Appendix 3

ALBERTA ENVIRONMENT
FISHERIES AND OCEANS CANADA
WATER MANAGEMENT FRAMEWORK:
INSTREAM FLOW NEEDS AND WATER MANAGEMENT SYSTEM
FOR THE LOWER ATHABASCA RIVER

---

The Water Management Framework

The purpose of the Framework is to protect the aquatic ecosystem of the lower Athabasca River while allowing development to occur. For Alberta, the maintenance of healthy sustainable aquatic ecosystems is enshrined in the Alberta Environmental Protection and Enhancement Act (EPEA) and the Alberta Water Act. Fisheries and Oceans Canada policy aims for no net loss of fish habitat on a project by project basis.

The Framework contains two components:

1. An IFN recommendation designed to provide protection of the aquatic ecosystem,
2. A water management system that allows withdrawals below the IFN recommendation but defines thresholds that trigger management responses. It also provides a path forward for future monitoring and research and a process to develop an adaptive management system based on that information.

Determining the IFN Recommendation

A comprehensive IFN examines the impacts that future water withdrawals and discharges will have on water quality, fish habitat, river geomorphology and riparian vegetation. The IFN is meant to provide guidance for the near future and is based on the historical flow record. A re-evaluation of the water management system threshold values will be required if the current flow record is no longer representative of future flows. Table A1 identifies the components and relevant studies on which the IFN is based.

Given the low total projected withdrawals, impacts are most likely to occur during low flow conditions and thus only pose significant threat to the water quality and biology (fisheries) components. Although some preliminary work was done on river geomorphology and riparian vegetation, it was not sufficient to make IFN recommendations. It was acknowledged that at the scale of withdrawals for the oil sands industry, river geomorphology and riparian vegetation is very unlikely to be affected. Current water quality modeling using extreme drought conditions (flows as low as 50 m³/s were modelled, minimum recorded flow from 1957-2004 is 75 m³/s) indicates that water quality is not likely to be a limiting factor because:

1. Release (discharge) from reclaimed areas and on-site storage of process affected water will not become significant until the future when most withdrawals will have subsided and,
2. The quality of future releases will meet guidelines for the protection of aquatic life.

Therefore, to date the IFN has been based on fish habitat results from reaches 2 to 5 during open water and reaches 4 and 5 during ice cover. Riparian vegetation requirements were considered when setting the cautionary threshold (CT) during the freshet by using the SSRB (South Saskatchewan River Basin) riparian minimum flow recommendation. It is expected that the CEMA work will continue and will address all components and all reaches of the lower Athabasca River.

Table A1. Studies providing the basis for establishment of the IFN.

Biology (Fish Habitat)

Methods for directly determining the impact of reduced water availability on the aquatic ecosystem are not available for the lower Athabasca River and to our knowledge are rare in the international scientific literature. Typical of most IFN studies, modelled fish habitat area versus flow relationships were used for the lower Athabasca River. While the use of fish habitat is the current best practice, future IFN development should include:

• Other ecosystem indicators such as riparian vegetation or benthic invertebrates if these are determined to be sensitive to changes in flow.
• The fish habitat modelling considered only depth and velocity. Other fish habitat characteristics of importance like substrate, particularly for spawning life stages, need to be investigated.
• The winter under-ice modelling CEMA and its partners developed for the lower Athabasca River was a first for IFN studies. There is interest in refining the models to more realistically simulate under ice and ice formation conditions

There is an Instream Flow Needs Technical Task Group (IFNTTG) within the CEMA umbrella that is addressing the future monitoring requirements and research needs. This work will be a key component for adaptation to the Phase 2 IFN.

Natural Flow Regime

The prevailing scientific literature on instream flow needs indicates that a natural flow regime is the best way to protect the aquatic ecosystem (Poff et al. 1997). Preserving seasonal cycles is important, as a river and the ecosystem it supports is a product of the energetic forces dissipated over the natural range of variability. The life histories of fish and other aquatic species are adapted to high and low flows at appropriate times of year. For example, high flows are required in spring for fish that spawn at this time while fish that spawn in the fall do best with typical low flows in fall.

Preserving year-to-year fluctuations is important for maintaining the diversity of species in the aquatic ecosystem. Some species do best in dry years while others do best in wet years. To maintain these species in their natural proportions, fluctuations in flow from year to year need to be preserved.

The natural flow regime is reproduced by using a percent withdrawal approach. As long as the percent withdrawal is not too severe, the IFN will mimic the natural hydrograph, providing wet and dry seasons and wet and dry years at the appropriate time.

