• Introduction
• Approach
• Results
• Discussion

Introduction

The LMPDS Study Team has contracted the Great Lakes Environmental Research Laboratory of NOAA (GLERL) to generate a series of alternate hydrologic scenarios based upon water supply sequences for each of the Great Lakes which plausibly represent a wider range of conditions over the near future. These scenarios include monthly mean water levels for each of the Great Lakes over the 50-year planning horizon of the LMPDS. Ultimately, potential damages occurring under each scenario will be evaluated.

The number of alternate hydrologic scenarios generated will strongly influence the level of effort needed to conduct economic impact analyses. The LMPDS Study Team identified five critical alternative hydrologic scenarios to be generated. These are:

• Existing recorded ranges on each lake, with similar frequencies of recorded extremes (Historic Range);
• Existing recorded ranges on each lake, with increased frequencies of extremes (Increased Frequency);
• Increased extremes in high water levels on each of the lakes, with an increase in the frequencies of high water levels on each (Extreme High);
• Increased extremes in low water levels on each of the lakes, with an increase in the frequencies of low water levels on each (Extreme Low); and
• Increased extremes in both high and low water levels on each of the lakes, with an increase in frequency of extremes for both high and low water conditions (Increased Range).

Approach

The various scenarios were selected by sliding a 50-year window through a stochastically derived data set consisting of a time series of 50,000 years of lake level data. The data were simulated by routing a time series of 50,000 years net basin supply data (Rassam et al., 1992) through a hydrologic response model of the Great Lakes system, which includes the modified regulation plans for Lakes Superior and Ontario (Lee et al, 1994). For each of the five scenarios, several 50-year blocks of data were identified that met the criteria of the proposed scenarios. Percentiles were calculated from the lake levels for the years of record for a selected site and the selected 50-year blocks. The selected site for Lakes Michigan-Huron was Harbor Beach, Michigan. The percentiles for the selected 50-year blocks were then plotted with the percentiles for the years of record. A representative 50-year block was chosen for each of the scenarios, based on a visual assessment of the plots that met the various criteria.

Results

The figure below shows the historical record for Lake Michigan-Huron with the extreme highs and lows identified. The figure also shows the plausible high and low water based on the 1% value of 50,000 years of simulated data.

The three figures below show the exceedance probability plots for the existing period of record and the five simulated scenarios. Two scenarios-the increased frequency of wet years and the increased frequency of dry years-define the High and the Low scenarios. To quantify a plausible still water level for the High scenario, the 1% exceedance elevation was calculated. The 1% exceedance elevation (583.41 feet, IGLD 1985) represents the water level that would be exceeded only 1% of the time under a scenario of increased frequency of wet years. Similarly, the plausible still water level for the Low scenario was determined as the 99% exceedance elevation. The 99% exceedance elevation (574.31 feet, IGLD 1985) represents the water level that would be exceeded 99% of the time under a scenario of increased frequency of dry years.











Although the still water analysis characterizes the plausible expected water levels under still water conditions, significant water level rises and drawdowns also occur as a result of storm conditions on the lake. The USACE regularly analyzes the storm rise probabilities on the Great Lakes and maintains a database of the monthly maximum storm rises. The storm rise values were calculated as the difference between the daily maximum water level and the monthly mean water level. Similarly, the storm drawdown values were calculated as the difference between the daily minimum water level and the monthly mean water level. These data were analyzed over the period of record to determine the 1% exceedance storm rise and the 99% exceedance drawdown water level difference. The 1% exceedance value was calculated as 1.07 feet at Holland, Michigan and 1.9 feet at Sheboygan, Wisconsin. The 99% exceedance (drawdown) value was calculated as 1.2 feet at Holland and 1.6 feet at Sheboygan.

Combining the High stillwater scenario elevation with the with the 1% exceedance storm rise produces what was defined as the Extreme High Scenario. Likewise, the Low stillwater scenario combined with the 99% exceedance storm drawdown elevation produces what was defined as the Extreme Low Scenario. These values are presented in Table 1.

Discussion

The 50-year blocks, representing the extremes for the non-regulated lakes, Lakes Michigan-Huron and Lake Erie, display a significant divergence from the data for the years of record. In addition, the 50-year blocks chosen as matching the years of record did reasonably well matching the years of record. It is understood that the levels reported for the Extreme High and Extreme Low scenarios are extremes, given the compound probabilities that were used to derive them. However, a comparison of the scenarios with data presented at a recent workshop on paleo-lake levels (Sellinger and Quinn, 1999) demonstrate that these lake levels represent plausible, future extreme conditions that could occur over the next 50 years. The Extreme High and Extreme Low lake levels will be used in this study to define the upper and lower bounds of plausible extreme lake levels with which to assess potential damages.

Download the Full GLERL Consulting Report on this topic from The LMPDS Document Clearinghouse

Click to return to the Lake Michigan Potential DamagesHome Page.