Processus Gravitaires Dans La Vallée Tasiapik (Nunavik) : Témoins Géomorphologiques De La Dynamique De Versant Récente Et Passée

Processus gravitaires dans la vallée Tasiapik (Nunavik) : témoins géomorphologiques de la dynamique de versant récente et passée

Mémoire

Samuel Veilleux

Maîtrise en sciences géographiques - avec mémoire Maître en sciences géographiques (M. Sc. géogr.)

Québec, Canada

Résumé

Ce projet de recherche, mené près d’Umiujaq (Nunavik), a été réalisé afin de documenter les processus gravitaires dominants qui se manifestent sur les versants escarpés de la vallée Tasiapik. Des relevés topographiques, granulométriques, morphométriques, pétrographiques et de végétation ont permis de caractériser 18 talus d’éboulis situés sur les versants sud-ouest et nord-est. Les résultats obtenus ont permis d’établir que l’éboulisation importante sur les versants de la vallée est un phénomène à la fois ancien, engendré par des processus paraglaciaires, et récent, résultat de processus périglaciaires toujours actifs aujourd’hui. Cela se traduit notamment par différents stades de développement des talus d’éboulis subactuels, illustrés par certains dépôts de versant frais et d’autres très anciens, mais aussi avec des topographies de pente très variables. Sur une échelle de temps plus courte, entre août 2017 et juillet 2018, les avalanches se sont avérées être un processus majeur, tel qu’observé sur les 14 000 photographies obtenues grâce à trois caméras à déclenchement automatique installées à l’été 2017. Ces avalanches, souvent déclenchées par des chutes de corniches à neige, ont été particulièrement fréquentes au printemps 2018, résultant de conditions météorologiques propices telles qu’une hausse rapide des températures journalières et des épisodes de pluie abondante. Des dépôts d’avalanches sales témoignent d’une grande capacité érosive, et incidemment de leur grand apport en débris vers les talus d’éboulis. Dans certains cas, les dépôts d’avalanche avoisinaient la route située en contrebas, démontrant ainsi un risque potentiel pour ses usagers.

iii

Abstract

This research project was conducted near Umiujaq (Nunavik) to document the main gravitational processes that occur on the slopes of Tasiapik Valley. 18 talus slopes on the southwest and northeast sides of the valley were characterized using topographic, granulometric, morphometric petrographic and vegetation surveys. Results show that talus formation in the valley is an ancient phenomena, due to paraglacial processes, and recent - and still ongoing – periglacial processes. This is evidenced by different development stages among the talus slopes, with fresh and very old debris covering the slopes, as well as contrasting slope topographies. On a shorter and more recent time scale, from August 2017 to July 2018, snow avalanches have proven to be a major process, as observed on the 14,000 photographs obtained using three automatic timelapse cameras installed in the summer of 2017. Snow avalanches were often triggered by a collapsing snow-cornice and were very frequent in the spring of 2018 due to favourable meteorological conditions such as a rapid increase in daily temperatures and abundant rainfall events. Dirty snow-avalanche deposits have shown the great erosive capacity of these snow avalanches, thus their important debris supply toward the talus slopes. In some cases, runout zones were located only a few meters from the road below, thus showing the potential risk for people travelling on the road.

Table des matières

Résumé ...... iii Abstract ...... iv Table des matières ...... v Liste des figures ...... vii Liste des tableaux ...... ix Remerciements ...... xi Avant-propos ...... xii Introduction générale ...... 1 I. Introduction ...... 1 II. Région et site d’étude ...... 3 Localisation ...... 3 Géologie régionale ...... 3 Géomorphologie quaternaire ...... 4 Climat et végétation ...... 5 III. Problématique, objectifs et hypothèses ...... 7 Énoncé du problème ...... 7 Objectif général ...... 7 Objectifs spécifiques...... 8 Hypothèses ...... 8 IV. Méthodologie ...... 9 Références ...... 13 Chapitre 1 ...... 17 Résumé ...... 18 Abstract ...... 19 I. Introduction ...... 20 II. Study area ...... 21 III. Methods ...... 24 IV. Results and interpretation ...... 28 1. Topography of slope deposits ...... 28 2. Relative dating of slopes ...... 30 3. Source and morphometry of slope deposits ...... 34

4. Debris runout ...... 40 5. Short-term slope dynamics ...... 42 V. Discussion ...... 43 1. Talus slope formation ...... 43 2. Slope debris redistribution ...... 45 3. Rockwall erosion ...... 48 VI. Conclusion ...... 51 References ...... 52 Chapitre 2 ...... 59 2.1 Résumé ...... 60 2.2 Abstract ...... 61 I. Introduction ...... 62 II. Study area ...... 63 III. Methods ...... 65 1. Time-lapse cameras ...... 65 2. Meteorological analysis ...... 65 3. Fieldwork ...... 65 IV. Results and interpretation ...... 67 1. Snow cornice formation and collapses ...... 67 2. Snow-avalanche events ...... 69 3. Snow-avalanche deposits ...... 73 Discussion ...... 76 1. Meteorological and topographic control ...... 76 2. Runout and proximity to the road ...... 79 3. Limitations ...... 81 Conclusion ...... 82 References ...... 83 Conclusion générale ...... 86

Liste des figures

Figure 1: Localisation de la vallée Tasiapik et les environs d'Umiujaq; géologie régionale et couverture sédimentaire dans la vallée Tasiapik; distribution des dépôts de versant et localisation des sites étudiés ...... 6 Figure 2: Localisation des caméras dans la vallée, avec leur angle de champ et leur cadre ...... 12 Figure 3: Location of Tasiapik Valley within the Umiujaq area; regional geology and quaternary sediments in Tasiapik Valley; distribution of the talus slopes and investigated slopes ...... 23 Figure 4: Location of the cameras along the SW side of Tasiapik Valley ...... 27 Figure 5: Longitudinal cross-section of the investigated slopes ...... 29 Figure 6: Vegetation stages along the talus slopes ...... 32 Figure 7: Slope development stage according to three parameters: 1) vegetation; 2) Ho/Hi index and 3) a combination of the two ...... 33 Figure 8: Proportion of the different lithologies along the talus slope ...... 36 Figure 9: Distribution of the debris size (a-axis) along the talus slopes. Close-up of the largest sampled debris along SW-09 profile and NE-09 profile ...... 37 Figure 10: Cross-section of the slopes (talus/rockwall) showing the reach and shadow angle for the farthest slope debris, and the hypothetical travel distance from the source area ...... 41 Figure 11: Snow-avalanche event exhibiting a dirty deposit, as the result of incorporated debris; discrete rockfall event and debris sliding onto the snow-covered talus ...... 42 Figure 12: Recent notches formed in the rockwall along SW-07 and NE-09 profiles; dense shrub cover at the base of the slope near SW-07 and SW-08 profiles and mid-slope near NE-05 and NE-06 profiles ...... 44 Figure 13: Cluster of basalt boulders located ~250 m away from the slope, overlying littoral sediments ...... 47 Figure 14: Erosion of the basalt layer at the top of the cuesta on the SW side, with tall monolith on the verge of falling, resulting in a sawtooth shape exposed on the rockwall; frost heave and extensive jointing of the basalt bedrock; basalt and sandstone layers overhanging quartz arenite layers on the NE side ...... 48 Figure 15: Location of Tasiapik Valley in the Umiujaq region; oblique view toward the north, showing the cuesta frontslope on the SW side ...... 64 Figure 16: Location of the cameras along the SW side of Tasiapik Valley ...... 66 Figure 17: Meteorological conditions during the cornice accretion period: temperature, snowfall, wind speed and wind direction ...... 67 Figure 18: Schematic cross-section and photographs representing the accretion, the creeping and collapsing/melting periods ...... 68 Figure 19: Snow avalanche events identified by the TAS2 camera in 2017-2018...... 71 Figure 20: Occurrence of cornice failures and snow avalanches during the winter/spring of 2017-2018 ...... 72

vii

Figure 21: 3D view of the SW side of Tasiapik Valley, highlighting the snow avalanche deposits observed in the field in June 2018, with topographic long profiles and on-site photographs for deposits respectively located downstream, mid-valley and upstream ...... 75 Figure 22: Schematic cross-section showing the snow redistribution by the wind, and the formation of a lee zone deposition on the slope below the snow cornice ...... 77 Figure 23: Snow cover on the SW side and NE side on June 12, 2018. Shaded SW side in June 2018 ...... 78 Figure 24: Wet snow-avalanche deposits with runout distances near the road ...... 79 Figure 25: Dirty snow avalanche deposits in June 2018. Comparison of a shorter runout distance from this event with another wet snow avalanche with a longer runout occurring in late May 2018 and a slab avalanche occurring in April 2018 ...... 80 Figure 26: Days with poor visibility on the three cameras installed in Tasiapik Valley ...... 81

viii

Liste des tableaux

Table 1: Topographic parameters of the investigated slopes...... 28 Table 2: Calculation of Ho/Hi index...... 31 Table 3: Debris size and morphometric parameters along the longitudinal profiles...... 38 Table 4: Calendar of snow-avalanche activity...... 70

Mental toughness, la dureté du mental. Pis icitte présentement, y’en a pas mal plus qu’on pense du mental.

Bob Chicoine Les Boys, 1997

Remerciements

J’aimerais remercier tous ceux et celles qui, de près ou de loin, ont contribué à l’aboutissement de ce mémoire. D’abord, un énorme merci à Najat Bhiry et Armelle Decaulne pour leurs nombreux et judicieux conseils, leur rigueur scientifique et leur disponibilité. Je les remercie également de m’avoir fait découvrir le Nunavik au cours des trois dernières années. Ensuite, je tiens à remercier Michel Allard et Thorstein Saemundsson pour leurs conseils et pour les discussions qui m’ont aidé à approfondir mes réflexions. Je remercie également les gens de la communauté d’Umiujaq et du parc national Tursujuq pour l’intérêt porté envers ce projet, ainsi que les collègues qui ont prêté assistance sur le terrain, dans des conditions parfois difficiles. Enfin, je souhaite remercier mes proches et mes ami(e)s qui ont su m’encourager et me donner un coup de main toujours très apprécié.

Avant-propos

Ce mémoire est rédigé sous forme de deux articles soumis à des revues scientifiques pour publication. Les chapitres 1 et 2 du mémoire reprennent l’essentiel de ces articles, rédigés en anglais, portant respectivement sur l’étude des processus d’éboulisation et des avalanches de neige. L’introduction et la conclusion générale du mémoire sont rédigées en français.

Les sections se détaillent comme suit :

Introduction générale

Chapitre 1 – Veilleux, S., Bhiry, N., & Decaulne, A. - Talus slope characterization in Tasiapik Valley (subarctic Québec): evidence of past and present slope processes. Soumis à la revue Geomorphology.

Chapitre 2 – Veilleux, S., Bhiry, N., & Decaulne, A. - Snow cornice and snow avalanche monitoring using automatic time-lapse cameras in Tasiapik Valley, Nunavik. Soumis à la revue Cold Regions Science and Technology.

Conclusion générale

xii

Introduction générale I. Introduction

Les hauts reliefs se font plus rares dans le Nord québécois que dans les grands massifs montagneux connus, la majorité du territoire étant constituée de plateaux, de bassins et de basses collines (MRNF, 2010). Ainsi, peu de chercheurs se sont intéressés à la dynamique et la géomorphologie des versants dans cette région, bien que cette thématique soit largement étudiée dans les milieux froids (Luckman, 2013). Les versants qui ont été étudiés sont situés en région isolée, notamment au lac Wiyâshâkimî (Bégin et Filion, 1985 ; St-Cyr, 1986 ; Marion et al., 1995 ; Decaulne et al., 2018) et à la presqu’île de Manitounuk (Belzile, 1984). Quelques hauts reliefs se profilent néanmoins à proximité d’établissements humains, et comme des lacunes subsistent quant à la compréhension de la dynamique de ces versants, il convient de s’y intéresser. L’avalanche mortelle de 1999 à Kangiqsualujjuaq est un cas concret où un manque dans la documentation des mouvements de versants a eu des conséquences directes sur la population (Hétu et al., 2008). Il est d’autant plus logique d’adresser l’enjeu de la vulnérabilité des populations face aux risques naturels dans le contexte actuel de croissance démographique et de création de parcs nationaux au Nunavik, où l’on peut envisager d’éventuelles hausses de fréquentation des lieux publics (Duhaime, 2007).

Parmi ces hauts reliefs, les cuestas hudsoniennes forment un escarpement qui longe la côte est de la baie d’Hudson, révélant des parois rocheuses dépassant 200 mètres de hauteur. Ces versants abrupts font partie intégrante des paysages du lac Tasiujaq et du village d’Umiujaq, et sont l’un des principaux attraits du parc national Tursujuq (ARK, 2007). Or, la présence de dépôts de versant au pied de cet escarpement témoigne d’une occurrence de processus gravitaires dans la région. C’est notamment le cas dans la vallée Tasiapik qui, bien que ne représentant qu’un petit échantillon de la région naturelle des cuestas hudsonniennes, constitue un corridor naturel faisant la liaison entre Umiujaq et le lac Tasiujaq, dont le rivage représente les limites du parc national Tursujuq. Conséquemment, il s’agit d’un endroit très fréquenté où circulent les employés et visiteurs du parc, les chercheurs, mais surtout les Inuit d’Umiujaq, pour qui cet endroit a une signification particulière puisqu’ils y pratiquent des activités traditionnelles telles que la chasse, la pêche et la cueillette de petits fruits (ARK, 2007). D’ailleurs, une route carrossable construite au milieu des années 2000 permet de faciliter l’accès à ce secteur, mais sa proximité avec la paroi rocheuse et les dépôts de versant pourrait engendrer des risques pour ses usagers.

Des campagnes de terrain durant trois étés consécutifs (2016, 2017 et 2018) ont permis d’investiguer les versants de la vallée Tasiapik. Ce projet de recherche vise à documenter les mouvements gravitaires, plus spécifiquement les éboulements rocheux et les avalanches, ainsi que leurs processus de déclenchement. L’étude se concentre sur l’évolution des versants au cours de l’Holocène, soit la période postérieure au retrait de l’inlandsis laurentidien, ainsi que sur une échelle de temps plus récente, soit par l’étude des dépôts avalancheux. Enfin, bien que ce projet aborde les risques associés aux mouvements gravitaires, il ne s’agit pas d’un exercice d’évaluation de la vulnérabilité, mais plutôt d’un mémoire visant l‘avancement des savoirs et la compréhension de ces aléas naturels en milieu subarctique.

II. Région et site d’étude

Localisation La vallée Tasiapik (56°33'N, 76°28'O) se situe à l’extrémité nord du lac Tasiujaq (anciennement nommé lac Guillaume-Delisle et Richmond Gulf), environ 5 km à l’est du village d’Umiujaq sur la côte est de la baie d’Hudson, au Nunavik (Figure 1a). La vallée fait environ 4,5 km de long et 1,5 km de large, suivant une orientation nord-ouest sud-est. La hauteur des versants varie entre 50 m à l’amont de la vallée à près de 230 m à l’aval, près du lac Tasiujaq. Ce dernier est connecté à la baie d’Hudson par un étroit passage nommé Tursujuq (Le Goulet) situé 40 km au sud d’Umiujaq. Le lac fait partie du parc national Tursujuq.

Géologie régionale La vallée Tasiapik est située à la rencontre de deux grands ensembles géologiques. D’abord, le socle précambrien de la province du Supérieur affleure au sud-est de la vallée, constitué de roches ignées et métamorphiques (granite, granodiorite et gneiss) d’âge néoarchéen, soit entre -2,73 et -2,68 Ga (Chandler et Schwarz, 1980 ; Chandler, 1988 ; Percival, 2007 ; Eaton et Derbyshire, 2010). Les versants sud-ouest et nord-est (colline Umiujaaluk) correspondent au groupe de Nastapoka et à la formation de Qingaaluk, une séquence volcano-sédimentaire paléoprotérozoïque (Figure 1b). Sous cette séquence se trouve le groupe de Richmond Gulf, aussi constitué d’un étagement de roches volcano-sédimentaires, mais n’affleurant qu’au sud-est la vallée (Stockwell et al., 1979 ; Chandler et Schwarz, 1980 ; Chandler, 1988 ; Eaton et Derbyshire, 2010). Cette séquence supracrustale a été mise en place lors de l’accrétion de la ceinture circum-supérieure, une marge passive paléoprotérozoïque, au craton du Supérieur pendant l’Orogénèse trans-hudsonienne (-1,8 Ga). Cet événement a d’ailleurs provoqué l’affaissement du graben du golfe de Richmond (Chandler et Schwarz, 1980 ; Chandler, 1988 ; Chandler et Parrish, 1989). Le pendage des couches forme un relief monoclinal asymétrique, appelé cuesta, se caractérisant par un revers faiblement incliné en direction de la baie d’Hudson vers l’ouest et d’un front abrupt orienté vers l’est ; cet escarpement s’étend sur 650 km le long de la côte est de la baie d’Hudson (Dionne, 1976 ; Guimont et Laverdière, 1980). La colline Umiujaaluk au nord-est, qui fait également partie du complexe de cuestas hudsoniennes, est une butte témoin dissociée de l’escarpement principal à la suite de processus érosifs. Les versants de la vallée Tasiapik expose, du haut vers le bas, des couches de basalte, de grès, de dolomie, d’arénite quartzitique et de calcaire laminaire. On distingue d’abondantes formes

d’érosion glaciaires telles que des stries, des cannelures et des broutures dans le substrat basaltique passablement poli situé au sommet des cuestas. Leur orientation témoigne d’un écoulement glaciaire vers l’ouest lors de la dernière glaciation (Craig, 1969 ; Hillaire-Marcel, 1976 ; Allard et Séguin, 1985).