At the same time, it is recognized that the potential impact from water withdrawal is greatest at low flows. The allowable percent withdrawal is therefore reduced at low flows to accommodate this increased ecosystem sensitivity. This is accomplished through the use of an ecosystem base flow (EBF) or cautionary threshold (CT). An EBF is a flow below which no withdrawals are recommended. It is based on the premise that at low flows, the aquatic ecosystem is more sensitive to water withdrawals.

Habitat Metrics

To evaluate the risk to fish habitat, a series of habitat metrics have been developed in Alberta to help define an IFN prescription. The method compares the change in suitability-weighted habitat area (weighted usable area or WUA) using the natural flow record compared to a proposed withdrawal scenario. To define the IFN, the maximum of the:

1. flow corresponding to an 80% habitat exceedence (HDA80 flow), or
2. the Q90 (to protect riparian vegetation following the SSRB recommendation)

was used for the cautionary threshold (CT) for each week of the year. The IFN was set at the largest percent reduction in flow that does not exceed any of the metrics assuming no water withdrawal below the CT. This approach and metrics were developed for the South Saskatchewan River Basin (SSRB). The approach has been generally adopted across Alberta but is referred to as the SSRB method. The metrics of the SSRB method are:

• Mean Loss – A 10% reduction in average habitat from natural habitat for the most sensitive fish life stage for the entire simulation period. This metric requires that high flows be stripped from the simulation period (simulation period is 1957-2004 for the lower Athabasca River). This metric checks overall or chronic impacts from water withdrawal.
• Maximum Weekly Loss – A 15% reduction in average habitat from natural habitat for any week of the year. The metric is calculated for the average change in simulated habitat for each week of the year for all years (1957-2004 flow record for each of week 1, 2, 3, etc.). It is a check to determine if certain seasons or weeks are being unduly impacted by the withdrawal.
• Maximum Instantaneous Loss – A 25% reduction in habitat from natural habitat for any week of any year. This metric independently evaluates each week of the period of record to look for any habitat bottlenecks. Every week of every year in the simulated withdrawal is compared to natural for the corresponding week (e.g., week 43, 1964). This metric is a check on habitat bottlenecks or acute habitat impacts from water withdrawal.

The premise for stripping high flows is that the model is not properly simulating fish habitat at these very high flows. All of the fish habitat curves peak at some value and then decline at higher flows. For example, Goldeye Adults in Reach 5 peak at 1500 m³/s. Reducing flow from 2000 m³/s to 1800 m³/s mathematically results in a habitat gain. The model boundaries are restricted by surveying methods and at high flows the margins of the river are not realistically simulating velocity and depth. In addition, the model does not address the moderating influence of floodplain vegetation on fish habitat at high flows.

For the Athabasca River, the method of stripping was modified slightly from the SSRB approach. All flows higher than the peak of the habitat-flow curve for each life stage were stripped from the period of record. The SSRB approach stripped all flows (high and low) in weeks where the median exceeded the peak of the habitat-flow curve.

Unlike the SSRB approach, the Phase 1 Framework allows withdrawals in the 80 to 100% exceedence range. Therefore, it is important to determine what habitat impacts are occurring during those more sensitive periods. The three SSRB metrics were applied to the 80 to 100% exceedence habitat range. However, high flows were not stripped from the period of record when calculating the 80 to 100% mean loss. The results for the overall and 80 to 100% exceedence habitat metrics are presented in Table 2 of the main body of the Framework.

The weight of opinion from international work supports between 10 to 20% reductions in flow at the 80% to 90% flow exceedence as the basis for assessing shifts to high-risk conditions (e.g., Brizaga and Arlington 2001, Tharme and King 1998, Clipperton et al. 2003). The same studies consistently indicate that there is a low flow below which withdrawals should not occur similar to the EBF concept in the SSRB approach. Again, each of these studies are consistent in suggesting that between the 80% and 95% flow exceedence there exists a range of potential impacts from measurable but recoverable to critical and unsustainable.

AENV, SRD (Alberta Sustainable Resource Development) and DFO jointly reviewed the habitat simulation data, the international contributions to IFN determination, and performed additional threshold assessment and statistical comparisons of habitat loss. These tests indicated increased risk of impacts to aquatic ecosystems begin to occur at the 80% habitat exceedence flow (HDA80 flow).

Summary of the IFN Threshold Values - Phase 1

The results of the habitat simulation indicated a 15% withdrawal is allowable until flow is lowered to values corresponding to the Cautionary Threshold (CT). The definition of the CT is the higher flow value of either the 80% habitat exceedence (HDA80) or the 90% flow exceedence (Q90). The 90% flow exceedence was adopted from the South Saskatchewan River Basin recommendation that this was a minimum required for maintenance of riparian vegetation. The HDA80 was calculated for each week and for each species and lifestage for which habitat suitability curves were available. The highest value, representing the most sensitive species or lifestage, was chosen for each week. For Reach 4, adult longnose sucker were most sensitive and therefore defined the HDA80 in most weeks except later summer weeks, which were determined by adult goldeye.

