International Lake Ontario-St. Lawrence River Study

About Us



Public Interest Advisory Group

Technical Work Groups

Reports and Minutes

Study Data


The Boardroom

International Joint Commission

Great Lakes Information Network
Web Site and Translation
by the Translation Bureau

Get Adobe Reader
Download Adobe Reader 7.0
Technical Working Groups

Predictability of Outflows

Performance Indicator Summary

Performance indicator: Predictability of Outflows

Technical Workgroup: Hydropower TWG

Research by: Hydropower TWG

Modeled by: To be modeled and tested by the PFEG

Activity represented by this indicator: This PI refers to the ability to look forward and predict whether outflows will be increasing or decreasing. Predictability of outflows allows for planned maintenance. It also contributes to the stability of the electricity markets.

Link to water levels: Water levels on Lake Ontario determine outflows. Lake Ontario rises and falls in a seasonal (predictable) pattern. A regulation plan that is premised on releasing water in relation to the water levels on Lake Ontario will be predictable.

Importance: This PI is important for hydropower, but it has importance for other interests as well. Maintenance at the power plants must be performed regularly. Often this requires that several units be removed from service simultaneously. For instance work on a transformer at Moses-Saunders will require a bank of units (4 unit tied together electrically) to be out of service. Power entities try to remove these units when outflows are lower and expect to remain lower. If flows increase unexpectedly during an outage then generation will be lost by running inefficiently at best and by spilling the water at worst. For Beauharnois, the best efficiency point is relatively low and each opportunity for unit maintenance must be used. The predictability is extremely important for proper scheduling.

Unexpected reductions in outflows will impact the price of electricity in the markets particularly if they occur during high demand periods such as a summer heat wave.

Performance Indicator Metrics: This is dependent on the flow routing techniques used in the plan. Each plan should be run to evaluate its reaction to a set of conditions. (Ie: How do different plans react to a sudden minor or major rain fall, at different periods of the year? Does one plan over react in comparison to the others in different situations?). This should be done by a detailed analysis of a limited number of plan options. The present global approach used by the PFEG does not provide for a good evaluation of this parameter. Primarily, a reaction (rate of flow increase or decrease) more rapid than 58DD, following a high inflow event, for example, should be considered as over-reaction and a bad performance.

This PI would likely be applied to the results of proposed plans to determine which plan responds in the most predictable manner. In a sense this would be a "sanity test". An example would apply to the high rainfall event following hurricane Frances last fall. For the week ending September 17, 2004 L. Ontario rose 9 cm (3.5 in) during a period when levels are usually declining. Plan flow reacted predictably by increasing flows however the increase was only 480 cms (17,000 cfs) If a proposed plan saw the same rise in one period and tried to eliminate the entire event immediately flows may have risen more than 900 cms (32,000 cfs) during the same period. This is assuming 300 cms (11,000 cfs) per week per 1 cm (0.4 in) on Lake Ontario. If the proposed plan assumed levels should be falling during this period then it may try to release even more than the 900 cms (32,000 cfs) mentioned above. The hydropower entities have listed this is an important PI but unpredictable flow fluctuations are disruptive to all entities. Following through with this example, flows were temporarily reduced by 400 cms (14,000 cfs) below plan flow one week later to allow boaters on Lake St. Lawrence to remove their boats.

Temporal Validity: Ongoing

Spatial Validity: n/a

Links with hydrology used to create the PI algorithm: Increasing levels on Lake Ontario normally predict increasing outflows and decreasing levels on Lake Ontario should be an indicator that outflows will decrease. Some events like ice control or flood control, which occur seasonally or frequently, are also predictable.

Algorithm: To be developed by the PFEG

Validation: Calculation of lost generation if flows suddenly increase during periods of unit maintenance.

Documentation and References: Hydropower TWG

Risk and uncertainty assessment: This PI attempts to reduce uncertainty of future outflows. If a regulation plan is selected that directs flows to fluctuate widely and in no predictable pattern then generation will be lost. This can be calculated. This uncertainty will cause unpredictable water level fluctuations in Lake St. Lawrence and downstream to Montreal. This will have negative impacts on property owners and recreational boaters.

Top of page