Raymond A. Assel

Recent Trends In Laurentian Great Lakes Ice Cover

A 39-winter (1963–2001) record of annual maximum ice concentration (AMIC), the maximum fraction of lake surface area covered by ice each year, is analyzed for each Great Lake. Lake Erie has the largest median AMIC (94%) followed by Lakes Superior (80%), Huron(63%), Michigan (33%), and Ontario (21%). The frequency distributionof AMICs is negatively skewed for Lakes Superior and Erie and positively skewed for Lakes Michigan and Ontario. Temporal and spatial patterns of typical and extreme AMICs is presented within the context of long-term average air temperatures and lake bathymetry. The variation of spatially averaged ice concentration with discrete depth ranges are discussed for each lake for the upper and lower end of the typical range of AMIC values. In general, ice concentration decreases with increasing depth ranges for a given winter. A decrease in the gradient of ice concentration with depths was also observed with an increase in the AMIC from winter 1983 to winter 1984. A temporal trend in the AMICs supports the hypothesis of three ice cover regimes over the past 39 winters. Approximately 44% of the highest quartile (10 highest) AMICs for the Great Lakes occurred during the 6-winter period:1977–1982 providing evidence of a higher ice cover regime during thisperiod relative to the 14 winters before them (1963–1976) and the 19 winters after them (1983–2001). Winter 1998 established new low AMIC extremes,and the AMIC averaged over the 1998–2001 winters is the lowest for theperiod of record on four of the five Great Lakes. These recent trends taken together are noteworthy as they may be harbingers of a period of even lower AMICs in the 21st Century.

HYDROCLIMATIC FACTORS OF THE RECENT RECORD DROP IN LAURENTIAN GREAT LAKES WATER LEVELS

Abstract An extreme low-water supply episode from 1997 to 2000 resulted in the largest 1-yr drop in Lakes Michigan–Huron and Lake Erie water levels (0.92 and 1.03 m, respectively) recorded since measurements began in the early 1800s. Lake Superior water levels were the lowest since 1925. Lakes Erie and Ontario also had relatively low levels. The episode was unusual, particularly when compared to the record-low water episode of the mid-1960s, in that the primary hydroclimatological driver was high air temperatures and not extremely low precipitation. The high air temperatures resulted in unusually high lake evaporation rates and decreased basin runoff. The drop in levels during this episode was compared to other 1–3-yr decreases throughout the period of record. A comparison of the 1997–2000 episode for Lakes Michigan–Huron with the 1960–64 episode, which led to record-low lake levels in 1964, shows that the various elements of the water balance have differing importance in the two episodes.

A one‐dimensional ice thermodynamics model for the Laurentian Great Lakes

Great Lakes hydrologic research requires the use of continuous-simulation daily ice cover models over long time periods in the absence of field observations. They must be physically based, rather than statistically based, for use under conditions different than those under which they were derived. But they also must match existing conditions for which data exist. A review discloses that existing ice dynamics models do not meet all of these criteria; a new one that does is based here on a prismatic ice pack heat balance, ice growth and temperature constraints, and thermodynamic flux terms from companion water heat balance and storage equations. The prismatic ice model is a good first step to understanding complex geometries and is supportable through the use of lake-averaged energy fluxes. The ice model is integrated into an existing lake thermodynamics and one-dimensional heat storage model, and the resulting combination is calibrated for Laurentian Great Lakes applications. Simulation experiments are used to analyze the models strengths and limitations and to explore its relevance. Comparisons between model output and existing data allow consideration of the ice climatology of the Great Lakes; the climatology description is extended through use of the new model. Promising potential model extensions include spatial extension, additional parameterizations for wind-ice movement, snow, and albedo, and inclusions of remotely sensed data.

Atmospheric teleconnection patterns and severity of winters in the Laurentian Great Lakes basin

