Why University of Manitoba?
The Department of Geological Sciences at University of Manitoba is ideally suited for this comprehensive multidisciplinary field and laboratory study of past hydrological conditions in these lakes because the analytical facilities and logistical/technical support are among the best in Canada. The Lake Sedimentology Laboratory maintains an excellent collection of field equipment required for this project, including boats, snowmobile, field chemical analyses kits, GPS units, underwater survey equipment, coring and sampling devices. The analytical facilities within the Department are first-rate. The Microbeam Facility in Geological Sciences is set up as a Regional Facility with a state-of-the-art electron microprobe, SIMS, LAM-HR-ICP-MS, GS-IRMS (including a gas bench system for C and O isotopic analyses of carbonates and waters), and a scanning electron microscope and image analysis system. The Geochemistry Laboratory contains ICP-OES and AAS. Powder x-ray diffraction facilities are available in the X-ray Laboratory and the Sedimentology Laboratories are set up with epi-florescence and phase contrast microscopes, automated particle size equipment and image analysis systems, and all core description, photography and subsampling facilities. The advantages of having this breadth of analytical equipment and technical expertise in-house are that I and my research colleagues can tailor our analytical procedures for the specific goals of this project and results can be produced with a minimum of lead time. Furthermore, the in-house facilities largely eliminate the necessity of utilizing external laboratories, which are often costly and an inefficient use of the Strategic Project funds.
Finally, I feel that my previous geolimnological research and that of my students’ on the modern sediments of the salt lakes of the Great Plains has now put us in a position of being able to make important advances in the understanding of the paleolimnological records in these lakes and to tackle the difficult task of discriminating between natural and anthropogenic fluctuations in these basins. In 1990 I began a long-term study of the water chemistry, sediments, sedimentary processes, and geochemistry of the lakes in this region. This work has made significant contributions to our knowledge of the character and variability of modern continental brine systems and the processes of deposition and diagenesis of lacustrine sediments in western Canada. With the impetus of the significant historical water level changes that are now evident in these lakes, applied geolimnological and paleolimnological research is now entering a very exciting phase that can make use of the past decade of research results from my laboratory.
Multidisciplinary Approach
I will use a multiple-indicator paleoenvironmental approach to formulate the history of changes in hydrology and ecological/environmental conditions within the basins and their catchments, and then combine these data with results from other published and ongoing studies to identify the climatic changes in the northern Great Plains region. To understand and reconstruct the often complex paleohydrological dynamics of lakes in the Great Plains, use of a multiple-indicator paleoenvironmental approach is essential. Reliance on only one or a few stratigraphic components (such only siliceous microfossils or mineralogy alone) is unlikely to provide a scientifically-defensible lake level history. For example, changes in physical indicators, such as an increase in grain size or a decrease in organic matter content, could be interpreted as evidence of decreased water levels, but without supporting evidence from sedimentary structures, sedimentary petrography, geochemical and isotopic signatures or endogenic mineral composition corroborating low water episodes, questions would remain as to whether local catchment disturbances (such as local erosion due to vegetation change, etc.) could ultimately have been responsible (see, for example, Campbell et al., 2000a; Hall et al., 2003; Teller and Last, 1990; Wolfe et al., 2006). Consequently, multiple independent lines of evidence from a variety of proxy indicators are required to establish a reliable lake level history and, in particular, to assess the role of climatic variability in observed and interpreted hydrological changes.
All interpretations will be based on the most sensitive and comprehensive methods available. Analyses of sediment cores, subaerial beach and shoreline sediments, as well as tufas and bioherms will be integrated with data from physical (sedimentary structures, texture and fabric, petrography, mineralogy), geochemical (major and trace element composition of beds and laminae, stable isotopes of biogenic and inorganic components), and limited biological indicators (microbialite composition). These data will be used to generate reliable inferences of past hydrological and environmental variability within each basin. Finally, multivariate- and time-series numerical analyses will be used to assess the hydrological and ecological responses to changes in climate during he past millennia.
Field
I will use a variety of field sampling techniques, including short (~30 cm) and long (1-2 m) core acquisition from boats during summer and from the stable ice cover during winter, SCUBA-based sampling of subaqueous microbialites, tufas, and bioherms, and geological facies mapping and paleoshoreline delineation. Two of the lakes, Manito and Deadmoose, will require a transect of core and sample locations because of the complexity of the basins’ morphology and because sedimentation rates are unlikely to be consistent and linear in these basins over the timeframe being investigated (Ginn et al., 2005; Last and Slezak, 1986). In addition, because Waldsea and Deadmoose lakes both contain permanently stratified bodies of water (Hammer, 1994), additional cores must be taken from areas located above and below the chemocline.