Géomorphologie quaternaire Les dépôts quaternaires dans la vallée Tasiapik témoignent d’une succession d’environnements sédimentaires postérieurs à la glaciation wisconsinienne (Figure 1b). Suite au retrait de l’inlandsis laurentidien à partir de 8200 ans cal. BP, l’affaissement glacio-isostatique a permis l’invasion des eaux marines de la mer de Tyrrell vers l’intérieur du continent. Les terres de la côte est de la baie d’Hudson ont été immergées à partir de 8000 ans cal. BP, jusqu’à une altitude d’environ 270 m au- dessus du niveau marin actuel (Hillaire-Marcel, 1976 ; Lavoie et al., 2012). De nombreuses crêtes de plages soulevées, édifiées lors de périodes de forte activité marine où les taux de relèvement glacio- isostatique et de remontée eustatique s’annulaient l’un par rapport à l’autre, sont visibles autour d’Umiujaq et du lac Tasiujaq (Fraser et al., 2005). Le taux de relèvement glacio-isostatique sur la côte est de la baie d’Hudson était d’abord rapide jusqu’à 6000 ans cal. BP, puis a progressivement diminué jusqu’à nos jours ; cet ajustement isostatique est néanmoins considéré comme étant l’un des plus rapides et importants au monde (Andrews, 1968 ; Hillaire-Marcel, 1976 ; Lavoie et al., 2012).

Dans la partie basse de la vallée Tasiapik se sont déposés des faciès d’eau profonde (silts et argiles) et des faciès d’eau peu profonde (sables). Dans la partie amont de la vallée, la présence de sédiments fluvioglaciaires témoigne d’une stabilisation de la marge glaciaire autour de 8000 ans cal. BP, qui a entre autres permis l’édification d’un fan de contact sous-glaciaire ainsi que de petites moraines frontales (Lajeunesse et Allard, 2003). De nos jours, diverses formes périglaciaires dominent le paysage de la vallée, tels que des buttes et plateaux cryogéniques ainsi que des palses (Allard et Séguin, 1987 ; Lavoie et al., 2012 ; Pelletier et al., 2018).

Enfin, de nombreux dépôts de versant se sont accumulés sur les deux versants de la vallée, au pied de hautes parois ou de ressauts rocheux. Pour la présente étude, 18 zones d’accumulation de dépôts de versant ont été étudiés (Figure 1c).

Climat et végétation La vallée Tasiapik est située en zone de pergélisol discontinu et est caractérisée par un climat subarctique. Les températures moyennes annuelles de l’air enregistrées entre 2013 et 2017 varient entre -5,6 et -4,2°C, mais avec de grands écarts, allant de 23°C à -36°C (Fortier, 2017). Les précipitations annuelles atteignent en moyenne 500 mm, dont 40% tombent sous forme de neige (Ménard et al., 1998).

La région d’Umiujaq se trouve à l’interface de la toundra arbustive et de la toundra forestière. La limite des arbres, qui délimite ces deux zones bioclimatiques, parcourt la vallée Tasiapik ; une mosaïque d’arbustes bas, d’éricacées et de lichens recouvre l’amont de la vallée alors qu’un couvert forestier dense occupe l’aval (Payette, 1983 ; Allard et Séguin, 1987 ; Provencher-Nolet et al., 2014 ; Pelletier et al., 2018). La strate arbustive, dominée par l’aulne crispé (Alnus crispa), le bouleau glanduleux (Betula glandulosa) et le saule (Salix sp.), a connu une expansion importante (arbustation) au cours du 20e siècle (Ménard et al., 1998 ; Provencher-Nolet et al., 2014 ; Pelletier et al., 2018). L’épinette noire (Picea mariana), aussi présente sous forme arbustive, constitue l’unique espèce arborescente. Enfin, plusieurs petites tourbières minérotrophes (fen) et mares thermokarstiques sont situées à proximité du rivage du lac Tasiujaq (Ménard et al., 1998 ; Pelletier et al., 2018).

Figure 1: Localisation de la vallée Tasiapik et les environs d'Umiujaq (A); géologie régionale et couverture sédimentaire dans la vallée Tasiapik (B); distribution des dépôts de versant et localisation des sites étudiés (C). Sources : MRNF, UMI orthomosaic, 2010.

III. Problématique, objectifs et hypothèses

Énoncé du problème Au Nunavik, peu d’études ont porté sur les mouvements gravitaires et la géomorphologie des versants. Parmi ces travaux, St-Cyr (1986), Marion et al. (1995) et Decaulne et al. (2018) au lac Wiyâshâkimî, situé dans le parc national Tursujuq 120 km à l’est de la vallée Tasiapik, et Belzile (1984) à la péninsule de Manitounuk à 100 km au sud d’Umiujaq, ont étudié les processus d’éboulisation sur des versants rocheux. Ces travaux ont été réalisés dans des lieux isolés et peu fréquentés. D’importants dénivelés pointent toutefois à proximité d’établissements humains et méritent que l’on s’y attarde en raison des risques naturels qui pourraient leur être associés. D’ailleurs, de nombreux processus avalancheux ont été observés dans quelques villages du Nunavik, notamment l’avalanche de Kangiqsualujjuaq en 1999 (Lied et Domaas, 2000 ; Germain, 2016).

La vallée Tasiapik, à proximité du parc national Tursujuq, est un lieu très fréquenté par les Inuit d’Umiujaq. Située à quelques kilomètres du village, elle fait office de corridor d’accès au lac Tasiujaq. La route pour s’y rendre longe le bas du versant, et la présence de talus d’éboulis à proximité de celle-ci suggère que ces versants sont - ou ont déjà été - actifs. Ainsi, sans le savoir, les usagers de cette route pourraient s’exposer à des risques. L’évaluation des risques passe néanmoins par la documentation et la compréhension des processus dynamiques se manifestant sur les versants. À première vue, l’étendue de ces processus est non négligeable considérant la vaste distribution des dépôts de versant dans la vallée, faisant partie intégrante du paysage. Enfin, telle qu’observé sur le terrain en juin 2018, la vallée semble propice au déclenchement d’avalanches de neige, et plus particulièrement du côté sud-ouest, à proximité de la route.

Ce projet de recherche cherche à répondre aux questions suivantes : 1) quels sont les processus gravitaires dominants? ; 2) comment se manifestent-ils dans la vallée selon les différents types de versants? ; et 3) comment ces processus sont-ils intervenus dans l’évolution holocène des versants?

Objectif général Ce projet de recherche a comme objectif principal de documenter les processus gravitaires se manifestant dans la vallée Tasiapik et leur incidence sur la géomorphologie des versants.

Objectifs spécifiques 1) Documenter la géomorphologie associée aux éboulis rocheux et le développement des talus d’éboulis dans la vallée Tasiapik à partir de relevés géomorphologiques, depuis la dernière déglaciation jusqu’à nos jours ;

2) Documenter l’occurrence des avalanches sur les versants de la vallée Tasiapik en mettant en relation leur déclenchement avec les différents paramètres topographiques et météorologiques sur l’évolution du couvert de neige.

Hypothèses À partir de résultats préliminaires et des observations faites sur le terrain, les hypothèses suivantes ont été émises :

1) La géomorphologie des talus d’éboulis est étroitement liée au contrôle géologique du versant, la séquence lithologique jouant un rôle prépondérant dans la dispersion des débris au bas du versant ;

2) La récurrence des périodes d’activité des versants semble obéir à deux modèles théoriques, soit le modèle de tarissement (exhaustion model) et le modèle bimodal (ou plurimodal) ;

3) Les conditions météorologiques sont favorables à la formation d’une corniche neigeuse au sommet des cuestas, et la chute de celle-ci est la principale cause de déclenchement d’avalanches dans la vallée.

IV. Méthodologie

La collecte des données s’est échelonnée sur quatre campagnes de terrain au cours des étés 2016, 2017 et 2018. Avant d’entamer les travaux sur le terrain, plusieurs talus ont pu être identifiés par photo-interprétation. Deux jeux d’orthophotographies provenant du Ministère des Ressources naturelles et de la Faune du Québec (MRNF) ont été utilisés, datant de 2004 (échelle 1/10 000, 25 cm de résolution) et de 2010 (échelle 1/10 000, 15 cm de résolution). Leur analyse a permis de constater une distribution inégale des dépôts de versant dans la région d’Umiujaq. Néanmoins, ceux- ci sont particulièrement abondants au front des cuestas ; deux secteurs de la vallée Tasiapik ont été identifiés, soit la colline Umiujaaluk (nord-est) et le versant sud-ouest. Ce dernier est d’autant plus intéressant qu’il est bordé par une route connectant Umiujaq au lac Tasiujaq.

Des relevés topographiques ont été réalisés le long de 18 transects longitudinaux sur les dépôts de versant à l’aide d’un DGPS (Differential Global Positioning System). Les points de cheminement ont été enregistrés à partir de l’apex des dépôts de versant jusqu’à leur base, perpendiculairement à la paroi rocheuse. La grande précision du DGPS permet d’exposer la forme générale des talus et d’obtenir des détails de microtopographie tels que l’inflexion et la texture des dépôts. Les données ont été traitées dans ArcGIS et Excel afin de produire des profils topographiques. Différents paramètres topographiques ont aussi été produits et analysés avec les points GPS, dont l’estimation du stade d’évolution à partir du rapport Ho/Hi, où Ho correspond à la hauteur du talus et Hi à la hauteur totale du versant (Francou, 1988 ; Sellier, 1992). Un versant dont le rapport s’approche de 1 témoigne d’un stade avancé en raison de la faible hauteur de la paroi résiduelle comparativement à la hauteur du talus.

Sur 12 de ces 18 transects longitudinaux ont été réalisés des relevés granulométriques et pétrographiques, en échantillonnant 25 fragments rocheux sélectionnés aléatoirement à chaque station d’échantillonnage. Ces stations sont disposées à intervalles d’environ 10-15 m le long des transects, depuis l’apex jusqu’à la base. Les débris ont été mesurés selon trois axes, soit longueur, largeur et épaisseur. Les mesures ont été compilées dans Excel et puis analysées afin d’en ressortir des statistiques descriptives. Des indices morphométriques ont également été calculés à partir de

ces mesures (Pérez, 1998 ; Hétu et Gray, 2000). L’indice d’aplatissement (Fi) est calculé :

푎 + 푏 퐹 = 푖 2푐 où a correspond à la longueur, b à la largeur et c à l’épaisseur du fragment mesuré (Cailleux, 1947).

Plus la valeur Ai est élevée, plus le débris a une forme aplatie. Puis, l’indice d’allongement (Li) est calculé :

푎 퐿 = 푖 푏 où a et b correspondent à la longueur et à la largeur du fragment mesuré (Schneiderhöhn, 1954).

Une valeur Li élevée indique que le débris tend à avoir une forme allongée. Enfin, l’indice de sphéricité (Si) est calculé :

1 푏푐 3 푆 = ( ) 푖 푎2 où a, b et c correspondent à la longueur, à la largeur et à l’épaisseur du fragment mesuré (Krumbein,

1941). Une valeur Si s’approchant de 1 indique que le débris tend à avoir une forme plus massive, soit sphérique dans le cas d’un fragment arrondi et cubique dans le cas d’un fragment anguleux. Ces paramètres renseignent sur le comportement de chute des fragments, alors qu’un débris sphérique pourra davantage rouler tandis qu’un débris allongé et plat aura tendance à glisser.

Les relevés pétrographiques ont permis d’obtenir la lithologie des débris rocheux à partir de cassures fraîches. La provenance des débris, soit locale (associée au versant) ou exogène (origine glaciaire) est étroitement liée à leur lithologie et à leur dispersion sur le talus. L’aspect des arêtes a aussi été relevé et renseigne sur la provenance des débris ; un fragment anguleux aura une provenance locale, ayant subi peu d’érosion, alors qu’un fragment arrondi aura subi davantage de transport, notamment par la glace ou l’eau. Ensuite, la couverture végétale a été décrite à chacune des stations d’échantillonnage. Pour ce faire, des valeurs hiérarchisées ont été attribuées à chaque station en fonction du type de végétation et de l’estimation visuelle du pourcentage de recouvrement de la végétation sur les débris :

1) Débris récent : couverture faible (0-20%), peu ou pas d’espèces de lichens ;

2) Débris récent : couverture faible à moyenne (20-40%), quelques espèces de lichens ;

3) Débris moyen : couverture moyenne (40-60%), plusieurs espèces de lichens ;

4) Débris ancien : couverture moyenne à élevée (60-80%), plusieurs espèces de liches et mousses ;

5) Débris très ancien : couverture élevée (80-100%), plusieurs espèces de lichens et mousses, potentiellement couvert d’arbustes bas.

Trois caméras à déclenchement automatisé ont été installées à l’été 2017 afin de suivre l’évolution du versant sud-ouest et ainsi distinguer les mouvements gravitaires qui ont lieu au cours d’une année, et particulièrement pendant l’hiver et au printemps. Deux caméras sont situées au pied de la paroi rocheuse près de zones d’accumulation de débris rocheux, alors que la troisième est située au sommet de la cuesta. Des photos sont prises à chaque jour et à toutes les heures entre 9h00 et 17h00 pour la période entre août 2017 et juin 2018, puis à toutes les quinze minutes entre juin 2018 et août 2018. Les conditions météorologiques qui prévalent lors de certaines périodes critiques peuvent être relevées grâce aux photos, et il est possible de corréler ces conditions avec les données météorologiques des stations climatiques SILA du Centre d’études nordiques d’Umiujaq, tout comme l’évolution du couvert nival au cours de l’hiver et au printemps. Enfin, des relevés héliportés ont été réalisés, au cours desquels il a été possible de photographier les versants sous divers angles, et d’y observer les particularités du terrain à plus petite échelle. Le sommet des cuestas a aussi été arpenté à pied afin de discerner les différents processus de démantèlement du substrat.

Figure 2: Localisation des caméras dans la vallée, avec leur angle de champ et leur cadre. Source : UMI orthomosaic (2010).

Références Allard, M., & Seguin, M. (1985). La déglaciation d’une partie du versant hudsonien québécois: bassins des rivières Nastapoca, Sheldrake et à l’Eau Claire. Géographie physique et Quaternaire, 39(1), 13-24.

Allard, M., & Seguin, M. K. (1987). The Holocene evolution of permafrost near the tree line, on the eastern coast of Hudson Bay (northern Quebec). Canadian Journal of Earth Sciences, 24(11), 2206-2222.

Andrews, J. T. (1968). Postglacial rebound in Arctic Canada: similarity and prediction of uplift curves. Canadian Journal of Earth Sciences, 5(1), 39-47.

ARK (2007). Projet de parc national des Lacs-Guillaume-Delisle-et-à-l’Eau-Claire. État des connaissances. Administration régionale Kativik, Service des ressources renouvelables, de l’environnement, de territoire et des parcs, Section des parcs, Kuujjuaq, Québec.

Bégin, C., & Filion, L. (1985). Analyse dendrochronologique d'un glissement de terrain de la région du Lac à l'Eau Claire (Québec nordique). Canadian Journal of Earth Sciences, 22(2), 175-182.

Belzile, M. C. (1984). Les versants rocheux périglaciaires à la presqu’île des Manitounouc Kuujjarapik, Nouveau-Québec. Mémoire de maîtrise. Département de géographie, Université Laval.

Cailleux, A. (1947). L’indice d’émoussé des grains de sable et grès. Revue de Geomorphologie Dynamique, 3, 78-87.

Chandler, F. W. (1988). The early Proterozoic Richmond Gulf Graben, East Coast of Hudson Bay, Quebec (Vol. 362). Geological Survey of Canada.

Chandler, F. W., & Schwarz, E. J. (1980). Tectonics of the Richmond Gulf area, northern Quebec—a hypothesis. Current Research, Part C, Geological Survey of Canada, Paper, 80, 59-68.

Chandler, F. W., & Parrish, R. R. (1989). Age of the Richmond Gulf Group and implications for rifting in the Trans-Hudson Orogen, Canada. Precambrian Research, 44(3-4), 277-288.

Craig, B. G. (1969). Late-glacial and postglacial history of the Hudson Bay region. In Earth Science Symposium on Hudson Bay (Vol. 68, pp. 63-77). Pap. Geol. Surv. Can.

Decaulne, A., Bhiry, N., Lebrun, J., Veilleux, S., & Sarrazin, D. (2018). Geomorphic evidence of Holocene slope dynamics on the Canadian shield–a study from Lac à l’Eau- Claire, western Nunavik. Écoscience, 1-15.

Dionne, J. C. (1976). Les grandes cuestas de la mer d'Hudson. GÉOS (Energie, Mines et Ressources Canada), 5(1), 18-20.

Duhaime, G. (2007). Profil socioéconomique du Nunavik. Chaire Condition Autochtone.

Eaton, D. W., & Darbyshire, F. (2010). Lithospheric architecture and tectonic evolution of the Hudson Bay region. Tectonophysics, 480(1-4), 1-22.

Fortier, R. (2017). Groundwater monitoring network from the Umiujaq region in Nunavik, Quebec, Canada, v. 1.3 (2012-2016). Nordicana D19, doi: 10.5885/45309SL- 15611D6EC6D34E23.

Francou, B. (1988). L’éboulisation en haute-montagne – Andes et Alpes –, six contributions à l’étude du système corniche-éboulis en système périglaciaire. Thèse d’État, Université Paris, 7.

Fraser, C., Hill, P. R., & Allard, M. (2005). Morphology and facies architecture of a falling sea level strandplain, Umiujaq, Hudson Bay, Canada. Sedimentology, 52(1), 141-160.

Germain, D. (2016). Snow avalanche hazard assessment and risk management in northern Quebec, eastern Canada. Natural Hazards, 80(2), 1303-1321.

Guimont, P., & Laverdiere, C. (1980). Le sud-est de la mer d'Hudson: un relief de cuesta. The Coastline of Canada, SB McCann (edit.), Geological Survey of Canada, Paper, 80-10.

Hétu, B., & Gray, J. T. (2000). Effects of environmental change on scree slope development throughout the postglacial period in the Chic-Choc Mountains in the northern Gaspé Peninsula, Québec. Geomorphology, 32(3-4), 335-355.

