Introduction

Lake Erie is the shallowest of the Great Lakes. Bathymetric charts (Fig. 1) reveal that it can be subdivided into three distinct areas: the western sub-basin, which extends from the western shoreline to the islands north of Sandusky Bay; the central sub-basin, which includes the area between the islands and the Long Point–Erie axis; and lastly the eastern sub-basin, which extends from the Long Point–Erie axis to the Niagara River outlet of the lake at Buffalo. From west to east the sub-basins get progressively deeper as may be seen on Figure 1.

Fig. 1. Bathymetric map of Lake Erie.

Ice forms in late November or early December and usually lasts until mid-April. Except for Reference StewartStewart (1973), little if any work has been done on currents or thermal patterns in winter. Reference HutchinsonHutchinson ([^c1957]) has pointed out that orientation of the lake in relation to the prevailing westerly winds of the area favors a net transport of surface water from the north shore toward the south shore and a northerly current along the bottom. Thus, upwelling along the north shore is to be expected. Reference HartleyHartley (1968) found a general south-east–to–north-west circulation in the summer with bottom currents moving toward the north shore between Pelee Point and Point Aux Pins (Rondeau Harbor), easterly currents along the north shore from Point Aux Pins and Point Burwell, and a northerly current swinging around Long Point into Long Point Bay. Thermal studies by Reference IrbeIrbe (1969) confirm these general patterns.

Since the winter of 1972–73, visible- and thermal-band images of North America from NOAA satellites have been available for the study of ice on the Great Lakes. The NOAA series of polar-orbiting satellites collect two thermal images and one visible image per day with a spatial resolution of about 1 km; the Geostationary Observational Environmental Satellites (GOES) collect visible (day-time) and thermal (day-time and night-time) infrared images every 30 min. NASA’s polar-orbiting Landsat satellites have an 18 d revisit cycle (which becomes a 9 d cycle when two satellites are orbiting), with superb spatial resolution (79 m) in the visible and reflected infrared regions. Satellite images from all three satellite systems were used in this study. Prior studies of Great Lakes ice cover that utilized NOAA satellite imagery include Reference Strong, Thomson, Thomson, Lane and CsallanyStrong (1973), Reference LeshkevichLeshkevich (1976), Reference McMillan and ForsythMcMillan and Forsyth (1976), and Reference WarthaWartha (1977).

Typical Break-Up Pattern

On Lake Erie in a typical winter, one may anticipate only one or two days of total ice cover before break-up begins. The first sign of break-up always appears in the area of Pelee Point, and extends along the east edge of Kelley’s Island, Pelee Point, and the small islands at the edge of the shallow (7 m deep) western basin, usually as a result of strong north westerly winds (Fig. 2). At about the same time, leads develop along the north shore, and the ice is compressed at right angles to the axis of the lake, resulting in ridging, thrusting, and a general pile-up of ice along the south shore. Break-up is thus initiated in the central sub-basin and in the western sub-basin by north-westerly or westerly winds. The western sub-basin ice, which is held in the sub-basin by a bottom ridge and a string of islands, does not break up at this time, although leads may develop along the north shore (Fig. 3).

Fig. 2. GOES image 7 February 1968, 16.30 Z. A typical onset of break-up, with a lead widening east of the islands and a westerly wind. The ice in the western basin is denoted by A; moving ice in the central basin by H.

Fig. 3. NOAA-4 VHRR (visible), 14 February 1976, 15.00 Z. Note the ice-free areas. Note the clearing of the western portion if the central sub-basin and the leads developing along the north and west shores of the western sub-basin.

The tension produced in the ice mass of the large central sub-basin causes fractures as well as ridges. When the north-west wind subsides or when the wind changes direction, and the ice is no longer under tension, floes form from the multifractured sheet. The broken ice in the central sub-basin is then free to drift and move with the current and wind. It tends to move eastward with the current toward the Niagara River (Fig. 4). The satellite images show large floes at this time, but as the ice moves eastward it becomes constricted between Long Point and Erie, Pennsylvania, where abrasion and additional fracturing occurs.