Some archived cores from my previous investigations of these lakes and from the GSC’s IRMA project (Vance, 1997a) can be used to help supplement sample requirements in the early stages of the work. However, given the tight time-line for this project, it is critical that most or all of the field work, core and sample collection be completed in the first six to eight months of the study. Field mapping, including all shoreline microbialite and nearshore biohermal sampling, will be done as soon as possible in the ice-free season of 2008. Short cores (20-30 cm in length) will be collected from the basins during early summer, 2008, using a modified K-B gravity corer (Glew, 1989). Longer sequences of lacustrine offshore sediment will be collected using a modified Livingstone piston corer (Wright, 1991) as soon as a stable ice cover forms on the lakes (probably December, 2008 to January, 2009). All core and sample material will be transported to the University of Manitoba and stored in a cold room at 4oC.
Laboratory
In the laboratory, cores and samples will be subjected to a series of analyses, based on the knowledge gained from my previous research, that will be used to identify environmental and limnological responses to past hydrological changes, and to assess the role of climatic variability. In all cases, I and my research colleagues involved in this project will follow well-established experimental protocols as summarized in various published methodological papers and handbooks (e.g., Francus, 2004; Hinton et al., 1995; Last and Smol, 2002a; Last and Smol, 2002b; Leng, 2006; Rutter and Catto, 1996; Samson et al., 2003; Sylvester, 2001; Tucker, 1988; Warner, 1990). Briefly, following core and sample photography, x-radiography, description and preliminary petrographic analysis, material will be subsampled at appropriate intervals to establish a stratigraphic chronology. This chronostratigraphy will help dictate the subsampling interval for subsequent analyses. Establishing a well-constrained chronology for the cores is absolutely critical to the success of this project and will depend mainly on accelerator mass spectrometry (AMS) 14C measurements of terrestrial material (plants or insects). Our previous experience in the Manito and Waldsea basins suggests that biogenic carbonates in these lakes do not possess significant hard-water effects (Ginn, 2007; Last and Schweyen, 1985). Thus, the abundant organic carbonates in these lakes are also a source of datable material. All 14C measurements will be determined at either Isotrace Laboratories (Toronto) or the KCCAMS Facility (University of California – Irvine) and converted to calendar years using INTCAL98 (Stuiver et al., 1998). Chronology for the past ~150 years will be determined on the basis of 210Pb dating using Flett Research Ltd (Winnipeg).
Once preliminary chronostratigraphy is established, subsampling will continue for detailed physical, mineralogical, petrographic, geochemical and limited biological analyses. Physical and sedimentological analyses will emphasize bedding features and sedimentary structures, mineralogical composition and petrography, moisture and organic matter contents, and texture and fabric. The rationale and importance of this suite of analyses to lake level history studies are discussed by Dean et al. (1999), Kemp et al. (2001), Last (2002a; 2002b), Schnurrenberger et al. (2003) and Teller and Last (1990). As reviewed by Last and Vance (1997) and Kemp (1996), many of these physical features and data can yield information about past limnological and hydrological conditions in the basin and catchment. For example, sedimentary structures can provide unambiguous information about the physical and chemical structure of the overlying water mass, the nature of the sediment transporting mechanisms, and the level of energy at the depositional site - factors which often can be quantitatively related to basin depth and lake-level fluctuations (e.g., Allen, 1982; Collinson, 1978; Collinson and Thompson, 1982; Curran, 1985; Demicco and Hardie, 1994). Non-annual laminae (from Waldsea and Deadmoose), cryptalgal laminae and mat structures (possibly both annual and non-annual from Manito), exposure horizons and incipient soil horizons, graded bedding and current-generated laminae (from Antelope) will likely form the basis of much of the paleoenvironmental interpretations. In addition, by undertaking digital imagery it will be possible to conduct quantitative time-series analyses of any laminated sequences at a high resolution to gain information about the frequency and cyclicity of key depositional and erosional processes.
Study of the various textural and fabric parameters in these lake sediments can lead to important quantitative information about: i) source (provenance) of the deposits, ii) the mechanisms responsible for the transport of the material, iii) past physical and chemical conditions at the depositional site, and iv) paleoclimatic and paleohydrological conditions within the surrounding watershed. Interpretation of textural data is not always unambiguous, thus requiring that these fundamental measurements be incorporated within a multiple-indicator approach. For example, at Pine Lake in south-central Alberta, Campbell et al. (2000a) provide convincing arguments for relatively coarse sediment reflecting intervals of increased stream-flow and humid conditions, and fine-grained sediment corresponding to periods of lower stream-flow, lower water levels and, therefore, increased aridity. Conversely, the opposite scenario (i.e., coarse sediment reflecting low lake levels and more arid conditions, and finer-grained sediment associated with high lake levels and humid climates) has been used as a template to decipher past hydrological conditions in Lakes Manitoba and Winnipeg in the eastern Great Plains (Henderson and Last, 1998; Last and Teller, 1983), Harris Lake, Saskatchewan (Last and Sauchyn, 1993), Ceylon Lake, Saskatchewan, (Last, 1990), Clearwater Lake, Saskatchewan (Last et al., 1998; Leavitt et al., 1999) and many other lakes on a global basis (e.g., Digerfeldt, 1986; Teller and Last, 1990). Textural interpretations may be further complicated by changing provenance of the material and different mechanisms of transport of sediments. For example, in Oro Lake, Saskatchewan, a relatively arid climate during the early and mid-Holocene resulted in an aeolian influx of fine-grained sediment to the basin (Last and Vance, 2002), whereas in Antelope Lake, the same arid conditions are probably associated with relatively coarse-grained sediments (Last and Vance, 1996; Vance, 1997a).