Hétu, B., Brown, K., & Germain, D. (2008). L’inventaire des avalanches mortelles au Québec depuis 1825 et ses enseignements. In Proceedings of the 4th Canadian conference on geohazards, Université Laval, Québec (pp. 20-24).

Hillaire-Marcel, C. (1976). La déglaciation et le relèvement isostatique sur la côte est de la baie d’Hudson. Cahiers de géographie du Québec, 20(50), 185-220.

Krumbein, W. C. (1941). Measurement and geological significance of shape and roundness of sedimentary particles. Journal of Sedimentary Research, 11(2), 64-72.

Lajeunesse, P., & Allard, M. (2003). The Nastapoka drift belt, eastern Hudson Bay: implications of a stillstand of the Quebec–Labrador ice margin in the Tyrrell Sea at 8 ka BP. Canadian Journal of Earth Sciences, 40(1), 65-76.

Lavoie, C., Allard, M., & Duhamel, D. (2012). Deglaciation landforms and C-14 chronology of the Lac Guillaume-Delisle area, eastern Hudson Bay: a report on field evidence. Geomorphology, 159, 142-155.

Lied, K., & Domaas, U. (2000). Avalanche hazard assessment in Nunavik and on Côte- Nord, Québec, Canada. Norwegian Geotechnical Institute.

Luckman B.H. (2013) Talus Slopes. In: Elias S.A. (ed.), The Encyclopedia of Quaternary Science, vol. 3, pp. 566-573. Amsterdam: Elsevier.

Marion, J., Filion, L., & Hétu, B. (1995). The Holocene development of a debris slope in subarctic Québec, Canada. The Holocene, 5(4), 409-419.

Ménard, É., Allard, M., & Michaud, Y. (1998). Monitoring of ground surface temperatures in various biophysical micro-environments near Umiujaq, eastern Hudson Bay, Canada. In Proceedings of the 7th International Conference on Permafrost. Yellowknife, Canada (pp. 723-729).

Ministère des ressources naturelles et de la faune (2010). Portrait territorial – Nord-du- Québec, 2010. Québec, Gouvernement du Québec.

Payette, S. (1983). The forest tundra and present tree-lines of the northern Québec- Labrador peninsula. Nordicana, 47, 3-23.

Pelletier, M., Allard, M., & Levesque, E. (2018). Ecosystem changes across a gradient of permafrost degradation in subarctic Québec (Tasiapik Valley, Nunavik, Canada). Arctic Science, (0), 1-26.

Percival, J. A. (2007). Geology and metallogeny of the Superior Province, Canada. In Mineral deposits of Canada: A synthesis of major deposit-types, district metallogeny, the evolution of geological provinces, and exploration methods (Vol. 5, pp. 903-928). Geological Association of Canada, Mineral Deposits Division. Special Publication No. 5.

Pérez, F. (1998). Talus fabric, clast morphology, and botanical indicators of slope processes on the Chaos Crags (California Cascades), USA. Géographie physique et Quaternaire, 52(1), 47-68.

Provencher-Nolet, L., Bernier, M., & Lévesque, E. (2014). Quantification des changements récents à l'écotone forêt-toundra à partir de l'analyse numérique de photographies aériennes. Écoscience, 21(3-4), 419-433.

Schneiderhöhn, P. (1954). Eine vergleichende Studie über Methoden zur quantitativen Bestimmung von Abrundung und Form an Sandkörnern (Im Hinblick auf die Verwendbarkeit an Dünnschliffen.). Heidelberger Beiträge zur Mineralogie und Petrographie, 4(1-2), 172-191.

Sellier, D. (1992). Évolution comparée de versants quartzitiques des Highlands d'Écosse et de Norvège centrale (Rates of quartzitic slopes evolution in the scottish Highlands and central Norway). Bulletin de l'Association de Géographes Français, 69(3), 236-241.

St-Cyr, N. (1986). Formation et évolution des versants rocheux des îles centrales du lac à l’Eau Claire, Québec subarctique. Mémoire de maîtrise. Département de géographie, Université Laval.

Chapitre 1

Talus slope characterization in Tasiapik Valley (subarctic Québec): evidence of past and present slope processes

Samuel Veilleux1,2, Najat Bhiry1,2 and Armelle Decaulne3

1Département de géographie, Université Laval, Québec, Canada 2Centre d’études nordiques, Université Laval, Québec, Canada 3Centre national de recherche scientifique, Laboratoire LETG, Université de Nantes, France

Résumé Des relevés topographiques, granulométriques, morphométriques, pétrographiques et de végétation ont été effectués sur les versants de la vallée Tasiapik, près d’Umiujaq (Nunavik) dans le but de documenter les processus gravitaires dominants et leur impact géomorphologique. Les talus d’éboulis font partie intégrante du paysage de la vallée; on les retrouve à la fois sur le versant sud- ouest et nord-est, au pied d’abruptes falaises rocheuses. La lithologie des dépôts de versant témoigne de leur provenance locale, soit de chutes discrètes issues de la paroi adjacente. Localement, la faible couverture de lichens et de mousses sur les talus indique que les débris sont récents, alors qu’ailleurs des couverts arbustifs ont colonisés les talus, démontrant la faible activité de nos jours, concomitante avec une progression récente de la végétation arbustive. Suite à la dernière déglaciation, les processus paraglaciaires ont potentiellement favorisé l’instabilité des versants. Les processus périglaciaires actifs de nos jours induisent un démantèlement important du substrat rocheux, et par conséquent rendent disponible une grande quantité de matériel. L’utilisation de caméras automatiques au cours de l’hiver 2017-2018 a permis d’observer de nombreux événements d’avalanche, mais peu d’événements d’éboulis. Une grande quantité de débris ont été transporté vers les talus au pied du versant par les avalanches printanières, fréquentes en mai et juin 2018. La neige a également permis une redistribution des débris sur les talus.

Mots-clés : talus d’éboulis, processus de versant, cuesta, Nunavik.

Abstract Topographic, granulometric, morphometric, petrographic and vegetation surveys were conducted on the slopes of Tasiapik Valley, near Umiujaq (Nunavik), to document mass wasting processes and their geomorphological impact. Talus slopes, widespread at the foot of the steep rockwalls of Tasiapik Valley, are an important landscape feature in the area. The lithology of the slope deposits attest their local origin, namely the result of rockfalls coming from the adjacent wall. Locally, poor lichens and mosses vegetation covering the clasts exhibits recently fallen debris; elsewhere, dense shrub cover has colonized the slopes demonstrating the low activity nowadays. Following the last deglaciation, paraglacial processes have potentially favoured slope instabilities. On-going periglacial processes have led to extensive dismantling of the rockface, enabling for debris supply. The use of automatic cameras during the winter 2017-2018 resulted in the observation of many snow-avalanche events; however few rockfall events have been observed. Spring snow avalanches have carried rock debris to the talus at the foot of the slope; snow also enabled debris redistribution on the slopes.

Key words: morphometry, slope dynamics, snow avalanches, periglacial, Nunavik.

I. Introduction Northern landscapes have undergone many changes since their deglaciation. In particular, paraglacial conditions (e.g. glacio-isostatic rebound) induced talus slope formation by supplying debris through the pressure release on rock fractures and freeze-thaw processes (Ballantyne and Benn, 1994; Matsuoka and Sakai, 1999; Ballantyne, 2002; Matsuoka, 2008).

Nunavik is part of the low-Arctic region of eastern Canada and its landscape consists of low hills, basins and plateaus. The few studies that have been conducted in this vast region have demonstrated the occurrence of slope processes on slopes less than 100 m high (Belzile, 1984; Bégin and Filion, 1985; St-Cyr, 1986; Marion et al., 1995; Germain and Martin, 2012; Germain, 2016). Recent studies (Decaulne et al., 2018; Bhiry et al., 2019) conducted at Wiyâshâkimî Lake in Tursujuq National Park (Nunavik) showed that talus slope formation started after deglaciation at about 4600 BP, and that slope processes are still active today. Some of the villages in Nunavik (Salluit, Kangiqsujuaq and Kangiqsualujjuaq) are located within glacial valleys with prominent slopes, while other villages (Umiujaq) are situated near high cuesta relief (~230 m). Accordingly, it is crucial to document slope dynamics and to evaluate associated risks on the local population, visitors and infrastructures. For instance, in Kangiqsualujjuaq (northeastern Nunavik), a dreadful snow avalanche hit the gymnasium of Satuumavik school during the 1999 New Year’s Eve celebrations, causing the death of 9 people and injuring 25 (Bérubé, 2000; Lied and Domaas, 2000; Germain, 2016). However, no extensive research has been conducted on slope processes in the Umiujaq area (including snow avalanches, landslide and rockfalls), their triggering factors, their occurrence and their runout distance. General conditions are conducive for bedrock dismantling and mass wasting, even with limited slope heights, but additional knowledge about slope processes is still required.

The main objective of the study is to document landforms organisation built by gravitational processes in Tasiapik Valley and their contribution to talus development based on geomorphological surveys. This study discusses slope evolution during the Holocene, from the retreat of the Laurentide Ice Sheet in the area to the present-day, highlighting the potential risk at the valley bottom.

II. Study area Tasiapik Valley (56°33' N, 76°28' W) is located 5 km east of the Inuit village of Umiujaq, on the east coast of Hudson Bay in Nunavik, Québec (Figure 3a). It is approximately 4.5 km long and 1.5 km wide, following a northwest-southeast orientation. At the southeastern end of the valley lies Tasiujaq Lake (formerly named Guillaume-Delisle Lake or Richmond Gulf), a 691 km2 brackish water body connected to the Hudson Bay by a narrow cataclinal channel called Le Goulet (ARK, 2007). The lake is part of Tursujuq National Park, created in 2013.

The regional geology is characterized by a Paleoproterozoic volcano-sedimentary sequence lying unconformably on the Precambrian shield (Figure 3b). The volcano-sedimentary sequence includes limestone, quartz arenite, dolomite and sandstone strata (Qingaaluk Formation) underlying a thick (~15 m) basalt layer (Nastapoka Group) dipping westward (Stockwell et al., 1979; Chandler and Schwarz, 1980; Chandler, 1988; Eaton and Derbyshire, 2010). This asymmetrical monoclinal relief (cuesta) consists of a gentle western slope and a steep eastern slope and extends over 650 km along the east coast of Hudson Bay (Dionne, 1976; Guimont and Laverdière, 1980). Tasiapik Valley lies at the frontslope of the cuesta on its southwestern side, whereas the northeastern side consists of a residual butte called Umiujaaluk Hill.

Quaternary deposits on the east coast of Hudson Bay are the result of a succession of sedimentary environments following the retreat of the Laurentide Ice Sheet at about 8200 cal. BP (Hillaire-Marcel, 1976; Allard and Séguin, 1985; Lavoie et al., 2012) (Figure 3b). Lowlands below 271 m a.s.l, the altitudinal limit of the postglacial Tyrrell Sea in Tasiujaq Lake area (Fraser et al., 2005; Lavoie et al., 2012), are covered by deep-water and shallow-water marine sediments, and littoral deposits (raised beaches) that were formed during stages of rapid glacio-isostatic uplift (Hillaire-Marcel, 1976). A glaciomarine fan complex lies in the upstream part of the valley. It consists of fluvioglacial material that was deposited during a stillstand of the ice margin around 8000 cal. BP (Lajeunesse and Allard, 2003b).

The study area has a cold subarctic climate and it is located in the discontinuous permafrost zone (Allard and Lemay, 2012). Mean annual air temperature recorded between 2013 and 2017 varies between -5.6 and -4.2°C, with maxima of 23°C and minima of -36°C (Fortier, 2017). Mean annual precipitation is approximately 500 mm, with 40% falling as snow (Ménard et al., 1998). The Umiujaq

area is located at the edge of the shrub and forest tundra zones; low shrubs, ericaceous plants and lichens cover the upstream part of Tasiapik Valley, while dense forest cover occupies the downstream part (Payette 1983). Shrub cover has expanded significantly (shrubification) during the 20th century (Ménard et al., 1998; Provencher-Nolet et al., 2014; Pelletier et al., 2018).

The SW side of Tasiapik Valley has a near-vertical rockwall. It is approximately 50 m high in the upstream part of the valley, but it increases to 230 m near Tasiujaq Lake in the downstream part. Slope deposits lie at the base of the escarpment, connecting the rockwall to the valley floor, but have also accumulated on basaltic rocky outcrops in the uppermost part of the rockwall (Figure 3c). The NE side features a step-like topography, with slope deposits either located at the base of the slope or perched on basaltic and sedimentary rocky outcrops. A gravel road connecting Umiujaq to Tasiujaq Lake follows the cuesta frontslope on the SW side.

Figure 3: Location of Tasiapik Valley within the Umiujaq area (A); regional geology and quaternary sediments in Tasiapik Valley (B); distribution of the talus slopes and investigated slopes (C). Sources: MRNF (A, B), UMI orthomosaic (2010) (C).

III. Methods For this study, 18 talus slopes were investigated, on the SW and NE sides of Tasiapik Valley (Figure 3c). Data were collected over four field campaigns during the summers of 2016 (August), 2017 (August) and 2018 (June and August). Several slope deposits were identified by satellite imagery prior to initiating fieldwork. Two sets of orthophotos from Québec’s Ministère des Ressources naturelles et de la Faune (MRNF) were used, one dating from 2004 (scale 1/10,000, 25 cm resolution) and the other from 2010 (scale 1/10,000, 15 cm resolution).

Topographic surveys were conducted along 18 longitudinal transects using a Leica DGPS (Differential Global Positioning System). Waypoints were recorded from the apex of the slope deposits to their base, perpendicular to the rock face. Data were processed in ArcGIS and Excel to produce topographic profiles, revealing the microtopographic features such as inflection and texture in accurate details. The estimation of the stage of evolution is carried out from the Ho/Hi ratio, where Ho corresponds to the height of the talus slope and Hi to the total height of the slope including the rockwall (Francou, 1988; Sellier, 1992). The ratio gives an overview of the exhaustion of the remaining rockwall (debris source) in concomitance with the talus slope formation on a longer timescale, namely since the last deglaciation. For example, a ratio approaching 1 indicates an advanced slope development stage due to the low height of the residual wall compared to the height of the talus slope.

Topographic data and satellite imagery were used to measure rockfall runout distances (horizontal travel distance) calculated from the source area to the farthest slope debris. In addition, the reach angle, calculated from the source-area to the farthest slope debris, and the shadow angle, calculated from the apex of the talus to the farthest slope debris, provide information about the extent of slope processes in the area.

On 12 of the 18 longitudinal transects, granulometric and petrographic surveys were conducted by sampling 25 randomly selected rock fragments at intervals of 10-15 m along the transects. Debris were measured along three axes: length (a-axis), width (b-axis) and thickness (c-axis). Measurements were compiled in Excel and then analyzed to produce descriptive statistics. Morphometric indices were also calculated from these measures (Pérez 1998, Hétu and Gray 2000).

The flattening index (Fi) is calculated as follows:

푎 + 푏 퐹 = 푖 2푐 where a corresponds to the length, b to the width and c to the thickness of the fragment (Cailleux,

1947). A high Fi value indicates that the debris has a flatter shape. The elongation index (Li) is calculated as follows:

푎 퐿 = 푖 푏 where a and b correspond to the length and width of the fragment (Schneiderhöhn, 1954). A high Li value indicates that the debris tends to be elongated. Finally, the sphericity index (Si) is calculated as follows:

1 푏푐 3 푆 = ( ) 푖 푎2 where a, b and c correspond to the length, width and thickness of the fragment (Krumbein, 1941). A value Si approaching 1 indicates that the debris has a more massive shape, spherical in the case of a rounded fragment and cubic for an angular fragment. These indices document the falling behavior of clasts, since spherical debris are prone to rolling, while elongated flat debris are more likely to slide. Petrographic surveys provide lithological data for the measured fragments. Their origin, either local (associated with the local slope development) or exogenous (generally from glacial transport and deposition), is closely related to their lithology, thus their general shape, and their position on the slope. The edges of the debris were characterized, with a view to determining their origin: an angular fragment has undergone very little erosion, indicating a short transportation distance/local source, while debris transported by glaciers or reworked in the Tyrrell Sea has a pronounced rounded shape.

Vegetation cover was described at each sampling station in order to assess recent and current process activity. Hierarchical values were attributed to each station based on the type of vegetation and the estimated percentage of coverage on the debris, providing relative age-estimates:

1) Fresh debris: no lichen species observed on the clast;

2) Recent debris: some lichen species observed on the clast;

3) Medium-aged debris: several lichen species partially cover the clast;

4) Old-aged debris: several species of lichens and mosses partially cover the clast;

5) Very old-aged debris: several species of lichens and mosses totally cover the clast; potentially also covered with low shrubs.

Vegetation classification values and Ho/Hi ratio values were used to estimate the stage of slope development. The addition of these two values gives an overview of the slope evolution from both short term (vegetation) and long term (Ho/Hi) perspectives. Values of 1 to 5 were assigned to each longitudinal profile according to their Ho/Hi ratio, following the Jenks natural breaks classification method (Jenks, 1967); a value approaching 1 indicates a low Ho/Hi ratio, thus a younger development stage. The vegetation values (i.e. the lowest - freshest - value per profile, ranked 1 to 5 according to the relative age estimate described above) were added to provide an overall development score. In addition, the age of shrubs at the bottom of talus along the SW-07, SW-08 and SW-09 profiles was determined using dendrochronology on 11 black spruce samples (Picea mariana).

Finally, in order to monitor slope movements on a shorter time scale, three Reconyx PC800 automatic time-lapse cameras were installed on the SW side of the valley in August 2017. One of the cameras (TAS 1) is located on the cuesta edge, above the rockwall and talus along the SW-07 and SW-08 profiles (Figure 4). The latter is covered by a second camera (TAS 2) that is located 300 m away from the rockwall. A third camera (TAS 3) is located further upstream near the base of the talus along the SW-06 profile. Approximately 14,000 photos were taken over a one-year period from August 2017 to August 2018 in the valley. Photos were taken during daytime at one-hour intervals until June 2018, then at 15 or 30 minutes intervals (depending on the location) until August 2018.