Abstract We analyzed the relationship between an index of Great Lakes winter severity (winters 1950–1998) and atmospheric circulation characteristics. Classification and Regression Tree analysis methods allowed us to develop a simple characterization of warm, normal and cold winters in terms of teleconnection indices and their combinations. Results are presented in the form of decision trees. The single most important classifier for warm winters was the Polar/Eurasian index (POL). A majority of warm winters (12 out of 15) occurred when this index was substantially positive (POL > 0.23). There were no cold winters when this condition was in place. Warm winters are associated with a positive phase of the Western Pacific pattern and El Niño events in the equatorial Pacific. The association between cold winters and La Niña events was much weaker. Thus, the effect of the El Niño/Southern Oscillation (ENSO) on severity of winters in the Great Lakes basin is not symmetric. The structure of the relationship between the index of winter severity and teleconnection indices is more complex for cold winters than for warm winters. It takes two or more indices to successfully classify cold winters. In general, warm winters are characterized by a predominantly zonal type of atmospheric circulation over the Northern Hemisphere (type W1). Within this type of circulation it is possible to distinguish two sub‐types, W2 and W3. Sub‐type W2 is characterized by a high‐pressure cell over North America, which is accompanied by enhanced cyclonic activity over the eastern North Pacific. Due to a broad southerly “anomalous” flow, surface air temperatures (SATs) are above normal almost everywhere over the continent. During the W3 sub‐type, the polar jet stream over North America, instead of forming a typical ridge‐trough pattern, is almost entirely zonal, thus effectively blocking an advection of cold Arctic air to the south. Cold winters tend to occur when the atmospheric circulation is more meridional (type C1). As with warm winters, there are two sub‐types of circulation, C2 and C3. In the case of C2, the jet stream loops southward over the western part of North America, but its northern excursion over the eastern part is suppressed. In this situation, the probability of a cold winter is higher for Lake Superior than for the lower Great Lakes. Sub‐type C3 is characterized by an amplification of the climatological ridge over the Rockies and the trough over the East Coast. The strongest negative SAT anomalies are located south of the Great Lakes basin, so that the probability of a cold winter is higher for the lower Great Lakes than for Lake Superior.

Atmospheric teleconnections for annual maximum ice cover on the Laurentian Great Lakes

Great Lake ice cover records for winters 1963–1990 were used to define anomalously high (low) average ice cover based on the seven highest (seven lowest) annual maximum ice covers. Analysis of the maximum ice cover reveals (i) a low (1964–1976); (ii) a high (1977–1982); and (iii) once again a low (1983–1990) ice cover regime. The high ice cover regime corresponded in part with a hiatus in El Nino–Southern Oscillation (ENSO) events and the beginning of an interdecadal change in Northern Hemisphere atmospheric circulation that started in the late 1970s. About 46% of the lowest quartile ice covers occurred during the mature phase (year+one winter) of El Nino. Only 1 year out of seven with the mature phase of El Nino between 1963 and 1990 was not associated with the lowest quartile ice cover, this was 1977, a pivotal year after which a new climatic regime in the Northern Hemisphere was established. Anomaly maps of 700 hPa geopotential height for the lowest quartile ice cover reveal a zonal flow pattern. Highest quartile ice cover was associated with meridonal circulation from the Arctic directed toward the Great Lakes. Significant differences occur for highest minus lowest quartile ice cover composite 700 hPa height anomaly maps in the Pacific Ocean, the west coast of North America, north Mexico, eastern North America, north central Siberia, western Europe and the adjacent North Atlantic. Correlations between first differences (year t+1 minus year t) of annual maximum ice cover and 700 hPa geopotential heights for winters 1963–1990 agrees with these teleconnections and were higher than the absolute time series correlations, indicating strong interannual teleconnections. Annual maximum ice cover was also significantly correlated with the tropical Northern Hemisphere teleconnection index.

Implications of CO2 global warming on great lakes ice cover

Statistical ice cover models were used to project daily mean basin ice cover and annual ice cover duration for Lakes Superior and Erie. Models were applied to a 1951–80 base period and to three 30-year steady double carbon dioxide (2 × CO2) scenarios produced by the Geophysical Fluid Dynamics Laboratory (GFDL), the Goddard Institute of Space Studies (GISS), and the Oregon State University (OSU) general circulation models. Ice cover estimates were made for the West, Central, and East Basins of Lake Erie and for the West, East, and Whitefish Bay Basins of Lake Superior. Average ice cover duration for the 1951– 80 base period ranged from 13 to 16 weeks for individual lake basins. Reductions in average ice cover duration under the three 2 × CO2 scenarios for individual lake basins ranged from 5 to 12 weeks for the OSU scenario, 8 to 13 weeks for the GISS scenario, and 11 to 13 weeks for GFDL scenario. Winters without ice formation become common for Lake Superior under the GFDL scenario and under all three 2 × CO2 scenarios for the Central and East Basins of Lake Erie. During an average 2 × CO2 winter, ice cover would be limited to the shallow areas of Lakes Erie and Superior. Because of uncertainties in the ice cover models, the results given here represent only a first approximation and are likely to represent an upper limit of the extent and duration of ice cover under the climate change projected by the three 2 × CO2scenarios. Notwithstanding these limitations, ice cover projected by the 2 × CO2 scenarios provides a preliminary assessment of the potential sensitivity of the Great Lakes ice cover to global warming. Potential environmental and socioeconomic impacts of a 2 × CO2 warming include year-round navigation, change in abundance of some fish species in the Great Lakes, discontinuation or reduction of winter recreational activities, and an increase in winter lake evaporation.

