.

Correlation of penetrometer Pavement performancehardness and CBR The plate test results and CBR determinations

Equation A8 relates the standard rammsonde provide the opportunity to compare actual pave-hardness to snow density, and eq AI0 relates the mentperformancewith calculated predictions. TheCBR to snow density. These two equations can be next section outlines the calculations using theused to derive a third equation, the CBR to method of Russell-Head et al. (1982).rammsonde hardness relationship:

Calculation of settlementCBR = 1.68 R°'432 (A12) Table A6 lists the pertinent properties of the test

pavement and plate. The calculation that followsand the inverse: uses the Burmister (1945) two-layer system theory

as given by Poulos and Davis (1974).R = 0.303 CBR23. (A13) The values of elastic modulus E were obtained

from Russell-Head et al. (1982), and the value of theThe R value is the standard rammsonde hard- displacement factor is from Poulos and Davis

ness as calculated by eq A6. The CBR value is from (1974).a test with a piston speed of 1 mm per minute. Thestandard American test speed is 0.050 inches per Settlement = 1.5 (pa / E2) IP (A14)minute (1.27 mm per minute). It is likely that the = 3.2 mm.27% higher piston speed will give different CBRvalues. Some allowance may need to be made Observed plate settlementwhen comparing CBR values if test speeds are not The plate movement with time is shown in Fig-the same. ures A27 and A28. As mentioned earlier, the sur-

face of the pavement was not flat, and there is moreTemperature and age effects settlement than if it had been flat. The elasticon snow strength modulus values in Russell-Head et al. (1982) are

The testing program did not address these prob- derived from CBR tests and therefore have a timelems directly, and littlLmquantitative data were factor inherent to the CBR test. Therefore the settle-collected at the site. The temperature effect on CBR ment calculation is applicable for a few minutes ofand rammsonde hardness cannot be ascertained at loading.this stage, although it is clear that there is a substan- The plate settlement after a few minutes was 4-5tial loss of strength at snow temperatures above mm, and as discussed earlier, is probably higher-1.50C. than that for a flat surface. Taking this factor into

An aging effect on strength was qualitatively account, the match with the calculated value isobserved at the site. It seemed to be temperature reasonable.related in that age-hardening occurred more quicklyat high temperatures. The only quantitative infor-mation, the increase in rammsonde hardness of Table A6. Plate and pavement parameters for theprocessed snow, is shown in Figure A23. The three- plate load test.fold increase in R indicates a probable doubling ofthe CBR value. The aging of processed snow is also CBR of pavement 40%likely to depend on the particle size distribution, as CBR of subgrade 10%

well as snow density and temperature. E of subgrade (E2) 38 MPaThe data set required to adequately quantify the E,/Ef 10.1

temperature and aging effects would be quite large. Pavement thickness, h 350 mm (including base course)The expense is probably not justified for the Casey Plate radius, a 300 mmrunway, where adequately high CBR values can be Dead weight 16.2 Mg

runway Plte load 159 kNeasily reached. However, the performance of the Plate pressure, p 62 kParunway at high ambient temperatures and high Displacement factor, lP 0.48incident radiation should be investigated. It is

47

The settlement calculated is for a few minutes pressed snow pavement at the Lanyon Junctionloading and will probably overestimate the tire rut site.ior a moving aircraft. On the other hand the settle-ment after prolonged loading will be a number of In situ 'testing of pavementtimes the calculated value. The density of the snow is the most important

Although there may be room for more refine- parameter for assessing pavement performance.ment in the pavement analysis, the general out- While it may not be easy to obtain cores from thecome of the plate test is clear. The test pavement is pavement, the direct measurement of the densityadequately strong for C-130 operation. The degree from core samples is probably the most reliableof adequacy suggests that hard-standing areas for way of checking the quality of the pavement.fully laden aircraft could be made from compressed The next best check is to perform rammsondesnow as well. profiling. A new penetrometer should be con-

structed with a 20- to 25-mm-diameter tip with theDesign parameters for snow same geometry as the rammsonde. Routine profil-pavement at Lanyon Junction ing through the pavement zone and into the sub-

The test pavement was shown to be adequately grade is a quick method of checking uniformity ofstrong for operation by wheeled C-130 aircraft. The the pavement thickness during the productionpavement course itself is fairly thin at 200 mm, and phase. The standard rammsonde values can bethe base course makes an important contribution to used to assess the present strength and also tothe overall strength of the pavement, monitor age-hardening. The density value should