Figure 4: Location of the cameras along the SW side of Tasiapik Valley. Field of view of each camera is shown on the left. Source: UMI orthomosaic (2010).

IV. Results and interpretation

1. Topography of slope deposits The SW profiles show a steeper slope gradient, with a mean angle of 25.3° and a median angle of 26.3°, while the NE profiles have a mean angle of 21.2° and a median angle of 17.5° (Table 1; Figure 5). A slight or distinct concavity of the talus slope is apparent on six of the nine SW profiles, while the SW-01 and SW-06 profiles are virtually linear. The SW-09 profile shows a more complex shape (linear proximal part and chaotic distal part). On the NE side, there is no distinct concave profile, yet seven of the nine profiles show either a slightly concave or a linear shape, while the NE- 04 and NE-07 profiles respectively show a complex and a convex shape. Four profiles (SW-07, SW- 08, SW-09 and NE-09) exhibit a strong basal concavity (Figure 5).

Table 1: Topographic parameters of the investigated slopes.

Profiles Slope angle (°) Mean angle (°) Median angle (°) Inflection SW-01 18.4 linear SW-02 28.1 slightly concave SW-03 28.3 slightly concave SW-04 19.3 concave SW-05 24.9 25.3 26.3 concave SW-06 27.1 linear SW-07 26.3 concave SW-08 30.0 concave SW-09 25.2 complex NE-01 14.2 linear NE-02 12.3 slightly concave NE-03 17.5 slightly concave NE-04 16.1 complex NE-05 17.4 21.2 17.5 linear NE-06 31.8 linear NE-07 27.1 convex NE-08 23.2 linear NE-09 31.6 slightly concave

Figure 5: Longitudinal cross-section of the investigated slopes. Source: UMI orthomosaic (2010).

2. Relative dating of slopes The mean Ho/Hi index on the SW (0.36) and NE sides (0.25) indicates that the remaining rockwall is generally higher than talus slopes (index < 0.5) (Table 2). However, the step-like topography on the NE side, compared to the near-vertical rockwall on the SW side, could mean that the NE side has reached a more advanced stage of development. Talus slopes near the southern margin of both sides (along the SW-07, SW-08, SW-09 and NE-09 profiles) are located under high vertical rockwalls with Ho/Hi index below 0.2, thus indicating a younger development stage. The SW-04 and SW-05 profiles indicate an older stage than the other SW profiles, with respective index values of 0.56 and 0.62.

Examination of the vegetation covering the surficial debris on the slope deposits revealed the presence and position of few fresh deposits. Most of the debris had a clear vegetation cover on the SW side (Figure 6). In the apical parts of the talus slopes, various lichens and/or mosses are abundant, while a discontinuous thin strip of herbaceous plants and low shrubs is located at the edge of the rockwall undisturbed by present-day slope activity. Distal parts also feature abundant lichens and mosses on most of the clasts along with thick mosses covering the slope deposits and low shrubs; this trend is especially evident on the SW-07 and SW-09 profiles. A similar trend was observed on the NE profiles, as most sampling stations in the apical part of the talus slopes show medium to old-age status, while sampling stations in the distal parts indicate older-age status. However, debris along the SW-08, NE-06 and NE-09 profiles appear to be more recent, with little overall coverage and the presence of few lichens. Some fresh debris were scattered along most of the profiles.

Table 2: Calculation of Ho/Hi index.

Talus base Talus apex Ho Rockwall elev. Ho/Hi Profiles Hi value elev. (m) elev. (m) value (m) index SW-01 171.3 187.6 16.3 222 50.7 0.321 SW-02 147.3 178.3 31.0 216 68.7 0.451 SW-03 146.3 176.8 30.5 208 31.7 0.494 SW-04 151.4 183.2 31.8 208 56.6 0.562 SW-05 148.0 187.9 39.9 212 64.0 0.623 SW-06 127.4 159.2 31.8 212 84.6 0.376 SW-07 46.5 82.0 35.5 214 167.5 0.212 SW-08 47.6 67.2 19.6 222 174.4 0.112 SW-09 49.6 75.9 26.3 272 222.4 0.118 NE-01 156.1 169.9 13.8 202 45.9 0.301 NE-02 157.2 167.3 10.1 252 94.8 0.107 NE-03 164.6 182.8 18.2 254 89.4 0.204 NE-04 163.0 184.3 21.3 270 107.0 0.199 NE-05 162.1 178.1 16.0 282 119.9 0.133 NE-06 225.6 260.8 35.2 282 54.4 0.624 NE-07 131.7 167.5 35.8 292 160.3 0.223 NE-08 134.6 184.0 49.4 302 167.4 0.295 NE-09 161.2 188.8 27.6 312 150.8 0.183

Figure 6: Vegetation stages along the talus slopes. By adding up the Ho/Hi index values and the vegetation classification values, we can estimate the developmental stage of the slopes. As shown on Figure 7, the SW-08 and NE-09 profiles seem to be at the youngest developmental stage among all the talus slopes. The Ho/Hi index and the vegetation classification values are consistent for some profiles, showing a concomitance for both parameters. For example, the talus slope along NE-09 profile has a low Ho/Hi index (0.183) and there is very poor lichen cover on the debris. However, the two parameters proved to be contradictory for some talus slopes, particularly for the SW-09 profiles, due to the overly high rockwall (increased debris supply potential) and the presence of well-developed vegetation (limited debris supply on the talus slope). Both of these findings indicate that the recent debris supply is sporadic.

Figure 7: Slope development stage according to three parameters: 1) vegetation; 2) Ho/Hi index and 3) a combination of the two. Source: UMI orthomosaic (2010).

3. Source and morphometry of slope deposits Three classes of debris were identified along the investigated profiles: 1) basalt, 2) sedimentary rocks (comprising dolomite, limestone, quartz arenite and sandstone), and 3) granitic gneiss (Figure 8).

On the SW side, sedimentary rock debris comprised 64.5% of the sampled clasts, whereas basalt debris accounted for 35.4% and gneiss for 0.07%. Sedimentary rock debris represented a larger proportion on the NE side, accounting for 89.4% of all clasts, while basalt and gneiss accounted for 6.4% and 4.2%. Assuming the top basalt layer is ~15 m thick throughout the valley; those values coincide with the large proportion of sedimentary rock strata available for debris supply on the exposed rockwall. Sedimentary rock strata account for 90% (~140 m) of the rockwall (~155 m) above the SW-07, SW-08 and SW-09 profiles. In the upstream part of the SW side, sedimentary rock strata account for 57% (~20 m) of the rockwall (~35 m). On the NE side, the top basalt layer has considerably receded on most of the investigated slopes, revealing rocky outcrops composed of sedimentary rock strata. However, the rockwall above the NE-06 profile is mainly composed of basalt (68% of total height).

By comparing the respective proportions of each lithology at sampling stations along the profiles, many of the slope deposits show consistent ratios of sedimentary rocks and/or basalt debris from the apex to the base of the slope. On the SW side, the proportion of basalt debris varies between 80% and 84% throughout the sampling stations on the SW-02 profile and between 14% and 26% on the SW-07 profile. On the NE side, the proportion of sedimentary rock debris varies between 80% and 88% on the NE-02 profile, while the NE-08 and NE-09 profiles show no difference as sedimentary rock debris compose 100% of the talus. However, some of the slope deposits show an increasing proportion of basalt material toward the foot of the talus. For example, along the SW-08 profile, the basalt debris percentage increases from 17% at the apex to 37% at the bottom of the slope; on the SW-09 profile, it increases significantly from 12% to 100%. Finally, on the NE-06 profile, the basalt debris are only located at the bottom of the talus, whereas the sedimentary rock debris comprises the entire apical part.

Clasts were measured to document the size and morphometry of the slope deposits, and their distribution along the deposits (Figure 9). On the SW side, the debris (a-axis) range in size from 30 to 154 cm on average (Table 3). However, standard deviation values show important disparities between the measured clasts, as the largest clasts range from 200 to 1900 cm. In addition, mean

flatness index values range from 2.36 to 4.22. Lengthening and sphericity indices do not vary much, ranging from 1.48 to 1.86 and from 0.54 to 0.65. On the NE side, sizes range from 83 to 178 cm on average, while standard deviation values show important disparities between the measured clasts. The largest clasts range in size between 350 and 950 cm, while the mean flatness index values range from 2.29 to 3.58. Lengthening and sphericity indices show similar values as the profiles on the SW side, ranging from 1.40 to 1.82, and from 0.57 to 0.69. By comparing the calculated index values along all profiles, the clasts tend to have a flatter and more elongated shape in the downstream part of the valley on the SW side (SW-07, SW-08 and SW-09), and along the NE-06 and NE-09 profiles. This morphology is more associated with sedimentary rock debris due to the dismantling of the thin sedimentary layers which represents the majority (84.5%) of the sampled debris along these profiles (given the larger proportion of sedimentary rock within the rockwall). For most of the profiles, the mean and median a-axis values tend to increase toward the base of the slopes. As for the morphometry indices, 50% of the profiles show an increasing trend for flatness and sphericity values toward the base of the slope, while the lengthening index of clasts increases from the apex to the base for 75% of the profiles.

Debris from every sampling station on the SW side were analyzed and described as being either angular or subangular, which indicates their local slope-related provenance, whereas the NE side shows a greater diversity of debris shape. For the NE-02 and NE-04 profiles, rounded/sub-rounded debris are abundant at the base and angular debris are found near the apex of the talus slopes, meaning that slope-related debris have accumulated in the proximal part and mixed with rounded heavily reworked debris toward the distal part.

Figure 8: Proportion of the different lithologies along the talus slope.

Figure 9: Distribution of the debris size (a-axis) along the talus slopes. Close-up of the largest sampled debris along SW-09 profile (A) and NE-09 profile (B). Source: UMI orthomosaic (2010).

Table 3: Debris size and morphometric parameters along the longitudinal profiles.

Profiles a-axis (cm) b-axis (cm) c-axis (cm) b/a c/b Flatness Lengthening Sphericity

median 45.00 32.00 19.00 0.70 0.63 1.96 1.43 0.65

mean 65.54 45.74 27.54 0.69 0.64 2.36 1.53 0.65

01 standard dev. 58.57 44.99 26.57 0.16 0.22 1.24 0.40 0.12

- maximum 450.00 350.00 180.00 0.98 1.00 10.42 2.73 0.93

minimum 11.00 7.00 3.00 0.37 0.11 1.10 1.03 0.38

range 439.00 343.00 177.00 0.61 0.89 9.31 1.70 0.55

median 60.00 37.00 21.00 0.63 0.64 2.10 1.58 0.62

mean 78.18 48.93 30.99 0.64 0.63 2.55 1.68 0.62

02 standard dev. 80.30 48.04 35.48 0.17 0.21 2.44 0.51 0.11

- maximum 600.00 350.00 290.00 1.15 1.04 27.50 3.64 0.87

minimum 8.00 5.00 1.00 0.28 0.05 1.17 0.87 0.26

range 592.00 345.00 289.00 0.88 0.99 26.33 2.77 0.61

median 84.00 50.00 23.00 0.67 0.57 2.21 1.50 0.60

mean 154.48 95.20 60.77 0.67 0.58 2.84 1.62 0.62

03 standard dev. 262.49 141.39 108.40 0.18 0.24 1.88 0.49 0.13

- maximum 1900.00 1000.00 800.00 1.00 1.11 13.75 3.32 0.91

minimum 8.00 7.00 3.00 0.30 0.10 1.07 1.00 0.32

range 1892.00 993.00 797.00 0.70 1.01 12.68 2.32 0.59

median 50.00 36.00 17.00 0.72 0.58 2.21 1.38 0.61

mean 77.14 54.89 29.70 0.72 0.56 2.69 1.48 0.64

06 standard dev. 78.61 55.69 34.26 0.17 0.24 1.36 0.45 0.13

- maximum 470.00 310.00 190.00 1.04 1.00 7.50 3.86 0.93

minimum 5.00 5.00 3.00 0.26 0.15 1.06 0.96 0.33

range 465.00 305.00 187.00 0.78 0.85 6.44 2.90 0.60

median 23.50 14.00 6.00 0.63 0.50 3.00 1.58 0.53

mean 30.08 17.27 8.24 0.62 0.89 4.22 1.81 0.54

07 standard dev. 24.17 11.83 7.91 0.18 1.33 3.78 0.76 0.13

- maximum 206.00 100.00 70.00 1.00 10.00 29.50 6.80 0.94

minimum 6.00 2.00 1.00 0.15 0.04 1.10 1.00 0.21

range 200.00 98.00 69.00 0.85 9.96 28.40 5.80 0.73

median 36.00 20.00 11.00 0.59 0.56 2.57 1.69 0.56

mean 48.16 27.31 14.29 0.60 0.57 2.96 1.86 0.57

08 standard dev. 42.23 26.74 12.83 0.18 0.22 1.53 0.80 0.12

- maximum 280.00 223.00 100.00 1.00 1.00 11.17 10.42 0.92

minimum 4.00 3.00 1.00 0.10 0.09 1.07 1.00 0.19

range 276.00 220.00 99.00 0.90 0.91 10.10 9.42 0.73

median 25.00 16.00 8.00 0.67 0.54 2.56 1.50 0.58

mean 69.45 37.23 21.53 0.66 0.54 3.28 1.67 0.59

09 standard dev. 189.69 89.36 53.67 0.18 0.25 2.38 0.61 0.13

- maximum 1700.00 700.00 400.00 1.00 1.00 20.00 5.36 0.92

SW minimum 5.00 4.00 1.00 0.19 0.07 1.11 1.00 0.22

range 1695.00 696.00 399.00 0.81 0.93 18.89 4.36 0.70

median 87.50 67.00 40.00 0.74 0.61 1.99 1.35 0.68

mean 146.87 108.63 66.67 0.78 0.61 2.32 1.40 0.69

standard dev. 150.64 110.64 78.16 0.35 0.21 1.42 0.38 0.14

NE maximum 630.00 580.00 420.00 3.18 0.96 11.00 2.70 1.24

minimum 20.00 11.00 3.00 0.37 0.11 1.20 0.31 0.39

range 610.00 569.00 417.00 2.81 0.85 9.80 2.39 0.85

median 85.00 55.00 34.00 0.67 0.61 2.07 1.50 0.63

mean 107.02 72.16 43.72 0.69 0.62 2.29 1.55 0.65

standard dev. 88.58 68.70 41.95 0.17 0.20 0.90 0.41 0.11

NE maximum 610.00 590.00 340.00 1.18 1.50 7.25 2.88 0.91

minimum 12.00 9.00 3.00 0.35 0.18 1.18 0.85 0.40

range 598.00 581.00 337.00 0.83 1.32 6.07 2.03 0.51

median 60.00 40.00 20.00 0.67 0.59 2.37 1.50 0.57

mean 83.00 47.47 25.97 0.64 0.57 3.17 1.82 0.60

standard dev. 72.94 39.96 22.93 0.21 0.24 2.31 0.90 0.17

NE maximum 350.00 250.00 120.00 1.00 1.00 13.00 5.83 0.95

minimum 10.00 8.00 4.00 0.17 0.11 1.08 1.00 0.21

range 340.00 242.00 116.00 0.83 0.89 11.92 4.83 0.74

median 75.00 45.00 25.00 0.67 0.63 2.17 1.50 0.60

mean 87.77 54.73 32.89 0.66 0.61 2.91 1.66 0.62

standard dev. 76.05 43.35 30.24 0.19 0.25 2.59 0.57 0.14

NE maximum 450.00 270.00 150.00 1.00 1.20 17.50 3.60 1.00

minimum 9.00 5.00 2.00 0.28 0.07 1.00 1.00 0.31

range 441.00 265.00 148.00 0.72 1.13 16.50 2.60 0.69

median 52.00 34.50 16.50 0.65 0.50 2.72 1.54 0.54

mean 92.89 54.51 27.47 0.64 0.53 3.58 1.79 0.57

standard dev. 124.67 75.11 45.60 0.24 0.27 3.53 0.74 0.14

NE maximum 950.00 640.00 420.00 1.94 1.67 32.63 4.03 1.31

minimum 11.00 6.00 2.00 0.25 0.05 1.17 0.51 0.23

range 939.00 634.00 418.00 1.70 1.62 31.46 3.51 1.09

4. Debris runout Several scattered clasts are located in the distal parts of the slopes. These are mostly located on the SW side and some of them were deposited only a few meters from the road. Their angular shape and lithology (mostly basalt) indicate their local slope-related provenance, by contrast with rounded fluvioglacial/glacial debris.

According to Corominas et al. (2003), the maximum reach angle for small-scale (1-10m3) rockfalls on an unobstructed path is 48°. Therefore, material falling from the uppermost source area in the rockwall (basalt layer) would theoretically be transported no further than the talus at the bottom of the slope, where the terrain levels out and the shrub vegetation is often denser (Figure 10). This perfectly matches the very large boulder accumulation on the SW-09 profile. However, the reach angle for the farthest basalt debris, located beyond the talus slopes, ranges from 24° to 40°. These lower reach angles show a greater horizontal displacement of fallen debris, resulting from either a large-scale rockfall (10-100 m3 for a 40° reach angle; 100-1000 m3 for a 33° reach angle; >1000 m3 for a 26° reach angle) (Corominas et al., 2003) or an external process. For example, there are a dozen large basalt boulders (a-axis > 100 cm) located in the distal part of the SW-07 profile at a 30° angle. However, the volume is not sufficient for such a displacement to result from a rockfall. The shadow angle ranges from 10° to 25°, with the lowest angles measured on the SW-07 (10°) and SW-09 (17°) profiles. These angles are lower than the minimum rockfall shadow angle ranging from 22° to 30° (Rapp, 1960; Govi, 1977; Lied, 1977; Hungr and Evans, 1988; Evans and Hungr, 1993). Therefore, it can be assumed that slope debris were deposited beyond the rockfall runout zone by another process. The hypothetical travel distance for the farthest debris ranges from 105 to 318 m, with the longest distances measured in the downstream part of the SW side.