Fig. 4. NOAA-4 VHRR (visible), 23 February 1976, 15.00 Z. Note the eastward migration of the mobile ice pack, the solid look of the ice near Buffalo and the grounded ice in the western sub-basin.

The western sub-basin, more than 60 m deep off Long Point, receives these floes. In Long Point Bay, however, the winds usually clear out the ice in late February (Fig. 4). The central sub-basin, then, clears systematically from west to east, while the western and eastern sub-basins remain ice covered.

At the eastern extreme of the lake, thawing begins in the Toledo area, aided by industrial and municipal heated water effluents and by the water of the Detroit River, which has also been warmed by industrial and municipal effluents. Soon the ice of the western sub-basin ice is totally detached from shore and is free to be moved by the wind, usually eastward toward the lake islands (Fig. 4). But the vast bulk of lake ice is still offshore at the eastern end of the lake. Commonly, ridges will form as the ice is forced eastward and is compressed by wind and current action.

The western basin is normally free of ice in mid- or late March (Fig. 5). The average opening date for ships in western Lake Erie (Cleveland) is 24 March Reference Boyce(Boyce, 1973). In Buffalo, however, the average historical opening of the harbor to shipping is 11 April, eighteen days later. We can see from Figures 3–5 why this is so. The ice, carried by wind and current, moves toward the outlet of the lake near Buffalo. Here it remains until it melts in April, impeding lake shipping not only in Buffalo harbor but also at the exit of the Welland Canal. In 1976, the ice at Buffalo was last recorded by NOAA-4 on 19 April.

Fig. 5. NOAA-4 VHRR (visible), 24 March 1976, 15.00 Z. The only ice in Lake Erie is the “plug” at Buffalo. Note the straight, wind-formed edge of the rotting ice.

1972–73 Ice Season Break-Up

Freeze-up began early in November 1972 Reference Assel(Assel, 1974). Below-normal temperatures in the first half of January brought ice formation, but a thaw in late January and early February eased ice conditions. On 8 February, the wind shifted to the north-west as a cold front moved through the area, and the mean daily temperature dropped to —5.6°C at Detroit. Figure 6 (9 February 1973) shows older ice accumulating in the central basin with ice accumulating west of Long Point as well. New clear ice is probably forming all along the north shore.

Fig. 6. NOAA-2 VHRR (visible), 9 February 1973, 15.00 Z. Older ice is accumulating in the central sub-basin. Clear new ice is probably forming all along the north shore and in the western sub-basin.

On the next day’s imagery (Fig. 7, 10 February 1973) ice could be seen forming along the south shore, and the area of mobile pack ice in the center of the lake had enlarged noticeably. The winds in the eastern basin, however, were from the north setting up a wind shear over the central basin. These winds actually moved the pack to the west as they became north-easterly blowing out of a large high-pressure center moving to the north. The effect of three days of easterly winds is seen in the distribution of the lake ice on 13 February (Fig. 8). Ice has been forced into the bay north of Long Point; the area between Long Point and the south shore from Buffalo to Erie is ice-free and the central basin continues to hold its mobile pack ice. The easterly winds persisted until 15 February. On 14 February the temperature rose above the freezing point, and on 15 February the wind again shifted to the north-west and the ice which by then totally covered Lake Erie was moved to the south. Evidence of this is seen on Figures 9 and 10 (17 February 1973 was the coldest day of the month in Cleveland). Here the shape of the north shore line is preserved in the ice, the older, snow-covered, and more reflective ice has been forced against the south shore, and between Long Point and Erie, Pennsylvania, the ice from Long Point and along the Canadian shore has been moved south and east forming a thick ice dam across the narrowest part of the lake at the west edge of the eastern sub-basin. Reference to the bathymetric chart (Fig. 1) shows how constricting this narrowing really is: a bottom ridge rises to within less than 15 m of the surface and extends to within 20 km of the south shore.

Fig. 7. NOAA-2 VHRR (visible), 10 February 1973, 15.00 Z. Compare with Figure 6. Now ice is forming along all shores. The wind is south-westerly in Cleveland but north-easterly in Buffalo inhibiting eastward movement of the ice.