Knowledge of the minerals comprising the inorganic components of the lake sediment can provide information related to the genesis of the sediments, transport mechanisms, and past limnological, hydrological and climatic conditions. These various mineralogical and petrographic analyses allow us to quantitatively decipher a tremendously wide array of paleoenvironmental and paleohydrological conditions. From a genetic perspective, the sediments of the study lakes are comprised of a mixture of allogenic (i.e., from outside the lake) material and endogenic (from within the basin) components. The allogenic fraction will likely consist of a mixture of clay minerals, quartz, carbonates, feldspars, and ferromagnesian minerals and relative proportions of these detrital mineral groups will generally be similar to that of the surrounding glacial debris. However, less chemically-stable constituents are susceptible to loss by weathering processes within the drainage basin. Last and Sauchyn (1993), Teller and Last (1981), and Henderson and Last (1998) used this feature, expressed as ratios of total feldspars/quartz, plagioclase/K-feldspars, and calcite/dolomite, to indicate chemical ‘weathering intensity’ in the drainage basins of Harris Lake, Lake Manitoba, and Lake Winnipeg. Weathering intensity values can then be related to paleoprecipitation and temperature in the watershed.
Even more specific and quantitative paleohydrological information can be derived from the endogenic fractions. Many studies in the Great Plains and elsewhere have documented the application of calcite, Mg-calcite, dolomite, and aragonite in paleolimnology (e.g., Dean, 1981; Dean and Fouch, 1983; Dean and Megard, 1993; Haskell et al., 1996; Kelts and Hsu, 1978). It has become the conventional method to use the stratigraphic variation of these endogenic carbonate species in lacustrine sequences to deduce past Mg/Ca ionic ratios and salinity of the lake water. Carbonate mineralogical data can be particularly useful in deducing past lake water composition and salinity when interpreted in conjunction with stable isotope analyses of either the inorganic carbonates or biological components as shown by Last and Vance (2002), Campbell et al. (2000a), Last et al. (1994) and Van Stempvoort et al. (1993) elsewhere in western Canada. Finally, the endogenic carbonate microcrystal morphology and fabric will be used to evaluate water column stratification, depth and paleochemistry in the Waldsea and Deadmoose basins following the methodology of Last et al. (2002b), Last and Vance (1997), and Greengrass et al. (1999). Fluid inclusion analyses will also be used in the coarsely crystalline salts of Waldsea and Deadmoose to help ascertain brine chemical and temperature changes (Lowenstein and Brennan, 2002).
The major geochemical indicators (major and trace elements and δ13C and δ18O of both endogenic inorganic carbonate and biogenic fractions) will be used to help track past changes in lake water balance, water composition, and source. Stable carbon and oxygen isotope systematics of carbonate minerals have been studied for decades and the utility of these isotope studies to lake studies is well documented (e.g., Benson et al., 1996; Casanova and Hillaire-Marcel, 1992; Cohen et al., 1997; Ito, 2002; Leng et al., 2006; Talbot, 1990). Despite the extremely widespread application of δ13C and δ18O in paleolimnological studies on a global basis, surprisingly little has been done on the lacustrine basins of western Canada. In particular, the tandem approach of using stable isotope analyses in conjunction with trace and major element geochemistry of the endogenic carbonates in the lakes has been shown to be an effective analytical technique (e.g., Fritz et al., 2000; Ingram et al., 1998; Valero-Garcés et al., 1997; Xia et al., 1997). This analytical approach will be used in Manito, Waldsea, and Deadmoose lakes to help constrain the paleolimnological and hydrological reconstructions.
Other paleolimnological studies in the northern Great Plains have successfully used sedimentary diatom and ostracode assemblages to identify fluctuations in lake water salinity and composition (e.g., Last et al., 1994; Last et al., 1998; Leavitt et al., 1999; Michels et al., 2007; Patoine and Leavitt, 2006; Porter et al., 1999; Risberg et al., 1999b). However, the elevated salinities and high alkalinities in the study lakes will limit the application of these proxies (Boden, 1985; Last, 1991a; Last and Kelley, 2002; Schweyen, 1984; Schweyen and Last, 1983). Thus, within the time constraints of this two-year project, biological investigations will be limited to identification of the cyanobacteria and other halophilic microbiota (Moore et al., 1983; Papineau et al., 2005; Reid et al., 2000) in the Manito Lake stromatolites, thrombolites and leiolites. These data can then be used, together with the structural morphology, position in the basin, and inorganic composition to decipher past ecological and water level conditions in this basin (cf. Dupraz and Visscher, 2005).