In June 2018, numerous dirty snow-avalanche deposits were observed on the SW side, but their terminus rarely exceeded the talus slope, thus supporting the assumption that snow-avalanche runout is generally limited to the foot of the talus. However, longer runout was observed for clean snow avalanches occurring in February through April. Extreme snow avalanche runout cannot be documented at this stage in the valley.

Tree-ring counting on 11 black spruce (Picea mariana) samples, dating back to 1900-1930, and located at the foot of the SW-07 and SW-08 profiles, showed clear signs of eccentric growth, such as

the formation of reaction wood during their lifespan. As these reaction wood periods are not significantly concomitant from one tree to another, wind or snow cover could hardly be the main controlling factors for these deformations. Instead, slope processes, such as snow avalanches, are likely causes. However, the limited number of samples does not support a precise chronology of events.

Figure 10: Cross-section of the slopes (talus/rockwall) showing the reach and shadow angle for the farthest slope debris, and the hypothetical travel distance from the source area. Source: MRFP.

5. Short-term slope dynamics Analysis of the photographs from the three automatic cameras has documented the slope dynamics from summer 2017 to summer 2018, with the record of one full winter. Snow avalanches were the main gravitational processes observed, occurring from November 2017 to June 2018; snow- avalanche events occurred more frequently from the end of May 2018. During this period, most of the snow-avalanche deposits exhibited a dirty appearance, because rock debris were incorporated in the dense and humid snow. This characteristic implies a flowing motion with frequent contact with the regolith (Figure 11a). These observations were validated in the field in June 2018, when about 20 wet snow avalanches were observed on the SW side of the valley, from June 11 to June 15.

Along the SW-07 and SW-08 profiles (TAS2), some larger rock debris (~1 m a-axis) were transported and deposited downslope by snow avalanches. Other debris were observed falling onto the snow- covered talus until early July 2018, concomitant with snow-avalanche events. Two significant rockfalls occurred one-hour apart on June 30 2018. Several debris could be observed sliding on the snow covered talus after their fall; they travelled ~30 m before settling at mid-slope (Figure 11b). During the same period, only a few small snow avalanches occurred in the area covered by camera TAS3, along the SW-06 profile, and no rockfall were observed. Out of the seasonal presence of snow, no movement was observed on the slopes from the analysis of the photographs.

Based on the observations during 2017-2018, slope movements occur more frequently in the spring. Prior to June 2018, no rocky deposits on the snow cover had been observed following snow- avalanches, which occurred sporadically from December 2017 to April 2018.

Figure 11: Snow-avalanche event exhibiting a dirty deposit, as the result of incorporated debris (A); discrete rockfall event and debris sliding onto the snow-covered talus (B).

V. Discussion

1. Talus slope formation Results from the morphometric and petrographic surveys suggest an accumulation of rock debris on the SW and NE sides resulting from successive discrete rockfalls that formed into scree slopes. The vast majority of the sampled debris is from a local source, namely the lithologies exposed on the rockwall, and they exhibit subangular to angular shapes.

Vegetation cover on the talus suggests that the most recent debris supply occurred along the SW-08, NE-06 and NE-09 profiles, as the debris were characterized by low coverage and few or no species of lichens. The debris likely originated from recent notches formed in the rockwall above the SW-08 (Figure 12a) and NE-09 profiles (Figure 12b). Along the other profiles (SW-01, SW-09, NE-02, NE- 07), the freshest debris are found in the proximal part of the talus, presumably supplied by small- scale rockfalls. The proximal part of talus slopes generally consists of smaller debris, whereas larger debris and well-developed vegetation are found on the distal part of most of the slopes. Thus, small- scale rockfalls are unlikely to reach the base of the talus, where the accumulation is associated with large-scale rockfalls that have occurred in the past. Nevertheless, debris-size-sorting on the talus slopes, with increasing debris size toward the base, could either be attributable to rockfalls (Kirkby and Statham, 1975; Statham, 1976; Church et al., 1979) or to talus reworking by snow avalanches (Rapp, 1960; Luckman, 1978; Luckman, 1988; Jomelli and Francou, 2000; Decaulne and Saemundsson, 2006). Basalt debris are larger and exhibit a more massive shape than sedimentary rock debris. They also have a higher fall height, resulting in greater travel distances toward the distal part of talus slopes, as observed along the SW-06, SW-08, SW-09, NE-02 and NE-06 profiles.

Nowadays, expanding shrub cover in Tasiapik Valley (Provencher-Nolet et al., 2014; Pelletier, 2015; Pelletier et al., 2018) could suggest a recent decline in debris supply, as rockfalls would not be frequent enough to limit shrub expansion on some slopes. On the NE side, a dense shrub cover (Figure 12c) separates perched and basal talus slopes, while shrubs established themselves at the very base of the talus slope along the SW-07, SW-08 and SW-09 profiles (Figure 12d). In both cases, shrubs have developed over highly weathered rock debris or rocky outcrops. Such a pattern of vegetation colonization highlights areas that lie beyond the reach of the most recent slope dynamics.

Figure 12: Recent notches formed in the rockwall along SW-07 (A) and NE-09 profiles (B); dense shrub cover at the base of the slope near SW-07 and SW-08 profiles (C) and mid-slope near NE-05 and NE-06 profiles (D).

2. Slope debris redistribution Scree slopes generally have a mean angle between 25° and 30°, as reported by Sauchyn (1986) and Francou and Manté (1990). Steeper scree slopes (>30°) have been studied in Québec, namely in Schefferville (Andrew, 1961), in Gaspésie (Andrews, 1961; Hétu, 1995; Hétu and Gray, 2000; Germain and Hétu, 2016) and on the central islands of Wiyâshakimî Lake (Decaulne et al., 2018). Scree slopes also have a segmented profile with a steeper gradient in the proximal part and a strong basal concavity in the distal part; this slope geometry was observed on several talus slopes (Figure 5). Given these findings, debris redistribution must be an ongoing process (Church et al., 1979; Francou and Manté, 1990) that is likely the result of snow-avalanche rework. Their impact results in a concave inflection of the talus slopes and an increase in debris size towards the base, and can be observed on most talus slopes (Rapp, 1960; Luckman, 1977; Luckman, 1978; Luckman, 1988; Jomelli and Francou, 2000; Decaulne and Saemundsson, 2006). The general concavity of talus slopes suggests that surficial debris redistribution is more important than debris supply from the rockwall. Instead, smaller debris tend to be trapped in the numerous surficial cavities that are found on the talus (Statham, 1976; De Blasio and Saeter, 2010).

Photographic monitoring of a portion of the SW side found evidence of discrete rockfall events in June and July 2018, but it has mainly highlighted numerous wet and dirty snow avalanches that occurred in the spring (April-June 2018). The latter have a greater erosive capacity due to higher friction at their base. They can also dislodge rock material from the bare rockwall, thus supplying the talus slope with new debris (Gardner, 1983a; Jomelli, 1999; McClung and Schaerer, 2006). During the same period, the snow-covered talus enabled recently fallen debris to slide down from the apex to mid-slope (~30 m). Sedimentary rock debris are prone to efficient sliding because of their flattened shape (the mean flatness index for sedimentary rock debris (3.04) is higher than basalt debris (2.49) on the investigated slopes), as reported by Pérez (1989) at Lassen Peak, California, and Hétu (1995) in Gaspésie, Québec.

Results of reach angle measurements have shown that scattered slope debris located far downslope originated from large-scale rockfall events along most of the talus slopes, as the angle is greater than 26° (Corominas et al., 2003). Furthermore, low shadow angles, especially along the SW-07 and SW- 09 profiles, provide evidence that their deposition is not entirely due to the rockfall itself, and that the transition from the talus to the valley floor requires an external transport agent (Domaas, 1994). The

lowest measured shadow angle was 17° in Norway (Domaas, 1994), whereas it generally ranges between 22° and 30° (Rapp, 1960; Govi, 1977; Lied, 1977; Hungr and Evans, 1988; Evans and Hungr, 1993). Debris falling onto an ice-covered talus slope are unlikely to be deposited at such a great distance, considering that Evans and Hungr (1993) reported a 24° shadow angle for a boulder falling onto a smooth glacier. In addition, as mentioned above, the current position of the farthest lying boulders does not support the hypothesis of snow-avalanche-related transport. The abrupt slope transition (steep rockwall to valley floor) is not conducive to extreme snow-avalanche runout distances (Bakkehoi et al., 1983; McClung and Schaerer, 2006). From the base of the talus, the slope angle decreases to ~4°, which means that a boulder that is located beyond this zone could not have been transported by a snow avalanche, being located outside of the runout zone.

Therefore, the deposition of debris would most likely be related to the deglaciation/postglacial marine episode, starting at around 8200 cal. BP in the area (Hillaire-Marcel, 1976; Lavoie et al., 2012). In particular, the presence of highly altered basalt boulders (Figure 11) overlying littoral marine sediments would suggest that their deposition occurred during the regression of the Tyrrell Sea. The great distance from the surrounding slopes could then be attributable to sea-ice-related processes, such as ice rafting or ice pushing. Finally, similar processes could have distributed glacially transported rounded boulders against the talus slopes in the lower part of the NE slope, where the vast majority of the subrounded and/or gneiss debris were observed.

Figure 13: Cluster of basalt boulders located ~250 m away from the slope, overlying littoral sediments.

3. Rockwall erosion Different dismantling mechanisms affect the various volcano-sedimentary lithologies exposed on the cuesta frontslope. The top basalt layer shows a distinct columnar polygonal jointing, resulting in the detachment and subsequent fall of large (>5 m a-axis) monoliths (Figure 14a). The edge of the basalt layer on the SW side reveals a sawtooth shape (Figure 14b), also documented by Belzile (1984) at the Manitounuk Peninsula, 100 km south of Umiujaq (55°42' N, 77°07' W). This pattern highlights the numerous monolith falls that occurred in the past. Ongoing periglacial processes such as gelifraction and frost heave caused extensive basalt dismantling (Figure 14c). Michaud and Dionne (1987) observed similar periglacial weathering in the basalt bedrock about 45 km south of Umiujaq (56°09' N, 76°36' W), resulting in block field development. Exhaustion of the underlying sedimentary rock layers could also cause the basalt layer to overhang the slope, eventually leading to rockfalls (Figure 14d).

Figure 14: Erosion of the basalt layer at the top of the cuesta on the SW side, with tall monolith on the verge of falling (A), resulting in a sawtooth shape exposed on the rockwall (B); frost heave and extensive jointing of the basalt bedrock (C); basalt and sandstone layers overhanging quartz arenite layers on the NE side (D). Photos: Decaulne, 2017.

The underlying sedimentary rock layers are prone to frost shattering due to their thin-bedded sub- horizontal structure and abundant fractures exposed on the rockwall, which means that these layers are also subject to rockfalls (Frayssines and Hantz, 2006; Mateos et al., 2012; Letortu, 2013; D’Amato et al., 2016). Notches have recently developed in the sedimentary rock layers, as shown on Figure 12a-12b, and have formed prominent debris cones. Subsequently, these notches will induce more discrete rockfalls from the overlying layers. Locally, rocky outcrops reveal small escarpments (~1-5 m high) that supply debris for nearby talus through short fall height and distances, especially on the NE side.

Freeze-thaw weathering is known to be a major factor in rockfall triggering, especially on deglaciated rockwalls (Matsuoka and Sakai, 1999; Ballantyne, 2002; Matsuoka, 2008). The air temperature data record for the Umiujaq region (CEN, 2018), from September 2002 to June 2018, suggests that freeze-thaw cycles are on average more frequent during the spring (27 - March to June) than in autumn (19 – September to December), but with important interseasonal differences. For instance, only 9 cycles occurred in autumn 2007, followed by 34 cycles during the next spring. However, according to Hales and Roering (2007), the frequency of freeze-thaw cycles would not be a determining factor in rockfall triggering. Instead, they propose a temperature threshold (-3°C to -8°C) whereby thin water layers can infiltrate and cause segregated ice to form in the fractures, leading to the subsequent dismantling of the bedrock. The air temperature data from October 2002 to May 2018 (total of 5688 days) suggests that there were 642 days that had a mean temperature between -3 and –8°C. They are more frequent in November (31.8%), April (20.9%), December (15.7%) and May (13.9%). At a depth of 100 cm in the basalt bedrock, the number of days within this threshold temperature range were more frequent in April (39.5%), December (35.2%) and January (11.9%) between 2001 and 2006 (Allard et al., 2016), which shows a potential delay from autumn to winter compared to air temperature. It should be noted that slope orientation and altitude could influence the temperature at the surface of the rockwall, and that significant differences could be expected between the SW (NE facing) and NE sides of the valley. Nevertheless, the available freeze-thaw cycle data and temperature threshold data concur with the seasonal trend (autumn/spring) for potential rockfall activity (Gardner, 1983b; Matsuoka and Sakai, 1999; Hales and Roering, 2007; Matsuoka, 2008).

On a longer time scale, paraglacial adjustment is another rockfall-triggering factor that should be considered on deglaciated slopes (Ballantyne and Benn, 1994; Ballantyne, 2002; Kellerer-Pirklbauer et al., 2010; Cossart et al., 2013; Grämiger et al., 2017). Deglaciation occurred at about 8200 cal. BP in the study area and induced rapid rates of isostatic uplift (Hillaire-Marcel, 1976; Lajeunesse and Allard, 2003a; Lavoie et al., 2012). This in turn caused an adjustment of the rock substratum and led to stress-release on the rock slopes resulting from the ice retreat, also known as ‘debuttressing’ (Ballantyne, 2002). The rockwall became susceptible to slope failures as a result of these changes.

Slope development results (Ho/Hi index) suggested that the SW side of the valley is at a younger stage of development, considering the height of the remaining rockwall compared to the step-like topography of the NE side. However, well-developed vegetation on the debris observed throughout the SW profiles (except for the SW-08 profiles) raises questions about the spatial distribution and frequency of the slope processes. The exhaustion model, in which slope failures occurring after ice retreat gradually decline until complete exhaustion of the rockwall (Cruden and Hu, 1993), would not fit in this case since local characteristics control the triggering of slope processes (i.e. periglacial processes). A bimodal/multimodal model (Krautblatter and Leith, 2015) would thus be more representative considering that the region has undergone periods of climate fluctuations since its deglaciation (i.e. Little Ice Age). However, the occurrence of extreme meteorological events, such as warm summers (Gruber et al., 2003) and heavy rainfalls (Rapp and Strömquist, 1976; Delonca et al., 2014) should also be considered, as these events can trigger rockfalls and contribute to the slope development. Furthermore, a winter/spring period with many snow avalanches could result in increased erosion on the slope, thus favoring rockfalls, and cleaning available debris accumulated in the rockwall chutes.

VI. Conclusion In this study, we provided evidence of slope activity resulting in talus formation in Tasiapik Valley. The results of the topographic, granulometric, morphometric, petrographic and vegetation surveys suggest that talus slopes throughout the study area are at different stages of development, with some ancient and recent slope deposits. In the field, this results in steep and concave talus slopes with fresh debris, but also coarse deposits with an openwork texture showing alteration and a more developed vegetation cover. In addition, the opposite SW and NE sides exhibit significant differences with respect to most of the parameters surveyed, which means that their evolution occurred at different time scales.

Following the last deglaciation, paraglacial adjustment could have enabled slope failures, initiating talus slope formation. Nowadays, evidence of periglacial processes (gelifraction, frost heave) that have interacted with extensive bedrock jointing has been highlighted. As a result, features such as monoliths that have detached from the basalt layer and are on the verge of falling, or notches within the sedimentary rocks, show that rockfalls could occur at any given time. However, dense vegetation cover, primarily in the form of dense shrubs, has developed on some of the slopes and at the base of the talus slopes, meaning that the slope process runout is limited at the present time.

The present-day slope dynamic has been documented with the use of automatic cameras over a one-year period during the 2017-2018 winter season. The reworking of snow avalanches on slope deposits appeared to be a significant factor in the redistribution of debris, especially in the spring when the wet snow avalanche deposits mostly consisted of dirty snow. Discrete rockfalls have occurred during the same period, and some of the fallen debris have been transported down the snow-covered talus slope.

References Allard, M., & Lemay, M. (2012). Nunavik and Nunatsiavut: From science to policy: An Integrated Regional Impact Study (IRIS) of climate change and modernization. ArcticNet Inc., Québec, Canada. 303 pp.

Allard, M., & Séguin, M. (1985). La déglaciation d’une partie du versant hudsonien québécois: bassins des rivières Nastapoca, Sheldrake et à l’Eau Claire. Géographie physique et Quaternaire, 39(1), 13-24.

Allard, M., Sarrazin, D., & L’Hérault, E. (2016). Borehole and near-surface ground temperatures in northeastern Canada, v. 1.4 (1988-2016). Nordicana D8, doi: 10.5885/45291SL-34F28A9491014AFD.

Andrews, J. (1961). The development of scree slopes in the English Lake District and central Quebec-Labrador. Cahiers de géographie du Québec, 5(10), 219-230.

Bakkehoi, S., Domaas, U., & Lied, K. (1983). Calculation of snow avalanche runout distance. Annals of Glaciology, 4, 24-29.

Ballantyne, C. K. (2002). Paraglacial geomorphology. Quaternary Science Reviews, 21(18-19), 1935-2017.

Ballantyne, C. K., & Benn, D. I. (1994). Paraglacial Slope Adjustment and Resedimenfation following Recent Glacier Retreat, Fåbergstølsdalen, Norway. Arctic and Alpine Research, 26(3), 255-269.

Bégin, C., & Filion, L. (1985). Analyse dendrochronologique d'un glissement de terrain de la région du Lac à l'Eau Claire (Québec nordique). Canadian Journal of Earth Sciences, 22(2), 175-182.