Fig. 8. NOAA-2 VHRR (visible), 13 February 1973, 15.00 Z. Compare with Figure 7. Note the ice in Long Point Bay. Note also the clear area south of Long Point. This pattern has resulted from the easterly winds of the previous three days.

Fig. 9. NOAA-2 VHRR (visible), 17 February 1973, 15.00 Z. Computer-enhanced. Ice cover is nearly 100%; the dark areas are newly-formed ice that is transparent, thin, and not snow-covered. Note the ice accumulation or “dam” at Long Point.

Fig. 10. Landsat-l MSS-7, 17 February 1973, 16.30 Z. This image shows in exquisite detail the ice conditions on the lake.

On 18 February a well-publicized satellite image of Lake Erie ice was taken by Landsat-1. Cloud-free conditions permitted a glimpse of the onset of break-up. Southerly wind at 10–14 miles h^–1 (0.5–0.6 m s^–1) began the break-up despite a daily mean temperature of only —9.4°C at Cleveland. Three days of south-westerly winds followed. Figure 11 shows the effect of the southerly wind. Tension cracks have opened up all along the south shore and along the islands that stretch across the lake. The ice is moving north and east.

Fig. 11. Landsat-l MSS-7, 18 February 1973, 16.30 Z. Southerly winds have shifted the ice to the north. Note the sharp angular breaks.

On 23 February the lake was apparently ice-free in the area north of Long Point and the north-western part of the central sub-basin. During the 5 d cloud-covered period the winds had backed around through west to north-west. But temperatures continued to be below freezing during this period, so new ice formed. On 24 February the north shore was all that was visible as the north-westerly winds blew off-shore, but the next day and for three days (Fig. 12) the winds were easterly; thus Buffalo and the north shore were ice-free.

Fig. 12. NOAA-2 VHRR (visible), 1 March 1973, 15.00 Z. Easterly winds for three days have produced an anomalous break-up pattern. The ice has been blown to the west. Note the large floes.

By this time ice floes were large (10–30 km wide) and mobile. On 28 February, the lake ice was pushed farther to the west with the north-eastern shore of the central basin and the eastern half of the eastern basin ice-free.

A very warm strong southerly wind initiated widespread thawing as March began. The temperature did not fall below freezing again until 18 March. On 6 March it rose to a record 23.3°C at Cleveland. On 7 March the VHRR images show the ice piled up on the east-facing shores of the Canadian side of the lake and “rotten” and melting ice almost exclusively in the central sub-basin. On 8 March (Fig. 13) another Landsat-1 image revealed the same in more detail. Buffalo continued to be ice-free. Indeed the port of Buffalo opened without ice breaker assistance on 20 March, a full 22 d ahead of the average historical opening, and 25 d ahead of the forecasted opening. March water temperature for Lake Erie averaged a phenomenal 5.4 deg above normal. None the less it was primarily the easterly wind that caused the early port opening at Buffalo, not the high temperature of the water.

Fig. 13. Landsat-1 MSS-6, 8 March 1973, 16.30 Z. High temperatures and a southerly wind have reduced the pack to a rotten, lacy look. Note sastrugi and the low reflectance of the ice.

Concluding Remarks

Satellite observations of lake ice in the visible portion of the spectrum provide a useful method of monitoring ice formation and break-up. Over a period of years the empirical observations exhibit variations that can be analyzed in terms of meteorological, geomorphic, and thermal factors. This rudimentary study simply documents the results of a typical break-up as well as a very atypical one. Because of the satellite’s continuing, repetitive coverage, not only can day-to-day changes be recorded but they can be compared to concomitant environmental changes.

In 1973, unseasonable easterly winds produced ice-free conditions in eastern Lake Erie 25 d earlier than forecast. In retrospect, the forecast could have been up-dated as early as 1 March to indicate an earlier-than-usual opening. In the author’s view, the continuing accumulation of satellite data on lake ice will undoubtedly contribute significantly toward improved forecasting of ice break-up and formation. One additional satellite sensor needed is the day-night, all-weather capacity, such as side-looking airborne radar (SLAR) currently available only on aircraft, or the synthetic aperture radar (SAR) on Seasat-A.