Lake A, Ellesmere Island (83°00’N, 75°30’W), is a near-perennially ice-covered, meromictic lake located at the northern margin of the Canadian High Arctic. In this extreme limnological environment, typically only a narrow moat (< 25 m) forms along the shoreline during summer, while the remaining ice may persist for many years. Anoxic conditions within the bottom waters (monimolimnion), strong chemical stratification, limited mixing and highly seasonal sediment inputs provide conditions conducive to the formation of varves (annual units of sediment). Varves, verified by radioisotope dating, are evident in the uppermost sediments and are typically composted of a silt unit with an overlying clay cap. A conspicuous feature of the sediments involves clastic aggregates with organic matter fragments. The composition and texture of these “pellets” differ from the surrounding sediments. Their size, composition and spatial distribution suggest that the most likely source for the pellets is ice-rafting from littoral areas. Since the deposition of the pellets is dependent on reduced ice cover, the pellets’ frequency within the sedimentary record represents a quasi-century scale proxy record of reduced ice cover constrained by a varve age model.

Using this data and marker beds between multiple cores, an age-depth model for the sedimentary record was developed and used to provide dates for the pellet frequency record. This record suggests that reductions in ice cover during 20th century summers were the most prevalent of the past millennium, with less significant periods of ice mobility from the mid-1400s to early-1700s. More persistent ice cover is inferred ca. 1000-1440 and during the 1700s and 1800s. Inferences of low ice cover during the 1900s stem from increased pellet concentrations near the top of the sedimentary record, which potentially correspond to recently observed near ice-off (2000) and reduced ice cover summers (2003 and 2006; D. Mueller, pers. comm.). Regional proxy melt season temperature and paleoecological records largely corroborate the Lake A proxy ice cover record, with some discrepancies likely due to the temporal resolution of the Lake A record. Of particular note is the coherence of all records during the 1900s, inferring above mean summer temperatures and corresponding reduced ice cover during much of the century. As biological records (e.g., diatoms) from arctic lake sediments are regarded as being dependent on ice cover variability, the development of sedimentary pellets as an independent proxy ice cover record provides a means of linking these two record types and testing ecological inferences about past ice cover variability in more detail.