Belzile, M. C. (1984). Les versants rocheux périglaciaires à la presqu’île des Manitounouc Kuujjarapik, Nouveau-Québec. Mémoire de maîtrise. Département de géographie, Université Laval.

Bérubé, J. (2000). Rapport d’enquête publique sur les circonstances des décès survenus à Kangiqsualujjuaq, Nouveau-Québec le premier Janvier 1999. Gouvernement du Québec, Bureau du coroner.

Bhiry, N., Decaulne, A., & Bourgon-Desroches, M. (2019). Development of a subarctic peatland linked to slope dynamics at Lac Wiyâshâkimî (Nunavik, Canada). The Holocene, 29(9), 1459-1467.

Cailleux, A. (1947). L’indice d’émoussé des grains de sable et grès. Revue de Geomorphologie Dynamique, 3, 78-87.

CEN (2018). Climate station data from the Umiujaq region in Nunavik, Quebec, Canada, v. 1.6 (1997-2018). Nordicana D9, doi: 10.5885/45120SL-067305A53E914AF0.

Chandler, F. W. (1988). The early Proterozoic Richmond Gulf Graben, East Coast of Hudson Bay, Quebec (Vol. 362). Geological Survey of Canada.

Chandler, F. W. & Schwarz, E. J. (1980). Tectonics of the Richmond Gulf Area, Northern Quebec – A Hypothesis. In Current Research, Part C, Paper 80-1C (p. 59-68). Geological Survey of Canada.

Church, M., Stock, R. F., & Ryder, J. M. (1979). Contemporary sedimentary environments on Baffin Island, NWT, Canada: debris slope accumulations. Arctic and Alpine Research, 11(4), 371-401.

Corominas, J., Copons, R., Vilaplana, J. M., Altimir, J., & Amigó, J. (2003). Integrated landslide susceptibility analysis and hazard assessment in the principality of Andorra. Natural Hazards, 30(3), 421-435.

Cossart, E., Mercier, D., Decaulne, A., & Feuillet, T. (2013). An overview of the consequences of paraglacial landsliding on deglaciated mountain slopes: typology, timing and contribution to cascading fluxes. Quaternaire. Revue de l'Association française pour l'étude du Quaternaire, 24(1), 13-24.

Cruden, D. M., & Hu, X. Q. (1993). Exhaustion and steady state models for predicting landslide hazards in the Canadian Rocky Mountains. Geomorphology, 8(4), 279-285.

D'Amato, J., Hantz, D., Guerin, A., Jaboyedoff, M., Baillet, L., & Mariscal, A. (2016). Influence of meteorological factors on rockfall occurrence in a middle mountain limestone cliff. Natural Hazards and Earth System Sciences, 16(3), 719-735.

De Blasio, F. V., & Sæter, M. B. (2010). Properties of talus slopes composed of flat blocks. Norsk Geografisk Tidsskrift–Norwegian Journal of Geography, 64(2), 85-93.

Decaulne, A., & Saemundsson, T. (2006). Geomorphic evidence for present-day snow- avalanche and debris-flow impact in the Icelandic Westfjords. Geomorphology, 80(1-2), 80-93.

Delonca, A., Gunzburger, Y., & Verdel, T. (2014). Statistical correlation between meteorological and rockfall databases. Natural Hazards and Earth System Sciences, 14(8), 1953-1964.

Dionne, J. C. (1976). Les grandes cuestas de la mer d'Hudson. GÉOS (Energie, Mines et Ressources Canada), 5(1), 18-20.

Domaas, U. (1994). Geometrical methods of calculating rockfall range. Norwegian Geotechnical Institute, Report, 585910(1).

Eaton, D. W. & Darbyshire, F. (2010). Lithospheric architecture and tectonic evolution of the Hudson Bay region. Tectonophysics, 480(1), 1-22.

Evans, S. G., & Hungr, O. (1993). The assessment of rockfall hazard at the base of talus slopes. Canadian geotechnical journal, 30(4), 620-636.

Fortier, R. (2017). Groundwater monitoring network from the Umiujaq region in Nunavik, Quebec, Canada, v. 1.3 (2012-2016). Nordicana D19, doi: 10.5885/45309SL- 15611D6EC6D34E23.

Francou, B., & Manté, C. (1990). Analysis of the segmentation in the profile of Alpine talus slopes. Permafrost and Periglacial Processes, 1(1), 53-60.

Fraser, C., Hill, P. R., & Allard, M. (2005). Morphology and facies architecture of a falling sea level strandplain, Umiujaq, Hudson Bay, Canada. Sedimentology, 52(1), 141-160.

Frayssines, M., & Hantz, D. (2006). Failure mechanisms and triggering factors in calcareous cliffs of the Subalpine Ranges (French Alps). Engineering Geology, 86(4), 256-270.

Gardner, J. S. (1983a). Observations on erosion by wet snow avalanches, Mount Rae area, Alberta, Canada. Arctic and Alpine Research, 15(2), 271-274.

Gardner, J.S. (1983b). Rockfall frequency and distribution in the Highwood Pass area, Canadian Rocky Mountains. Zeitshrift fur Geomorphologie, NF, 27, 311-324.

Germain, D. (2016). Snow avalanche hazard assessment and risk management in northern Quebec, eastern Canada. Natural Hazards, 80(2), 1303-1321.

Germain, D. & Martin, J. (2012). The vulnerability of northern cities to weather-related hazards: Case studies from the Province of Quebec, eastern Canada. Advances in Environmental Research. Volume 22. 143-160.

Germain, D., & Hétu, B. (2016). Hillslope processes and related sediment fluxes on a fine-grained scree slope of Eastern Canada. Source-to-Sink fluxes in undisturbed cold environments, 79-95.

Govi, M. (1977). Photo-interpretation and mapping of the landslides triggered by the Friuli earthquake (1976). Bulletin of the International Association of Engineering Geology-Bulletin de l'Association Internationale de Géologie de l'Ingénieur, 15(1), 67-72.

Grämiger, L. M., Moore, J. R., Gischig, V. S., Ivy‐Ochs, S., & Loew, S. (2017). Beyond debuttressing: Mechanics of paraglacial rock slope damage during repeat glacial cycles. Journal of Geophysical Research: Earth Surface, 122(4), 1004-1036.

Gruber, S., Hoelzle, M., & Haeberli, W. (2004). Permafrost thaw and destabilization of Alpine rock walls in the hot summer of 2003. Geophysical Research Letters, 31(13).

Guimont, P., & Laverdiere, C. (1980). Le sud-est de la mer d'Hudson: un relief de cuesta. The Coastline of Canada, SB McCann (edit.), Geological Survey of Canada, Paper, 80-10.

Hales, T. C., & Roering, J. J. (2007). Climatic controls on frost cracking and implications for the evolution of bedrock landscapes. Journal of Geophysical Research: Earth Surface, 112(F2).

Hétu, B. (1995). Le litage des Éboulis Stratifiés Cryonivaux en Gaspésie (Québec, Canada): Rǒle de la Sédimentation Nivéo‐Éolienne et des Transits Supranivaux. Permafrost and Periglacial Processes, 6(2), 147-171.

Hillaire-Marcel, C. (1976). La déglaciation et le relèvement isostatique sur la côte est de la baie d’Hudson. Cahiers de géographie du Québec, 20(50), 185-220.

Hungr, O. & Evans, S.G.(1988). Engineering evaluation of fragmental rockfall hazard. Proceedings. of 5th International Symposium on Landslides, Lausanne, 1988, 685-690.

Jenks, G. F. (1967). The data model concept in statistical mapping. International yearbook of cartography, 7, 186-190.

Jomelli, V. (1999). Dépôts d’avalanches dans les Alpes françaises: géométrie, sédimentologie et géodynamique depuis le Petit Age Glaciaire. Géographie physique et Quaternaire, 53(2), 199-209.

Jomelli, V., & Francou, B. (2000). Comparing the characteristics of rockfall talus and snow avalanche landforms in an Alpine environment using a new methodological approach: Massif des Ecrins, French Alps. Geomorphology, 35(3-4), 181-192.

Kellerer-Pirklbauer, A., Proske, H., & Strasser, V. (2010). Paraglacial slope adjustment since the end of the Last Glacial Maximum and its long-lasting effects on secondary mass wasting processes: Hauser Kaibling, Austria. Geomorphology, 120(1-2), 65-76.

Kirkby, M. J., & Statham, I. (1975). Surface stone movement and scree formation. The Journal of Geology, 83(3), 349-362.

Krautblatter, M., & Leith, K. (2015). Glacier-and permafrost-related slope instabilities. The High-Mountain Cryosphere; Huggel, C., Carey, M., Clague, JJ, Kaab, A., Eds, 147-165.

Krumbein, W. C. (1941). Measurement and geological significance of shape and roundness of sedimentary particles. Journal of Sedimentary Research, 11(2), 64-72.

Lajeunesse, P., & Allard, M. (2003a). Late quaternary deglaciation, glaciomarine sedimentation and glacioisostatic recovery in the Rivière Nastapoka area, eastern Hudson Bay, Northern Québec. Géographie physique et Quaternaire, 57(1), 65-83.

Lajeunesse, P., & Allard, M. (2003b). The Nastapoka drift belt, eastern Hudson Bay: implications of a stillstand of the Quebec–Labrador ice margin in the Tyrrell Sea at 8 ka BP. Canadian Journal of Earth Sciences, 40(1), 65-76.

Letortu, P. (2013). Le recul des falaises crayeuses haut-normandes et les inondations par la mer en Manche centrale et orientale: de la quantification de l'aléa à la caractérisation des risques induits. Doctoral dissertation, Université de Caen Basse- Normandie.

Lied, K. (1977). Rockfall problems in Norway. ISMES Publication, 90, 51-53.

Lied, K., & Domaas, U. (2000). Avalanche hazard assessment in Nunavik and on Cote- Nord, Quebec, Canada. Norwegian Geotechnical Institute.

Luckman, B. H. (1977). The geomorphic activity of snow avalanches. Geografiska Annaler: Series A, Physical Geography, 59(1-2), 31-48.

Luckman, B. H. (1978). Geomorphic Work of Snow Avalanches in the Canadian Rocky Mountains. Arctic and Alpine Research, 10(2), 261-276.

Luckman, B. (1988). Debris accumulation patterns on talus slopes in Surprise Valley, Alberta. Géographie physique et Quaternaire, 42(3), 247-278.

Marion, J., Filion, L., & Hétu, B. (1995). The Holocene development of a debris slope in subarctic Québec, Canada. The Holocene, 5(4), 409-419.

Mateos, R. M., García-Moreno, I., & Azañón, J. M. (2012). Freeze–thaw cycles and rainfall as triggering factors of mass movements in a warm Mediterranean region: the case of the Tramuntana Range (Majorca, Spain). Landslides, 9(3), 417-432.

Matsuoka, N. (2008). Frost weathering and rockwall erosion in the southeastern Swiss Alps: Long-term (1994–2006) observations. Geomorphology, 99(1-4), 353-368.

Matsuoka, N., & Sakai, H. (1999). Rockfall activity from an alpine cliff during thawing periods. Geomorphology, 28(3-4), 309-328.

McClung, D., & Schaerer, P. A. (2006). The avalanche handbook. The Mountaineers Books.

Michaud, Y., & Dionne, J. C. (1987). Altération des substrats rocheux et rôle du soulèvement gélival dans la formation des champs de blocaille, en Hudsonie. Géographie physique et Quaternaire, 41(1), 7-18.

Payette, S. (1983). The forest tundra and present tree-lines of the northern Québec- Labrador peninsula. Nordicana, 47, 3-23.

Pelletier, M. (2015). Geomorphological, Ecological and Thermal Time Phase of Permafrost Degradation, Tasiapik, Nunavik (Québec, Canada). Master thesis, Université Laval.

Pelletier, M., Allard, M., & Levesque, E. (2018). Ecosystem changes across a gradient of permafrost degradation in subarctic Québec (Tasiapik Valley, Nunavik, Canada). Arctic Science, (0), 1-26.

Pérez, F. L. (1989). Talus fabric and particle morphology on Lassen Peak, California. Geografiska Annaler: Series A, Physical Geography, 71(1-2), 43-57.

Rapp, A. (1960). Recent development of mountain slopes in Kärkevagge and surroundings, northern Scandinavia. Geografiska Annaler, 42(2-3), 65-200.

Rapp, A., & Strömquist, L. (1976). Slope erosion due to extreme rainfall in the Scandinavian mountains. Geografiska Annaler: Series A, Physical Geography, 58(3), 193-200.

Sauchyn, D. J. (1986). Particle size and shape variation on alpine debris fans, Canadian Rocky Mountains. Physical Geography, 7(3), 191-217.

Statham, I. (1976). A scree slope rockfall model. Earth Surface Processes, 1(1), 43-62.

Chapitre 2

Snow cornice and snow avalanche monitoring using automatic time-lapse cameras in Tasiapik Valley, Nunavik

Samuel Veilleux1,2, Najat Bhiry1,2 and Armelle Decaulne3

2.1 Résumé Des caméras à déclenchement automatique installées le long du versant sud-ouest de la vallée Tasiapik, près du village d’Umiujaq, ont permis de documenter de nombreux événements d’avalanche, leurs différents modes de départ et les types de dépôts impliqués au cours de l’hiver 2017-2018. Une dynamique corniche-avalanche a pu être observée, en plus des avalanches de plaque et de neige non-cohésive, et des dépôts de neige propre et sale. Au sommet de la cuesta, une caméra a suivi le développement d'une corniche de neige à partir de novembre 2017, ainsi que les nombreuses chutes de corniche au cours de l’hiver et au printemps. Deux autres caméras ont suivi l’évolution du versant situé sous la corniche, où une synchronicité entre les chute de corniche et le déclenchement des avalanches a été observée. L'activité avalancheuse est demeurée faible jusqu'à la fin mai, malgré quelques événements importants. Les avalanches printanières à la fin mai et en juin se caractérisent par une neige mouillée et sale transportant des débris au pied des pentes, ne s’arrêtant parfois qu’à quelques mètres de la route longeant le versant.

Mots-clés : avalanche, corniche à neige, chute, versant, Nunavik.

2.2 Abstract A series of automatic time-lapse cameras installed along the southwestern side of Tasiapik Valley, near the village of Umiujaq, documented several different departure modes and types of snow involved in snow avalanches during the winter of 2017-2018. These included cornice-avalanche dynamics, slab and loose snow avalanches, and clean and dirty snow-avalanche deposits. At the top of the selected cuesta, a camera monitored the development of a snow cornice beginning in November 2017, detecting multiple cornice failures over the winter and spring. The track and deposition area of the runout paths were monitored from two cameras downslope, revealing the concomitance of snow cornice fall and snow avalanche triggering. Snow avalanche activity remained infrequent until the end of May, with only a few large events. Spring snow avalanches in late May and June are characterized by wet and dirty snow avalanches carrying debris at the foot of the slope and by runout zones reaching the road along the slope.

Key words: snow avalanche, snow cornice, failure, slopes, Nunavik.

I. Introduction Snow avalanches have been a sensitive topic for the Inuit of Nunavik ever since the avalanche that occurred on January 1, 1999, when 9 people lost their lives during the New Year celebrations at the Kangiqsualujjuaq Satuumavik School. The coroner’s report recommended that the potential for snow avalanches should be assessed for all Nunavik villages (Bérubé, 2000; Hétu et al., 2008). A report carried out on behalf of Québec’s Ministère de la Sécurité Publique stated that avalanche events have also occurred in the villages of Inukjuak, Ivujivik, Kangirsuk, Kangiqsujuaq, Quaqtaq and Salluit, where the topography and meteorological conditions are suitable (Lied and Domaas, 2000). Most of the villages listed above and their close surroundings are located near notable slopes and some infrastructures are potentially at risk. There is a pressing need to document these natural hazards, especially in a land-planning context, given the demographic growth in Nunavik (Duhaime, 2007; Carbonneau et al., 2015; Germain, 2016; L’Hérault et al., 2017).

In Tasiujaq Lake and the nearby Tasiapik Valley area close to the village of Umiujaq, the cuesta relief (asymmetrical relief with a steep frontslope) exposes steep rocky escarpments that are prone to slope movements (Veilleux, 2019a). Notable among these movements are snow avalanches, which have likely occurred for some time in the area, based on the testimony of local residents of Umiujaq. In addition, previous talus slope investigations in Tasiapik Valley have provided evidence of debris remobilization (Veilleux et al., 2017, 2019a), likely caused by snow-related processes.

Although few infrastructures are found in Tasiapik Valley (e.g., there is a single road connecting Umiujaq to Tasiujaq Lake), it is an important area for the Inuit who practice traditional activities such as fishing, hunting and berry picking. Many people also visit the area, particularly since the creation of Tursujuq National Park in 2013, and thus could be exposed to snow avalanches. The objective of this study is to examine the relationship between snow avalanche events and their triggering causes in Tasiapik Valley by considering the impact of meteorological conditions and topography parameters on the evolution of snow-cover instability.

II. Study area Tasiapik Valley (56°33' N, 76°28' W) is located 5 km east of the Inuit village of Umiujaq, on the east coast of Hudson Bay in Nunavik, Québec (Figure 15a). It is approximately 4.5 km long and 1.5 km wide, following a northwest-southeast orientation. At the southeastern end of the valley lies Tasiujaq Lake (formerly named Guillaume-Delisle Lake and Richmond Gulf), a brackish 691 km2 lake connected with the Hudson Bay by a narrow cataclinal channel called Le Goulet (ARK, 2007).

The regional geology is characterized by an asymmetrical monoclinal relief called a cuesta, revealing a gentle west-dipping slope and a steep eastern frontslope (Dionne, 1976; Guimont and Laverdière, 1980). The Hudson Bay cuestas consist of a Paleoproterozoic volcano-sedimentary sequence lying unconformably on the Precambian shield (Stockwell et al., 1979; Chandler and Schwarz, 1980; Chandler, 1988; Eaton and Darbyshire, 2010). The valley slopes are located at the front of the cuesta; slope height increases southward from 50 m (upstream) to 230 m (downstream) near Tasiujaq Lake (Figure 15b).