Laurentian Great Lakes Ice and Weather Conditions for the 1998 El Niño Winter

Abstract Winter 1997/98 occurred during one of the strongest warm El Nino events, and the Great Lakes experienced one of the least extensive ice covers of this century. Seasonal maximum ice cover for the combined area of the Great Lakes was the lowest on record (15%) relative to winters since 1963, a distinction formerly held by winter 1982/83 (25%), which was also an exceptionally strong El Nino winter. Maximum ice covers set new lows in winter 1997/98 for Lakes Erie (5%), Ontario (6%), and Superior (11%), tied the all—time low for Lake Huron (29%), and came close to tying the all—time low on Lake Michigan (15%; all—time low is 13%). Here the authors compare seasonal progression of lake—averaged ice cover for winter 1982/83, winter 1997/98, and a 20—winter normal (1960—79) derived from the NOAA Great Lakes Ice Atlas and discuss the 1997/98 ice cover in detail. Winter air temperatures in the Great Lakes were at or near record high levels, storms were displaced farther to the south over eastern North Ameri...

Fall and Winter Thermal Structure of Lake Superior

Temperature surveys were made along the normal upbound (westward) and down-bound (eastward) shipping lanes across Lake Superior to document fall and winter thermal structure of that lake. This work was done as part of the Congressionally-funded Demonstration Program to Extend the Navigation Season on the Great Lakes and the St. Lawrence Seaway. Surveys were made aboard ore carriers using a portable bathythermograph (BT) system and expendable BT probes. Surveys usually took 2 to 4 days to complete. Twenty-one surveys were made during the winters of 1973 to 1976 and 25 surveys were made during the falls of 1976 to 1979. Mean seasonal temperature trends identified from these data include: (1) approximately exponential increase in fall mixed layer depth through early to mid-November, (2) maximum value of average mixed layer and upper 25-m layer temperatures between the end of August and mid-September, (3) maximum value of average water column temperature in late September, (4) isothermal conditions between mid-November and mid-December, (5) completion of fall overturn in December and winter restratification in December or January depending primarily upon winds, (6) average winter [January to March] monthly mixed layer depth between 60 m and 100 m and, (7) minimum value of average water column temperature in late March. Midlake and nearshore thermal regimes were identified. These thermal regimes show agreement in trend with lake bathymetry, wind fetch, and lake circulation patterns. Deeper areas with longer wind fetch in both thermal regimes have the deepest mixed layers during winter. Areas having the combination of greater depth and larger wind fetch, midlake areas in most cases, also tend to have higher column temperatures and ice cover of short duration in winter and lower column temperatures in summer relative to adjacent areas.

Winter 1994 Weather and Ice Conditions for the Laurentian Great Lakes

Abstract The Laurentian Great Lakes developed their most extensive ice cover in over a decade during winter 1994 [December-February 1993/94 (DJF 94)]. Extensive midlake ice formation started the second half of January, about 2 weeks earlier than normal. Seasonal maximal ice extent occurred in early February, again about 2 weeks earlier than normal. Winter 1994 maximum (normal) ice coverages on the Great Lakes are Lake Superior 96% (75%), Lake Michigan 78% (45%), Lake Huron 95% (68%), Lake Erie 97% (90%), and Lake Ontario 67% (24%). Relative to the prior 31 winters (1963–93), the extent of seasonal maximal ice cover for winter 1994 for the Great Lakes taken as a unit is exceeded by only one other winter (1979); however, other winters for individual Great Lakes had similar maximal ice covers. Anomalously strong anticyclonic circulation over the central North Pacific (extending to the North Pole) and an abnormally strong polar vortex centered over northern Hudson Bay combined to produce a circulation pattern...

GREAT LAKES ICE THICKNESS PREDICTION

Abstract Weekly ice thickness data, collected from 24 bay, harbor, and river sites on the Great Lakes, were correlated with freezing degree-day accumulations to develop regression equations between ice thickness and freezing degree-days. The data base at ice measurement sites was 3 to 8 winters in length. Ths standard error of estimate varied for individual regression equations and averaged between 7 and 8 cm for five forms of regression equations. Because the regression equations are empirical, the range of input data used to predict ice thickness should be limited to the range of values used in the derivation.