The question of allowing a thinner pavement to be used to estimate the aged strength.be used in the design of the Lanyon Junction site The CBR tests are neither quick nor easy, butarises. The most difficult problem is the lack of they do give the strength of the snow directly. Theknowledge about the softening and melting during field method should not be used, and samplesthe warmest days. The denser the pavement, the should be obtained from site and tested in a ma-less this problem will be, but it may be easier at first chine in a cold laboratory.to build a thicker pavement than a denser one.

The test pavement appears to be a reasonablemodel for the design of the runway at Lanyon CONCLUSIONS ANDJunction. Table A7 summarizes the parameters for RECOMMENDATIONSrunway construction, according to present knowl-edge. These requirements may be modified in the The density of the in situ snow is greater thanlight of further field tests. had been previously estimated. Earlier work sug-

gested that the pavement thickness would be in theTable A7. Parameters forpavement con- 0.3- to 0.8-m range. The total pavement thicknessstruction at Lanyon Junction (with an (pavement plus base course) of compacted snowaverage sub-grade density of 0.48 Mg with an average density of 0.63 Mg m-3 is nowm-3). estimated to be about 300 mm.

Relationships between rammsonde hardness,Minimum pavement thickness 200 mm CBR and density have been found for in situ snow,Minimum base-course thickness 100 mm and these compare well with data from other sites

Minimum base-course density 0.58 Mg M-3 and laboratory studies.

The construction equipment was able to pro-duce a base course and pavement with sufficiently

The requirements of a hard-standing area may high density. A section of pavement was tested tobe met by the pavement as specified in Table A7. C-130 loads with a 600-mm-diameter plate loadedSettlements of a few centimeters while a fully laden by 16.2 Mg, giving a pressure of 562 kPa. Theaircraft is stationary would seem acceptable. If a average settlement after 2.5 hours was less than 10smaller settlement is demanded, the snow density mm. The calculated settlement for a fully ladencould be increased by additional rolling at high aircraft stationary for a few minutes is less than 5roller mass and high tire pressures during the mm, which is in good agreement with the observedwarmer days of the summer period. There does not plate behavior.appear at this stage to be a need for additional The equipment for snow testing was generallyreinforcement (e.g. aluminum matting) of a com- quite suitable for assessing the strength character-

48

istics of the in situ and processed snow. The out- REFERENCEScome of the testing program is that the pavement asproduced in the trial construction is adequately Burmister, D. M. (1945) The general theory ofstrong for wheeled C-130 operation. stresses and displacements in layered soil systems.

The results presented here are recommended as Journal of Applied Physics, 16 (2): 89-96.the basis for a design specification of the pavement Cameron, R. L., 0. H. Loken and J. R. T. Molholmfor the Lanyon Junction runway for operation by (1959) Wilkes station glaciological data, 1957-1958.wheeled aircraft. The density profile should be Ohio State University Research Foundation Projectused as the primary parameter for compressed 825, Report No. 1, Part 3.snow runway design and construction control. Hiley, A. (1925)A rational pile drivingformula and

Laboratory CBR tests should be performed regu- its application in piling explained. Engineeringlarly during the construction period to directly (London), 119: 657,721.assess the snow strength. A cold laboratory should Niedringhaus, L. (1965) Study of the rammsondebe maintained on site for this and other snow for use in hard snow. U.S.A. Cold Regions Researchtesting purposes. and Engineering Laboratory, Technical Report 153.

The rammsonde penetrometer should be used Poulos, H. G. and E. H. Davis (1974) Elastic solu-as the primary depth profiling instrument. A small- tions for soil and rock mechanics. New York: Johndiameter rammsonde with the standard tip geome- Wiley and Sons.try should be developed for profiling dense snow. Russell-Head, D. S., W. F. Budd and P. J. MooreThe instrument should be used regularly during (1982) Compressed snow and ice airstrip construc-construction to check pavement thickness and age- tion in Antarctica. Melbourne University Pro-hardening. gramme in Antarctic Studies, Report No. 52.