Tasiapik Valley is located in the discontinuous permafrost zone and has a subarctic climate. Mean annual air temperatures recorded between 2013 and 2017 vary between -5.6°C and -4.2°C, with maxima at 23°C and minima at -36°C (Fortier, 2017). Mean annual precipitation is approximately 500 mm, with 40% falling as snow (Ménard et al., 1998). The Umiujaq area is located at the edge of shrub and forest tundra zones; low shrubs, ericaceous plants and lichen cover the upstream part of the valley, while there is dense forest cover in the downstream part (Payette, 1983; Allard and Séguin, 1987; Provencher-Nolet et al., 2014; Pelletier et al., 2018). Shrub cover has expanded significantly (shrubification) during the 20th century (Ménard et al., 1998, Provencher-Nolet et al., 2014, Pelletier et al., 2018). Small peatlands and thermokarst ponds are located near the shore of Tasiujaq Lake, in the southernmost part of the valley.

Figure 15: Location of Tasiapik Valley in the Umiujaq region (A) (Source of background image: MRNF.); oblique view toward the north, showing the cuesta frontslope on the SW side (B).

III. Methods

1. Time-lapse cameras To detect slope movements, and more specifically snow avalanches, over the course of 12 months, three Reconyx PC800 Hyperfire time-lapse cameras were installed in the summer of 2017 near talus slopes on the southwestern side of Tasiapik Valley (Figure 16). The use of time-lapse photography to monitor snow-related processes and evolution has been adopted in other mountainous/cold regions such as Spitsbergen (Vogel et al., 2010, 2012; Eckerstorfer et al., 2013), the Swiss Alps (van Herwijnen et al., 2013; van Herwijnen and Fierz, 2014), Norway (Laute and Beylich, 2014) and the western United States (Hendrikx et al., 2012; Munroe, 2018). Two cameras are located at the foot of the talus slopes, and one is located at the top of the cuesta. Photos were taken every hour from 9:00 am to 5:00 pm for the period between August 2017 and June 2018. Between July and August 2018, the photographs covered a longer daytime period, from 6:00 am to 8:00 pm, and were taken every 15 or 30 minutes (depending on the camera location). In total, more than 14 000 photographs were taken from August 7, 2017 to August 15, 2018.

2. Meteorological analysis The meteorological conditions observed in the photographs were compared with data from the SILA Centre d’études nordiques weather stations located in the valley (VDTSILA) and at Umiujaq airport (UMIROCA). The meteorological data reveal meteorological variations and the evolution of the snow cover, which provides evidence of snow avalanche triggering. While temperatures are registered by the cameras and noted on each photograph, the actual air temperature differs due to the impact of solar radiation on the camera sensor (which is maintained in a metal case). Therefore, the meteorological data (including the temperature) from the SILA weather stations were preferred over the photographic temperature data.

3. Fieldwork The knowledge of the terrain acquired from fieldwork was used to determine snow depth on the plateau and at the foot of the slopes; photographs taken in situ during the summer also helped to estimate snow depth, thanks to well-known scales (human-scale, length and thickness of large boulders, height of vegetation, etc.). Snow avalanche deposits were examined during 6 days of fieldwork, from June 11 to June 16, 2018. A total of 29 deposits were identified over the fieldwork period. On-site characterization included avalanche descriptions and photographs, while GPS

coordinates were taken for every deposit. Snow avalanche deposit-mapping was used to track their dispersion throughout the valley. Finally, numerous snow avalanches were witnessed in real-time during fieldwork, facilitating the identification of their triggering and flowing processes.

Figure 16: Location of the cameras along the SW side of Tasiapik Valley. Field of view of each camera is shown on the left. Source of background image: UMI Orthomosaic (2010).

IV. Results and interpretation

1. Snow cornice formation and collapses In 2017-2018, a snow cornice developed on the cuesta ridgeline (SW side) of Tasiapik Valley. It started to form in early November 2017 and completely disappeared in early July 2018, lasting for a total of 248 days. Snow precipitation began in early October 2017, although it only remained on the ground from early November 2017 when daily temperatures were below 0°C (Figure 17a). Observations reveal a rapid growth of the cornice in November, as blizzard-like conditions were frequent (19 days out of 22) from November 1 to November 20 (Figure 17b). Snow accumulation totalled 118 cm during this period, while wind speed averaged 6.2 m.s-1, with max speed reaching up to 14.0 m.s-1 (Figure 17c). Data from the UMIROCA station indicates a predominantly easterly wind for this period, while the VDTSILA station indicates a prevailing wind from the southwest (Figure 17d). However, both stations indicate SW-WSW prevailing winds on the snowy/blizzard days. These values approach the theoretical optimal wind direction for cornice accretion on a NE-oriented slope (~212°), considering that the gently-inclined cuesta plateau is SW-oriented with virtually no wind obstacles, thus enabling efficient snow transport.

Figure 17: Meteorological conditions during the cornice accretion period: temperature (A), snowfall (B), wind speed (C) and wind direction (D). Source: CEN (2018), Environment Canada.

Evidence of snow displacement by wind action on the upwind cuesta plateau was observed in the time-lapse photographs. Drifting snow conditions occurred from November to May and led to sastrugi formation (ridge-like features oriented in the wind direction). Oriented with the SW prevailing winds, the irregular dune-like formations form parallel ripples facing NE, as observed during the winter months. The depth of the snowpack is variable on the cuesta plateau, as depressions collected up to 1.5 m of snow, while the windswept higher ground remained bare rock all year long. On the cornice, the snowpack reached its maximum depth (~1.5 m) in early January 2018. From then on, snow precipitation was less abundant, as the depth of snow cover at the VDTSILA station was consistent between January (~65 cm) and February/March (~84 cm). As snowfall data are unavailable for this period in Umiujaq, data from the Kuujjuarapik climatic station suggest a similar trend, recording only 50.7 cm between January and March. During this period, SSE to SSW prevailing winds contributed to the horizontal expanse of the cornice. This continuous snow loading resulted in the progressive deformation of the cornice, predominantly the tilting and creeping of the leading edge and roll face, as shown on Figure 18. This downward motion was observed in the photographs on both the TAS1 and TAS2 cameras. No cornice failure was witnessed until March. However, we should mention that low sunlight, frequent blizzard conditions and frost on the camera considerably restricted visibility. Nevertheless, the snow cover in the upper part of the slope under the cornice appeared to have remained intact before March, suggesting that collapses were infrequent. Thereafter, collapses occurred more frequently, especially from late May through June.

Figure 18: Schematic cross-section and photographs representing the accretion, the creeping and collapsing/melting periods.

2. Snow-avalanche events Most of the snow avalanches (36 of a total of 39 events) were observed on the TAS2 camera. Only three events were observed on the TAS3 camera: on April 26, April 28 and May 28, 2018 respectively (Table 4).

On the TAS2 camera, five events were identified before March 2018. These occurred on November 15, 2017, December 3, 2017, December 25, 2017, January 30, 2018 and February 19, 2018 (Figure 19). Low visibility makes it difficult to clearly distinguish the types and shapes of snow avalanches that occurred during this period; however, snow balls rolling on the snow-covered talus slope could be observed in November and December 2017. A slab avalanche occurred on March 5, 2018, leaving a clear crown and bed surface in the uppermost part of the slope. The event was likely triggered by a cornice collapsing on the same day, as observed on the TAS1 camera. Loose-snow avalanches occurred in late March 2018 and early April 2018, but they seem to have been set off at mid-slope. On April 17, 2018, a snow avalanche occurred after the failure of a large portion of the cornice; a distinct scarp left by the collapsing cornice is visible, as well as the space left vacant by the tilted roll face of the cornice. Another snow avalanche occurred on April 25, 2018, as a track seemed to have set off from the top of the slope. However, there is no visible evidence of a fallen cornice. A slab avalanche (April 27, 2018) and a mid-slope loose-snow avalanche (April 30, 2018) are also clearly visible in the time-lapse images. From May 27, 2018 to June 16, 2018, a total of 25 snow avalanche events occurred. Every cornice collapse during this period occurred on days when snow avalanche deposits were observed, showing a moderate positive relationship (r=0.58) between the two.

According to the UMIROCA and VDTSILA climate stations, winter conditions prevailed prior to May 27, 2018. Maximal and minimal daily air temperatures remained below 0°C and the majority of precipitation fell as snow (Figure 20). These conditions contributed to the thickening of the snow cover on the slopes, thus creating favourable conditions for snow avalanches. Beginning on May 27, 2018, warmer temperatures and abundant rainfall episodes combined to cause multiple cornice failures and snow avalanches in the following days. A total of 8 collapses and 23 snow avalanche events were observed until June 16. Thereafter, no snow avalanches were observed, although sporadic falls of rock debris occurred as well as debris sliding on the snow-covered talus. Rapid melting of the snow led to the retreat of the cornice and no failures occurred during this period.

Ephemeral meltwater streams were observed in late June and July on the slope, flowing down from the cornice to the talus.

With regard to the valley, snow avalanche activity was significant during fieldwork in June 2018. At least 20 events were witnessed, the majority of which were triggered by cornice failures. Heavy rainfall (13.1 mm) on June 11, followed by five days of good weather (no rainfall and temperatures above zero), likely increased the weight of the cornice and thus its tilting motion. These meteorological conditions were similar to those prevailing on May 27 and the following days.

Table 4: Calendar of snow-avalanche activity.

Snow- Time of Aspect of Aspect of the Liquid Starting Camera Date cornice deposit the snow in Sliding surface deposit roughness water mode failures observation the deposit November 15, 2017 No 14:00 Rounded snow balls Clean Absent Unknown On snow cover December 3, 2017 No 09:00 Rounded snow balls Clean Absent Unknown On snow cover December 25, 2017 No 09:00 Rounded snow balls Clean Absent Unknown On snow cover January 30, 2018 No 09:00 Rounded snow balls Clean Absent Unknown On snow cover February 19, 2018 No 09:00 Rounded snow balls Clean Absent Unknown On snow cover March 5, 2018 Yes 09:00 Rounded snow balls Clean Absent Slab On snow cover March 28, 2018 Yes 09:00 Unknown Clean Absent Loose On snow cover April 12, 2018 No 10:00 Unknown Clean Absent Loose On snow cover April 12, 2018 No 12:00 Unknown Clean Absent Loose On snow cover April 17, 2018 Yes 12:00 Angular snow blocks Clean Absent Slab On ground April 24, 2018 No 09:00 Rounded snow balls Clean Absent Loose On snow/ground April 27, 2018 No 09:00 Rounded snow balls Clean Absent Slab On snow cover April 30, 2018 No 09:00 Unknown Clean Absent Loose On snow cover May 27, 2018 No 11:00 Rounded snow balls Dirty Absent Loose On snow cover May 28, 2018 No 09:00 Rounded snow balls Dirty Absent Loose On snow/ground May 28, 2018 No 09:00 Rounded snow balls Dirty Absent Loose On snow/ground May 28, 2018 No 10:00 Rounded snow balls Dirty Absent Loose On snow/ground

May 28, 2018 Yes 15:00 Rounded snow balls Dirty Absent Loose On snow/ground May 30, 2018 Yes 15:00 Angular snow blocks Clean Absent Loose On snow/ground

TAS2 May 30, 2018 No 16:00 Rounded snow balls Dirty Absent Loose On snow/ground May 30, 2018 No 17:00 Angular snow blocks Dirty Absent Slab On snow/ground May 31, 2018 Yes 09:00 Unknown Clean Absent Loose On snow cover May 31, 2018 Yes 17:00 Rounded snow balls Clean Absent Loose On snow/ground June 6, 2018 No 09:00 Rounded snow balls Dirty Absent Loose On snow/ground June 6, 2018 No 12:00 Rounded snow balls Clean Absent Loose On snow/ground June 6, 2018 Yes 16:00 Angular snow blocks Clean Absent Loose On snow/ground June 6, 2018 No 17:00 Unknown Clean Absent Loose On snow/ground June 7, 2018 No 10:00 Rounded snow balls Dirty Absent Loose On snow/ground June 7, 2018 No 10:00 Rounded snow balls Dirty Absent Loose On snow/ground June 8, 2018 No 10:00 Unknown Clean Absent Loose On snow/ground June 10, 2018 No 09:00 Rounded snow balls Clean Absent Loose On snow/ground June 11, 2018 Yes 13:00 Unknown Dirty Absent Loose On snow/ground June 12, 2018 No 18:15 Unknown Dirty Unknown Loose On ground June 12, 2018 No 20:00 Unknown Dirty Present Loose On ground June 14, 2018 No 06:00 Unknown Dirty Unknown Loose On snow/ground June 16, 2018 No 19:15 Unknown Dirty Absent Unknown On ground

April 26, 2018 No 09:00 Rounded snow balls Clean Absent Unknown On snow cover April 28, 2018 No 09:00 Unknown Clean Absent Unknown On snow cover

TAS3 May 28, 2018 No 09:00 Rounded snow balls Clean Absent Unknown On snow cover

Figure 19: Snow avalanche events identified by the TAS2 camera in 2017-2018.

Figure 20: Occurrence of cornice failures and snow avalanches during the winter/spring of 2017-2018. Temperature and precipitations are also shown. Source: CEN (2018), Environment Canada.

3. Snow-avalanche deposits During the fieldwork on June 14, 2018, a total of 29 deposition zones were identified, extending along the entire length of the SW side of the valley (Figure 21). There were no deposition zones on the opposite NE side, as the snow cover had melted considerably by that time. Runout zones were clearly outlined for all snow avalanche deposits; the starting zone and track were sometimes hard to identify as snow had completely melted on the steepest parts of the slope. The snow cornice was still visible in June 2018 along most of the cuesta ridgeline, although it had considerably receded in the previous weeks.

The deposits found in the runout zones of the June 2018 snow avalanches consist of densely packed, wet snow with incorporated rocky debris, giving a dirty appearance to the deposits. The deposits generally exhibit a lobate shape, but they can also have a multilobate shape. In addition, snowballs resulting from the collapse of the cornice rolled down the slope. By comparison, deposits observed before May 27 consisted of clean snow, originating either from loose-snow or slab avalanches.

Deposits located in the downstream part of the valley show an elongated form, extending from the snow cornice down to a rocky outcrop located mid-slope (Figure 21a). Runout distances range between 275 and 350 m, and the mean runout angle (alpha angle) is 37.2°. The width of the deposits range from 28 to 80 m (푥̅ = 47 m), with a mean width/length (W/L) ratio of 0.15.

The steeper slope mid-valley led to less snow accumulation, and snow avalanche tracks are not always visible. However, the snow cover is still sufficient in the uppermost part of the slope to allow us to identify the starting zone (Figure 21b). Various forms were observed, including elongated, fan- shaped, juxtaposing and coalescing deposits. Runout distances range from 120 to 260 m, with a mean runout angle of 44.2°.The width ranges from 12 to 140 m (푥̅ = 58 m) and mean W/L ratio is 0.26.

In the upstream part of the valley, a lower height and gentler slope helped to preserve the snow cover, but also led to shorter runout distances, ranging from 49 to 112 m, with a mean runout angle of 31.2° (Figure 21c). The width ranges from 16 to 140 m (푥̅ = 40 m), with a mean W/L ratio of 0.52. Deposits exhibit a short lobate shape, with large (up to 5 m long) snow blocks located in the runout

zone, as shown on Figure 21c. In addition, many snowballs released from the uppermost part of the slope rolled down on the snow-covered slope, leaving well-defined tracks.

The road connecting Umiujaq to Tasiujaq Lake runs along the entire length of the SW side of Tasiapik Valley and is therefore threatened by snow avalanches due to its proximity to the slope. The many fallen boulders found near the road embankment on both sides of the road demonstrate the potential risk associated with mass movements on the slope (Veilleux, 2019a). This shows that snow avalanches have already had a significant impact in the area. The average distance between snow avalanche deposits and the road is 112.2 m, with generally greater distances in the downstream part (235 m) and shorter distances in the upstream part of the valley (85 m). However, the snow avalanche event of April 17, 2018 did reach the road (as observed on the TAS2 camera), as did another mid-valley event whose runout zone is located only a few meters from the road (Figure 21b).

Discussion

1. Meteorological and topographic control The results of the study have demonstrated that meteorological conditions were favourable for the accretion of a snow cornice at the top of the cuesta frontslope in November 2017. Moreover, the cornice lasted for 248 days, meaning that conditions were also conducive for continuous snow supply throughout the winter and spring. Considering the fact that snowfall is limited in the winter due to the freezing of Hudson Bay in late December 2017, thus causing the shift toward a more continental climate (Wilson, 1971), the effect of snow transport by wind is considerable.

Li and Pomeroy (1997) established that efficient dry-snow transport, typical of arctic and subarctic regions, occurs when wind speed is within a threshold between 4 and 11 m/s, whereas McClung and Schaerer (2006) established that a 5 to 10 m/s threshold wind speed could initiate saltation for cold loose snow and induces the formation and growth of snow cornices. From November 1, 2017 to June 1, 2018, 67.9% of the days had an average wind speed and 81.6% had a maximum wind speed within the 4-11 m/s threshold. Those percentages may be underestimated, since the conditions at the top of the cuesta do not fully reflect the data used from the UMIROCA and VDTSILA climatic stations. In addition, 41.8% (VDTSILA) or 48.3% (UMIROCA) of those days had efficient wind direction, considering a WNW to SSE direction span for efficient snow supply to the cornice. The presence of sastrugi on the snow surface, generally oriented in a NE-SW direction, indicates continuous snow transport and erosion by winds until the end of May 2018. In addition, the gently westward-dipping cuesta plateau is nearly unobstructed, which enables efficient wind travel from Hudson Bay toward the cuesta ridge.