Below the CT there exists a critical flow threshold termed the Potential Sustainability Threshold (PST). The PST was defined as the 95% flow exceedence. An example of the derived management zones in relation to typical flows in the lower Athabasca River is presented in Fig. A1.

Figure A1: Lower Athabasca River IFN indicating flows for the mean year (1958-2004)

Note: The intent of this figure is to graphically display the values presented in Table 2 of the main body of the Framework. In the green zone, a maximum of 15% of the instantaneous flow would be available for withdrawal, not the difference between instantaneous flow and the green-yellow line (the CT).

The Cautionary Threshold - Yellow zone

The yellow zone is where impacts may begin to appear, providing the green zone restrictions are met. It is believed that water withdrawal impacts in this zone, if they occur, would be short-lived. When flows are within the yellow zone, maximum cumulative withdrawals should not exceed 10% of instantaneous river flow. Given the difficulty in implementing instantaneous restrictions the method was modified to use 10% of the mid-point between the green-yellow and yellow-red boundaries. The difference of this approach from an instantaneous approach during winter weeks is within rounding errors. The 10% reduction is consistent with international scientific literature where effects are expected when hydrologic properties are altered by more than 10% between the 80% and 90% flow exceedence (Brizga and Arthington 2001). The 10% reduction is further restricted by maximum withdrawal limits. Maximum cumulative withdrawals are limited to 15 m³/s during winter ice-covered conditions and 5% of the HDA80 flow or 34 m³/s, whichever is less during spawning and 34 m³/s during the remainder of the ice-free time period.

Winter weeks are weeks 1 through 15 (Jan 1 to early April) and weeks 44 through 52 (late Oct. to Jan.), spawning occurs during weeks 16 through 24 (April through early June) and the remaining open water weeks (summer) are weeks 25 through 43.

Figure A2

The Potential Sustainability Threshold - Red Zone

The yellow - red demarcation is termed the Potential Sustainability Threshold (PST) and was defined as the 95% exceedence flow. This is a zone where withdrawal impacts are potentially significant and long-term, depending on duration and frequency withdrawals. Consequently, withdrawals have to be carefully managed in this zone. Two analyses were used to determine the total cumulative withdrawal allowed in the red zone:

1. A confidence interval approach for determining limits of significant change to ensure the severity of existing low flow conditions are not increased significantly.
2. An assessment of change in frequency and duration of red and yellow conditions to ensure the length of exposure to low flow does not increase significantly.

To assess the significance of increasing the severity of low flow conditions, the historic range of weekly mean flows was used to construct 90% confidence intervals (CI90) for the habitat available to the most sensitive species during winter (Longnose Sucker - LNSC). The range in habitat enclosed in the CI90 was converted back to the corresponding range in flow. The range in flows was consistently between +/- 6.6 and 7.7 m³/s for winter weeks 1 through 13 and up to 25 m³/s when all winter weeks are included. For ease of implementation the one-sided confidence interval flow for each week was converted to a percentage of the median weekly flow. This ranged between 4% and 8% not including the most extreme variable spring weeks and averaged 5.23% of median weekly flow across all winter weeks, excluding weeks 44 and 45. Weeks 44 and 45 could not be included because flows occurring during these two weeks are outside the range (too high) of the winter habitat curves used to predict habitat loss. Not including these two weeks results in a more protective water management framework because the extreme variation in these weeks would increase the CI interval. A withdrawal limit within the PST of 5.2% of median weekly flow was considered to be within a reasonable level of statistical detectability. This approach effectively employs statistical significance as a proxy for biological significance. In addition to the 5.2% withdrawal limit within the PST, AENV and DFO propose an upper limit of 15 m³/s ice-covered conditions and 5% of the HDA80 flow or 34 m³/s, whichever is less during spawning and summer ice-free time periods.

To validate this approach, AENV and DFO considered changes in frequency and duration of yellow and red conditions under incremental reductions in median weekly flow of 1 to 15 m³/s. A flow reduction of 5.2% of median weekly flows to a maximum of 15 m³/s results in a maximum increase of red zone frequency and duration less than 15%, and a 4% mean decline in habitat availability for combined yellow and red conditions.

Climate Change

The development of the IFN and water management system is based on available information on historical flows in the Athabasca River. If flows were to generally decrease due to climate change, restrictions of the Framework would be invoked more frequently. If flows were to generally increase, there would be fewer instances that restrictions on withdrawals would be necessary.