Rapid changes in time and space in the micro-relief of an ablating glacier surface and radiation-induced melt within the uppermost ice layer, termed the “weathering crust”, seriously affect the accuracy of the short-term ablation measurements. The various measuring techniques commonly used (stakes, ablatometers, ablatographs) and some new methods (measurement of discharge from a small supra-glacial drainage basin, and mass loss directly measured on core samples) are critically reviewed and assessed in the light of these phenomena. The implications for studies of heat and mass balance are discussed.It appears that the direct measurement of mass flux is the most accurate means of assessing short-terni ablation rates. The errors in short-term ablation measurements by any method are largely compensatory and consequently do not influence long-period mass-balance estimates.

ABSTRACT. R a pid changes in time and space in the micro-relief of an ablating glacier surface and radiation-induced melt within the uppermost ice layer, termed the "weathering crust" , seriously affect the accuracy of the short-term ablation measurements. The various measuring techniques commonly used (stakes, ablatometers, ablatographs) and some new methods (measurement of discharge from a small supraglacial drainage basin, and mass loss directly measured on core samples) are critically reviewed and assessed in the light of these phenomena. The implications for studies of heat a nd mass balance are discussed.It appears that the direct measurement of m ass Aux is the most accurate means of assessing short-term a blation ra tes. The errors in short-term ablation measurements by any m ethod are largely compensatory and consequently do not inAuence long-period m ass-balance estima tes . RESUME. Erreurs des mesures d'ablation cl court terme de la surface de glace enfonte. Des changements rapides dans le temps et l'espace du micro-relief de la surface d'un glacier soumise a l'a blation et la fusion due a la radiation dans la croute superficielle de la glace. affecte serieusement la precision des mesures d'ablation a court terme. Les techniques variees utilisees comm unement (balises, ablatometres, ablatographes) et quelques methodes nouvelles (mesure du debit d'un petit bassin de drainage superficiel , et perte de masse mesuree directement sur des carottes) sont revues d ' une maniere critique a la lumi ere de ces pheno menes.Les implications pour les etudes d'energie et de m asse sont discutees.Il apparait que la mesure directe du Aux d e masse est le moyen le plus prec is pour ev iter l'effet de mesures a court terme de la vitesse d'a blation. Les erreurs de mesures d'ablation a court terme pour chaque methode se compensent largement et en consequence n'i nAuence pas les timations du bilan de masse pour un e longue periode d'observation .

Data on the physical properties of seasonal alpine snow have been collected from the Beartooth Mountains near Cooke City, Montana (elevation ≈3000 m) and the Bridger Range near Bozeman, Montana (elevation ≈2200 m). Systematic measurements of snow density, temperature, structure, ram and Canadian hardness, centrifugal tensile strength and shear strength measured with a shear box and several types of shear vanes are included. Test results were grouped according to gross snow types (cohesive fine-grained “winter” snow. depth hoar, new snow, etc.) and whether the snow was wet or dry. Then interrelations between the different test parameters were studied. A plot of ram number versus density for winter snow gave a log-linear relation similar to that suggested for polar snows. Both shear-vane and centrifugal-tensile results when plotted as a function of porosity are well described by the negative exponential relation suggested by Ballard and Feldt. Depth hoar and wet snow invariably have lower strength values at any given density. There is an excellent one-to-one agreement between values obtained with the shear vane and the shear box.Several field experiments were performed to study the sources of error in making in-situ mechanical tests on snow without utilizing a pit wall. Statistical analysis of the results shows that the main factor contributing to the experimental scatter is lateral inhomogeneity in the snow cover. There was no significant difference between the results of different operators. The standard deviation of a group of strength tests is shown to be directly proportional to the mean value of the group. This indicates that a logarithmic transformation should be made in handling snow strength results in order to stabilize the variance. It is emphasized that the systematic relations between snow properties invariably become obscured when different snow “types” are indiscriminately grouped together.

Simultaneous m easurements were made of both th e strength a nd structural p roperti es of a d ry, seasona l snow-pack. T he d ensities of the snow studi ed ra nged fro m J 26 to 407 kg m-3 . T h e mecha ni ca l strength of this snow as determin ed by in situ shea r-wave test a nd a centrifuga l tensile tes t increases exponentiall y with the increase in size of interg ran ul a r sands. Bonding increase at a ra te substantiall y g rea ter tha n that predicted by the classical sintering eq ua tions. It is suggested that this is due to the wid e variety of stresses present in snow with these relatively low d ensities .R ESU ME. La croissance de liens et I' augmentation de la resistance mecanique dans une couche saisonniere seche. D es m esures simulta nees furent effectuees a la fois des proprietes d e resista nce e t de structure d ' un e couche de n eige saisonniere seche. Les densites d e la neige etudiees e taient comprises entre J 26 et 407 kg m-J. L a resistance mecanique d e cette neige determin ee par test d 'onde de cisaill ement e t un tes t de tension centrifuge, c rait d ' un e maniere exponenti elle avec l'augm entation de la tai ll e d es li asons intergran ul aires. L es li ens d e cohesions a ugmentent d 'un e quantite superieure a cell e prevue pa r les equ a tions cl assiques de tassement. 11 es t suggere que eel a est du a la la rge varie te d e contra intes d e rel a ti vites prese ntes dans la neige a fa ibles de nsites . ZUSAMMENFASSUNG. D as Wach stlllll von B rucken und die Zwwhme der mechanischen Festigkeit in einer trockenelljahreszeitlichen ScllIleedecke. Es wurden gleichzeitige M essungen d er F estigkeit und del' strukturellen Eigenscha ften in ein er trackenen, j a hreszeitli chen Schn cedecke durchgefuhrt. Die Dichte des unte rsuch ten Schn ees reichte von J 26 bis 407 kg m-3 . Di e mechan ische F es tigkeit d ieses Schnees, die durch ein e Scherwell enuntersuchung in situ und d urch ein e zentrifuga le Deh nungsu ntersuchu ng bestimmt wurde, wachst exponenti ell mit der G rossenzun a h me del' Brucken zwischen den K o rn ern. Die Bindung nimmt mil ein er Gesch windigkeit zu, die deutlich grosser is! a ls di e durch die klassischen Sinterungsgleichungen vora usgesagte. Es wird vermu te t, dass di es a uf die grosse Variationsbreite d el' in Schn ce mit so relati v geringer Di chte vorhande nen Span nungen zurli ckzuflihren ist.