Snow transport by wind enables the sorting of snow particles on the cornice. Larger, denser particles are deposited on the root of the cornice, whereas smaller and less cohesive particles are deposited on the roof (Naruse et al., 1985; Benson and Sturm, 1993; Sokratov and Sato, 2001) and on the lee- side slope (McClung and Schaerer, 2006). The roof of the cornice is sensitive to rapid changes in meteorological conditions, especially during the spring. McCarty et al. (1986) suggested that the roof of the cornice can become isothermal (no temperature gradient within the snow cover), thus reacting rapidly to weather variations. This situation occurred in late May 2018, as the air temperature had risen above 0°C and precipitation shifted from snow to rain, resulting in multiple cornice failures. For instance, on May 27, 2018, the air temperature reached 5.4°C, while rainfall totalled 21 mm, leading

to five cornice failures in the subsequent four days. This rain event, combined with warmer temperatures, contributed to the overloading of the cornice, as it tilted significantly during those days. The effect of longer daylight in spring could also exacerbate the meteorological control of warm temperature and rain on the cornice. A similar temporal pattern was observed by Vogel et al. (2012) and Eckerstorfer et al. (2013) in Spitsbergen, with a strong propensity for cornice failures in the spring.

During the winter, blowing snow over the cornice is deposited on the lee-side slope below (Figure 22). The accumulation of those small non-cohesive particles on the slope thus favours snow avalanche triggering by building-up a wind slab (McClung and Schaerer, 2006). This NE-oriented slope also benefits from shade, especially during the winter when there is less daylight (Figure 23c). These factors induce the development and preservation of weak snow layers (McClung and Schaerer, 2006). The crown of these slab avalanches is often located in the uppermost part of the slope, under the cornice, as shown on Figure 22.

Figure 22: Schematic cross-section showing the snow redistribution by the wind, and the formation of a lee zone deposition on the slope below the snow cornice. Two slab avalanches are shown on the bottom right, with red lines highlighting the crowns.

On the NE side of Tasiapik Valley, by contrast, there is much less snow avalanche activity due to its SW orientation. In June 2018, the snowpack had mostly melted as the slopes were in direct sunlight during the day (Figure 23b), whereas the shaded SW side still had a deep snow cover with abundant snow avalanche deposits (Figure 23a). In addition, no snow cornice was visible at the top of the NE side at that time. The wind pattern discussed earlier may be responsible for the potential absence of snow cornice on this side of the valley. However, snow avalanches may have occurred earlier in the winter, without being observed.

Results also suggest a strong synchronicity of cornice failures and snow avalanches, as the first often triggered the second, especially in the spring. Moreover, during fieldwork in June 2018, over 20 snow avalanche events were observed on the SW side of the valley, and most of them were triggered by collapsing cornices. The latter, characterized by a loud sound, were particularly frequent after the warm and rainy conditions of June 11, 2018.

Figure 23: Snow cover on the SW side (A) and NE side (B) on June 12, 2018. Shaded SW side in June 2018 (C).

2. Runout and proximity to the road Prior to melting in June 2018, the road passing through Tasiapik Valley is not plowed, although it is used in winter by local people travelling with snowmobiles. During the rest of the year, there is a lot of travel on the road, as it is mostly used by the local population, researchers and tourists (Tursujuq National Park).

Wet snow avalanches have limited runout, especially mid-valley, due to their slow flow and the abrupt transition between the slope and the valley floor. The slope’s change in angle from ~74° to ~4° causes the rapid deceleration of the wet snow avalanche and compaction of the snow in the runout zone (McClung and Schaerer, 2006). However, at this location, the road is closest to the rockwall (60-80 m), suggesting that some events could potentially reach the road. In June 2018, snow avalanches were witnessed there, and densely packed runout zones have nearly reached the road. One stopped only ~5 m before the road (Figure 24). Further downstream, slab avalanches were observed on the TAS2 camera in March, April and May 2018. These have longer runouts than wet snow avalanches and nearly reached the road, which is located more than 200 m from the rockwall (i.e. on April 17, 2018).

Figure 24: Wet snow-avalanche deposits with runout distances near the road. Photo taken on June 12, 2018.

Other fan-shaped and coalescent snow avalanche deposits have more limited runouts, as they do not reach the base of talus (Figure 25). In this case, deposits are located far from the road and the terrain separating them is virtually flat; the runout is also notably shorter than other deposits from the snow avalanches that occurred earlier in the spring. The multilobe shape is attributable to the high water and debris content of the deposit, similar to a mudslide geomorphology according to Jomelli and Bertran (2001). Due to their flow in direct contact with the ground, their erosive capacity on the rockwall is stronger (Gardner, 1983; McClung and Schaerer, 2006) and most of the wet snow avalanche deposits contain a dirty and dense snow, mainly fine particles but also boulders, which could potentially endanger people travelling on the road or nearby. Moreover, the presence of scree slopes at the foot of the rockwall indicates the occurrence of rockfalls in the area (Veilleux et al., 2019b, submitted). Rockfall events occurred in June 2018 during the peak period of snow avalanche activity, showing the synchronicity of both slope processes. It is also likely that snow avalanches played an important role in supplying debris for the numerous talus slopes in Tasiapik Valley.

Figure 25: Dirty snow avalanche deposits in June 2018. Comparison of a shorter runout distance from this event (A) with another wet snow avalanche with a longer runout occurring in late May 2018 (B) and a slab avalanche occurring in April 2018 (C).

3. Limitations This study has some limitations that affect the interpretation of the data. First, the use of automatic cameras in such a harsh environment inevitably involves a large number of unusable photographs due to poor visibility (i.e. frost, blowing snow, low sunlight/shade). Between November 1, 2017 and May 31, 2018, visibility was poor for 95 days (44.8% of total days) on the TAS1 camera, 36 days (17.0%) on the TAS2 camera, and 65 days (30.7%) on the TAS3 camera (Figure 26). Conditions of poor visibility often occurred on multiple consecutive days, which means that snow avalanche events could have been missed due to drifting snow rapidly covering the deposit, wind erosion of the snow surface and sublimation. In addition, since the TAS2 and TAS3 cameras only have a wide frame view of ~200 m, they represent only a small sample of the valley’s 5 km long SW side. Conditions observed prior to fieldwork in June 2018 could also be locally favourable (TAS2) or unfavourable (TAS3) to cornice failures and snow avalanches, although it is impossible at the moment to extrapolate these events (or non-events) at the valley scale. In addition, data collected and analyzed at this time only covers a one-year period, and thus it is impossible to obtain an interannual pattern at this stage. Finally, the last consideration concerns the meteorological data from the UMIROCA and VDTSILA stations. As the first is located at Umiujaq airport and the second is located on the valley floor, both stations do not reflect the actual conditions occurring near the ridgeline. Due to higher altitude and the ridge-like position, temperatures could be colder than nearby Umiujaq or the valley floor, and wind could be stronger (as experienced in the field). Therefore, the data used for this study are not entirely representative of the actual conditions at the top of the cuesta.

Figure 26: Days with poor visibility on the three cameras installed in Tasiapik Valley.

Conclusion This study has highlighted the dynamic between a snow cornice and snow avalanche events during the winter and spring of 2017-2018 in Tasiapik Valley. The study found evidence of meteorological and topographic conditions that were conducive to snow cornice accretion and failures, as well as snow avalanche triggering on the SW side of the valley. A strong propensity for spring (May-June) wet snow avalanches was observed, resulting from a rapid shift in weather conditions towards warmer temperatures and rainfall instead of snowfall. These observations were confirmed by the in- situ investigation of snow avalanche deposits in June 2018. Abundant debris-filled snow avalanche deposits have shown their great erosive capacity and thus their significance for debris supply on talus slopes. In addition, rockfall events have been observed during the peak period of snow avalanche activity, revealing a synchronicity between the two slope processes.

The runout zones of a slab avalanche occurring in April 2018 and wet snow avalanches occurring in June 2018 are located only a few meters from the road. Although risk assessment is not the objective of this study, it is likely that snow avalanches could hit the road and endanger its users, especially mid-valley where the road is located only 60 m from the slope.

The use of automatic cameras has helped us to better understand the slope dynamic over a one-year span, although frequent blizzard and fog events have often limited the visibility. However, further camera monitoring for the years to come is required in order to assess the interannual recurrence of snow cornice formation/failure and snow avalanches. From informal discussions with local people in Umiujaq, we were told that the 2017-2018 winter was particularly harsh, which could have induced the formation of a cornice and triggered more snow avalanches. Nevertheless, meteorological records from the previous years (2012 to 2017) do not show a different trend from the 2017-2018 observation year. Hence, this would demonstrate the potential for cornice failures and snow avalanches every year, although photo-monitoring over the next few years will certainly provide further clarification.

References Allard, M., & Seguin, M. K. (1987). The Holocene evolution of permafrost near the tree line, on the eastern coast of Hudson Bay (northern Quebec). Canadian Journal of Earth Sciences, 24(11), 2206-2222.

Benson, C. S., & Sturm, M. (1993). Structure and wind transport of seasonal snow on the Arctic slope of Alaska. Annals of Glaciology, 18, 261-267.

Bérubé, J. (2000). Rapport d’enquête publique sur les circonstances des décès survenus à Kangiqsualujjuaq, Nouveau-Québec le premier Janvier 1999. Gouvernement du Québec, Bureau du coroner.

Carbonneau, A-S., L’Hérault, E., Aubé-Michaud, S., Taillefer, M., Ducharme, M-A., Pelletier, M. & Allard, M., (2015). Production de cartes des caractéristiques du pergélisol afin de guider le développement de l'environnement bâti pour huit communautés du Nunavik. Rapport final. Québec, Centre d'études nordiques, Université Laval. 127 p.

CEN (2018). Climate station data from the Umiujaq region in Nunavik, Quebec, Canada, v. 1.6 (1997-2018). Nordicana D9, doi: 10.5885/45120SL-067305A53E914AF0.

Chandler, F. W. (1988). The early Proterozoic Richmond Gulf Graben, East Coast of Hudson Bay, Quebec (Vol. 362). Geological Survey of Canada.

Chandler, F. W., & Schwarz, E. J. (1980). Tectonics of the Richmond Gulf area, northern Quebec—a hypothesis. Current Research, Part C, Geological Survey of Canada, Paper, 80, 59-68.

Dionne, J. C. (1976). Les grandes cuestas de la mer d'Hudson. GÉOS (Energie, Mines et Ressources Canada), 5(1), 18-20.

Duhaime, G. (2007). Profil socioéconomique du Nunavik. Chaire Condition Autochtone.

Eaton, D. W., & Darbyshire, F. (2010). Lithospheric architecture and tectonic evolution of the Hudson Bay region. Tectonophysics, 480(1-4), 1-22.

Eckerstorfer, M., Christiansen, H. H., Rubensdotter, L., & Vogel, S. (2013). The geomorphological effect of cornice fall avalanches in the Longyeardalen valley, Svalbard. The Cryosphere, 7(5), 1361.

Fortier, R. (2017). Groundwater monitoring network from the Umiujaq region in Nunavik, Quebec, Canada, v. 1.3 (2012-2016). Nordicana D19, doi: 10.5885/45309SL- 15611D6EC6D34E23.

Gardner, J. S. (1983). Observations on erosion by wet snow avalanches, Mount Rae area, Alberta, Canada. Arctic and Alpine Research, 15(2), 271-274.

Germain, D. (2016). Snow avalanche hazard assessment and risk management in northern Quebec, eastern Canada. Natural Hazards, 80(2), 1303-1321.

Guimont, P., & Laverdiere, C. (1980). Le sud-est de la mer d'Hudson: un relief de cuesta. The Coastline of Canada, SB McCann (edit.), Geological Survey of Canada, Paper, 80-10.

Hendrikx, J., Peitzsch, E., & Fagre, D. B. (2012). Time-lapse photography as an approach to understanding glide avalanche activity. In Proceedings of the 2012 International Snow Science Workshop (pp. 872-877).

Jomelli, V., & Bertran, P. (2001). Wet snow avalanche deposits in the French Alps: structure and sedimentology. Geografiska Annaler: series A, physical geography, 83(1‐ 2), 15-28.

L’Hérault, E., Boisson, A., Allard, M., Aubé-Michaud, S., Sarrazin, D., Roger, J., & C. Barrette (2017). Détermination et analyse des vulnérabilités du Nunavik en fonction des composantes environnementales et des processus physiques naturels liés au climat. Rapport final. Réalisé pour le compte du Ministère des Forêts, de la Faune et des Parcs, Gouvernement du Québec. Centre d’études nordiques, Université Laval, 160 p.

Li, L., & Pomeroy, J. W. (1997). Estimates of threshold wind speeds for snow transport using meteorological data. Journal of Applied Meteorology, 36(3), 205-213.

Lied, K., & Domaas, U. (2000). Avalanche hazard assessment in Nunavik and on Côte- Nord, Québec, Canada. Norwegian Geotechnical Institute.

McCarty, D., Brown, R. L., & Montagne, J. (1986). Cornices: their growth, properties, and control. In: International Snow Science Workshop, Lake Tahoe, 1986. p. 41-45.

McClung, D., & Schaerer, P. A. (2006). The avalanche handbook. The Mountaineers Books.

Munroe, J. S. (2018). Monitoring snowbank processes and cornice fall avalanches with time-lapse photography. Cold Regions Science and Technology, 154, 32-41.

Naruse, R., Nishimura, H., & Maeno, N. (1985). Structural Characteristics of Snow Drifts and Cornices. Annals of Glaciology, 6, 287-288.

Payette, S. (1983). The forest tundra and present tree-lines of the northern Québec- Labrador peninsula. Nordicana, 47, 3-23.

Pelletier, M., Allard, M., & Levesque, E. (2018). Ecosystem changes across a gradient of permafrost degradation in subarctic Québec (Tasiapik Valley, Nunavik, Canada). Arctic Science, (0), 1-26.

Sokratov, S. A., & Sato, A. (2001). The effect of wind on the snow cover. Annals of Glaciology, 32, 116-120.

Stockwell, C. H., McGlynn, J. C., Emslie, R. F., Sanford, B. V., Norris, A. W., Donaldson, J. A., Fahrig, W. F. & Currie, K. L. (1979). Géologie du Bouclier canadien. In Douglas, R. J. W. & Tremblay, L. P., Géologie et ressources minérales du Canada (p. 117-119). Energie, mines et ressources Canada. van Herwijnen, A., Berthod, N., Simenhois, R., & Mitterer, C. (2013). Using time-lapse photography in avalanche research. In Proceedings of the International Snow Science Workshop (pp. 950-954). van Herwijnen, A., & Fierz, C. (2014). Monitoring snow cornice development using timelapse photography. In Proceedings of the International Snow Science Workshop (pp. 865-869).

Vogel, S., Eckerstorfer, M., & Christiansen, H. H. (2010). Cornice dynamics above Nybyen in Svalbards high arctic landscape. In Proceedings of the International Snow Science Workshop (pp. 785-790).

Vogel, S., Eckerstorfer, M., & Christiansen, H. H. (2012). Cornice dynamics and meteorological control at Gruvefjellet, Central Svalbard. The Cryosphere, 6(1), 157-171.

Wilson, C. (1971). Atlas climatologique du Québec. Service de l’Environnement Atmosphérique du Canada, Toronto, Ont.

Conclusion générale

Cette étude a permis de documenter les processus gravitaires dominants qui se sont produits dans le passé, depuis la déglaciation ainsi que ceux qui ont lieu tout récemment (2017-2018) au front des cuestas hudsoniennes dans la vallée Tasiapik, près d’Umiujaq (Nunavik).

Les relevés topographiques, géomorphologiques et écologiques (végétation de surface) réalisés sur les talus d’éboulis ont mis en évidence différents stades de développement. Nous avons identifié des dépôts de versant anciens, caractérisés par des débris fortement végétalisés, souvent altérés, ainsi qu’une texture de dépôt plus ouverte. Ces dépôts anciens ont été mis en place suite au retrait de l’Inlandsis laurentidien dans la vallée, potentiellement lors de la période de transgression/régression marine de la mer postglaciaire de Tyrrell. À l’inverse, des dépôts plus récents ont été identifiés, souvent dépourvus de végétation de surface et caractérisés par un talus à forte pente. La distribution des débris est intimement liée à la lithostratigraphie de la paroi, notamment en ce qui a trait aux différents processus de démantèlement du substrat, le basalte se fracturant en blocs massifs alors que les couches sédimentaires subhorizontales se débitent en fragments aplatis et de plus petite taille. Ce démantèlement est causé par des processus périglaciaires toujours actifs aujourd’hui, tels que la gélifraction et le soulèvement gélival. Sur une échelle de temps plus longue, la vallée a assurément subi un ajustement paraglaciaire suite au retrait de l’Inlandsis laurentidien à partir de 8000 ans BP, ayant permis la formation des talus d’éboulis qu’on voit actuellement sur le terrain. De nos jours, bien que des chutes de blocs surviennent sporadiquement, un couvert arbustif dense situé au pied des talus - et également à mi-pente - suggère que l’ampleur des processus actuels est limitée.

Par ailleurs, la neige semble avoir un rôle important dans la redistribution des débris à la surface des talus. Des caméras à déclenchement automatique, installées à l’été 2017, ont monitoré le versant sud-ouest de la vallée Tasiapik pendant près d’une année, soit du 9 août 2017 au 15 août 2018. Le principal processus observé, les avalanches, ont été fréquentes (39 événements observés sur caméras, une vingtaine sur le terrain), particulièrement vers la fin du printemps. Dans plusieurs cas, elles ont été déclenchées par des chutes de corniche survenant sur la crête de la cuesta. Les conditions météorologiques, combinées à la topographie particulière du site tel un revers de cuesta faiblement incliné permettant un transport efficace de la neige par le vent, ont favorisé la formation puis la rupture de la corniche. Les avalanches printanières transportent une grande quantité de

débris rocheux vers le bas de la pente, résultant en des dépôts sales et compacts. Certains dépôts ne sont situés qu’à quelques mètres de la route, témoignant du danger potentiel encouru par ses usagers. Les images fournies par les caméras au cours des prochaines années aideront à mieux comprendre la dynamique avalancheuse dans la vallée, mais pourront aussi documenter la répétitivité interannuelle de ces processus.