Snow tests should be performed in accordance Russell-Head, D. S., W. F. Budd and P. J. Moorewith a set of standard procedures to ensure com- (1984) Casey snow runway data report, 1983-84.patibility with the design specifications. Proof roll- Melbourne University Programme in Antarcticing to simulate full C-130 loading should also be Studies, Report No. 63.used at the final stages of the initial construction to Waterhouse, R. (1966) Reevaluation of theverify the performance of the entire length of run- rammsonde hardness equation. U.S.A. Cold Re-way. gions Research and Engineering Laboratory, Spe-

Further laboratory work should be done on cial Report 100.pavement weakening by summer melting. A sec- Weller, G. E. (1967) Radiation fluxes over an Ant-tion of test pavement should be made early in the arctic ice surface, Mawson, 1961-62. ANARE Sci-construction period and monitored for strength entific Reports, Publication No. 96.throughout the summer melt period.

49

APPENDIX B: A DESIGN AND TESTING MANUAL FOR THECONSTRUCTION OF COMPRESSED-SNOW RUNWAY PAVEMENTS*

SUMMARY The characteristics of the snow surface may beextremely varied from site to site, but typical snow

This appendix describes methods of designing densities range from 0.3 to 0.5 Mg m-3, with littleand testing compressed-snow pavements for air- change in broad average density over several meterscraft use. The methods are based on the results of of depth. In addition there are tendencies for hori-laboratory and field investigations. The design of a zontal crusts and thin layers of harder or densercompacted-snow pavement is governed by the snow to occur at the current surface and at olderaircraft type and the strength of the site material. surfaces at depth. These surface features have largeThe data collected during the 1983-84 summer site horizontal dimensions in comparison to their thick-investigation and construction trials at Lanyon ness. The snow or crust hardness depends not onlyJunction on the Law Dome, near Casey, Antarctica, on the density but also on the temperature and theare used in an example design of a pavement for age-hardening (caused bysintering, which cementsuse by C-130 aircraft at that site. Generalization to the snow grains together).other sites and aircraft is indicated. Obviously snow is not a conventional pavement

Wheel settlements are calculated using Burmis- or subgrade material. However, extensive testingter's elastic theory of two-layer systems. The elastic has shown it to be quite suitable as a subgrade andmoduli needed for the analysis are obtained from as a base material for reworking to form a pave-the results of California Bearing Ratio (CBR) tests. ment. The technique of preparing an upper layerThe modulus value of snow depends mostly on its by disaggregation, compaction and rolling pro-density but is also influenced by its temperature vides the basis for a two-layer system of upperand age. Settlement charts are presented for a range pavement and lower subgrade which, it has beenof pavement thicknesses and densities and strength found, conforms adequately to the conventionalof the lower layer (subgrade). Techniques are out- two-layer approach of conventional flexible pave-lined to measure the strength of the subgrade. ment design (Yoder 1959), provided appropriate

The methods for testing the pavement during design criteria are adhered to.and after construction are described. The primary For wheeled vehicles, deep settling or boggingpavement parameters are density, thickness and in soft snow is an obvious problem. A more serioustemperature. Procedures for monitoring the problem occurs, however, when a wheel breaksstrength of the pavement by rammsonde profiling through a surface crust. One would want to avoidand CBR tests are given. Plate bearing testing and the situation where a hard surface supports aproof rolling are the best methods of simulating wheeled vehicle to travel on in parts but, when theaircraft wheel loading. A testing schedule for the wheel does break through, deep settling in therunway and taxiway pavement and hard-standing softer snow below results in severe bogging.apron pavement is outlined. This report aims to provide a concise summary

of the techniques of design and testing for com-pacted-snow runways for wheeled aircraft. The

INTRODUCTION techniques of pavement design and testing forcompacted-snow pavements retain as much con-

This appendix addresses the problem of the vention as possible and avoid the introduction ofdesign and testing of compacted-snow runways arbitrary tests. The results of detailed laboratoryfor wheeled aircraft for sites on the Antarctic ice and field studies of snow in its natural and com-sheet or similar locations. A separate report has pacted states and the performance of a constructedbeen prepared for the techniques of construction of trial strip pavement area at the Lanyon Junctioncompacted runways (Appendix C). site on Law Dome, near Casey, Antarctica, bring

theory, research and practice into conjunction.The detailed testing programs and analyses are

'Modified slightly from a report publishe