ROCK GLACIERS AND PERIGLACIAL ROCK-ICE FEATURES IN THE EASTERN SIERRA NEVADA; UPDATES ON CLASSIFICATION, MAPPING, CLIMATE RELATIONS, AND MONITORING CONSTANCE MILLAR 1, DAVID CLOW 2, JESSICA LUNDQUIST 3, AND ROBERT WESTFALL 1 1 USDA Forest Service, Sierra Nevada Research Center, Berkeley, CA 2 USGS, Water Resources Division, Denver, CO 3 NOAA, Earth Systems Laboratory, Boulder, CO BACKGROUND Rock glaciers and related periglacial rock-ice (R-I) features are common but little-studied landforms in high mountains where conditions are relatively dry and adequate sources of shattered rock exist. While many of these features in temperate mountains appear to be Pleistocene relicts, a large proportion shows indications of activity, suggesting persistent embedded or underlying ice. Recent retreat of “clean glaciers” worldwide has been related to global warming. Rock-mantling and rock matrices, however, insulate rock glaciers and R-I features, and their activity and equilibrium with climate appear to considerably lag clean glaciers. They remain potentially important mountain reservoirs, yet overlooked in hydrologic research. For this and other reasons, research attention to the distribution, activity, genesis, and hydrology of rock glaciers has increased in recent years. Debate on the glaciogenic vs. periglacial (permafrost) origins of these features (Clark et al. 1998) has been confounded by inconsistent terms and definitions, lack of a sufficient classification, and by the varying expression of processes that occur under climatic and environmental conditions in different regions. Rock glaciers and R-I features are abundant in cirques and canyons of the Sierra Nevada south of Lake Tahoe, especially along the eastern escarpment, where they occur in diverse forms (Figs. 1 & 2). Several large valley rock glaciers have been studied for their climatic and neo-glacial relationships (Clark, pers. comm.; Konrad & Clark 1998), and a few Pleistocene relict rock glaciers are indicated on high-resolution geologic maps of the Sierra. These features are mentioned in passing in Sierran scientific and popular natural-history literature, but aside from Clark and his students’ research, no systematic studies have been published. While several papers from Europe and North America describe types of rock glaciers, define terms, and offer classifications, none proved sufficient to identify and map the range of features in the Sierra Nevada. We undertook work described here in hope that field-based mapping and a field- friendly taxonomy will assist future mapping, for example, from aerial photography. With the reconnaissance-level monitoring we are beginning, we further hope to encourage glacial, hydrologic, and ecologic research on these landforms. We include a wider range of features in our classification than is usually discussed to highlight potential physical interrelationships of these features and possible hydrologic significance of overlooked forms. GOALS: Relative to rock glaciers and associated periglacial R-I features of the Sierra Nevada: • Develop a taxonomic classification and nomenclature • Compile maps and geo-referenced databases derived from field-mapping • Analyze geographic and climatic relations of the features • Form tentative hypotheses of process and origins • Initiate reconnaissance monitoring of rock glacier and R-I feature 1) movement, 2) meltwater, including flow, seasonal persistence, temperature, and chemistry and 3) lichen and plant cover and ages. METHODS Fig 1. Goethe Rock Glacier Fig 2. Rock Glacier near Taboose Pass Mapping and Classification During field seasons of 2000-2005, Millar mapped discrete R-I features whose location and morphology suggested glacial or periglacial origins – e.g., rock glaciers, active moraines, protalus ramparts, solifluction slumps, patterned-ground circles, nets, stripes, and intermediate features. Conventional clean glaciers and persistent snowfields without associated decomposed rock mantling were not included. Each feature was field-mapped at coarse resolution on topographic maps and later characterized with digital maps (TOPO!) for latitude and longitude (center of feature), elevation range, and local aspect. Feature size was described in four ranks (400 ha, 50ha, 5 ha, 0.5 ha); three shape ranks were scored (wider than long, ~equal length & width, longer than wide). Current activity (embedded ice) is difficult to determine by casual observation. A feature was tentatively scored Active if it was: 1) within the elevation range of clean glaciers and persistent snowfields; and had 2) oversteepened front and sides (rock glacier types only); 3) angular rocks with no or little lichen growth; 4) plant cover absent or minimal; and 5) persistent spring or stream or presence of phreatophytes (e.g., willows, rushes) at the feature’s toe. Based on observations of ~300 features, a preliminary taxonomy and nomenclature were developed. This was tested by revisits to ~20% of the originally mapped features, and addition of ~100 new features, which led to the revised classification presented here. Climate Location data for all R-I features were imported into GIS as point coverages (lat/long of centers). These were intersected with 4km2 gridded data from the PRISM climate model (Daly et al., 1994), and adjusted to site-specific elevations using lapse rates from PRISM. We extracted layers for annual, January, and July minimum and maximum temperatures, respectively, and annual, January, and July precipitation, respectively, for the period of record, 1960-1999. The PRISM grids were converted to polygons and sequentially intersected with the locations of the R-I features, grouped by the six Location Classes (see Results). GIS analyses were done in ARC/Info. To determine differences among Location Classes, we subjected the merged R-I-feature/PRISM-climate data to discriminant analysis. We classified the analysis by Location Class with the climatic measures as variables, maximizing R-I-feature differences in multivariate climate space. We then computed mean climatic data from the PRISM model for the classified groups. Monitoring Reconnaissance monitoring was initiated in summer, 2005 (Table 1). We report preliminary results for water chemistry only; other elements await downloading and/or first measurements in summer 2006. Water Chemistry. On Oct 24-25, 2005, we sampled water from springs and streams emanating at the bases of six R-I features for chemical analysis. Water sampling and analytic methods followed Clow et al., 2002. To assess differences among sites, we used principal component analysis (PCA), with ln-transformed data to normalize the distribution. We used PCA and discriminant analysis on ln- transformed data to resolve clusters and then test differences of the rock-ice meltwater chemistry data to values previously analyzed from Yosemite National Park precipitation (6 samples, 2003-04), snow (2 samples, 2005), river (123 samples, 2003-05), and lake (15 samples, 2003-05) sources (Clow et al., 2002). Water Temperature and Depth. Water temperature is being recorded by ten small dataloggers (iButtons), programmed for four readings/day and seven larger level-loggers (solinsts), programmed to read every half hour. We placed all instruments in shallow pools (15-30 cm) at the spring heads or where meltwater streams first surfaced from R-I features (Figs. 3 and 4). Readings will give indications of temperature persistence and variability, timing in change of state from liquid to frozen, and timing of spring/stream drying, if any – all of which will help interpret the internal hydrologic structure of the features. Leveloggers also record water pressure, which, together with barometric pressure measured from barologgers installed at Tuolumne Meadows, Yosemite NP (2600 m; Lundquist, ongoing research), will enable calculation of water depth. Rock-Ice Feature Movement. We installed rudimentary transects to monitor movement of boulders on the steep front of one rock glacier (RGV, see Results), upper surfaces of two other rock glaciers (RGC), and across the bases of three boulder streams (BSC). Paint spots were sprayed on boulders every 1.7 m along an azimuth perpendicular to the inferred flow direction, with base-points on bedrock off the edge of the R-I features. Transects will be re-visited to determine if paint spots moved. Fig 3. Spring, Kidney Lk Boulder Stream Fig 4. Stream surfacing, Kuna Crest Rock Glacier RESULTS & DISCUSSION Classification and Mapping From field surveys, 395 R-I features were mapped (Appendix 1). A taxonomic classification, in which geomorphic relationships are inferred by taxonomic relationship, is proposed (Table 2). The classification describes three hierarchic levels: 4 Condition States, 6 Location Classes, and 18 Position Types. Condition States are the major categories of R-I features; Location Classes describe the location of features within the Condition State; and Position Types are the terminal categories. Type localities for each of the Position Types are identified, and a subset is illustrated (Figs 5-12). The classification is based on morphology and location, does not rely on inference of glacial vs. periglacial origins, and is intended for easy field use. Fig 5. RGC-V, Mt. Gibbs Fig 6. RGC-M, Mt. Kuna Fig 13. Slope aspects for Location Classes Summary of Main Groups Geographic location, size, and shape means for the Condition States and Location Classes are given in Table 3, and aspects are indicated by wind rose diagrams for the Location Classes in Fig. 13. Rock Glaciers/Periglaciers: