Bow River Basin - Water Supply Forecasts

Recorded March to September 2003 natural runoff volume in the Bow River basin was below-average to much-below-average, ranging between 72% and 93% of average, and ranking from 12th to 36th lowest on record (1912 to 1995 data) (Table 4). The monthly natural volume forecasts produced for the Bow River basin were within 6.8% of recorded values, on average, for the March through September runoff period.

Going into last winter, the southern half of the basin was wet but the northern half was dry. Winter precipitation was very low in all areas until large snowfalls occurred in March and April, however only the snowpack in the basin above Canmore was improved to normal or better depth. Snow throughout the basin melted very rapidly, producing more runoff than expected early in the summer. Much-below-normal summer precipitation reversed this wet trend very quickly, with July through September runoff volumes ranking among the ten lowest in 85 years of record. August through September was even drier, with record low or near-record low volumes measured during this period in all basins north of the Elbow River basin. The Elbow and Highwood River basins ranked near the tenth lowest on record for both periods. For the runoff year (March through September), volumes in 2003 were much higher than in 2001.

Bow River at Banff

The March to September 2003 recorded natural volume for the Bow River at Banff was much-below-average at 902,000 dam³ or 84% of average and ranks 13th lowest on record (based on the 1912-95 data) (Table 4).

Seven monthly (February to August) natural volume forecasts were produced for Banff for the March to September period. As illustrated in Figure 9, the recorded March to September 2003 natural volume at Banff (dark blue line) was above the forecasted values (brown triangles) for the first three months. The recorded volume was within the probable range for five of seven forecasts (all but July and August) and greater than the reasonable minimum for six out of seven forecasts (all except July).

Initial forecasts were low due to the lack of snow accumulation. The forecast increased substantially in May as a result of a major snowstorm in late April, resulting in an average snowpack. High amounts of runoff in June due to rapid snowmelt caused another increase in the forecast in July. However, the recorded volume ended up outside the forecast probable range during July and August and below the reasonable minimum during July as the July through September recorded volume was fourth lowest on record, and the August through September volume was second lowest, although both were similar to the volume recorded during the same time period in 2001. Summer precipitation was generally much-below-normal in the upper basin.

Forecasts ranged within 2.1% to 10.7% of the recorded volume, with February being the most accurate forecast and six of the forecasts within six percent of the recorded value. Monthly forecasts were, on average, within 4.8% of the recorded March to September volume.

Lake Minnewanka Inflow

The March to September 2003 recorded natural volume for the inflow to Lake Minnewanka was below-average at 175,000 dam³ or 93% of average and ranks 36th lowest in 85 years of record (1912-95 data) (Table 4).

Seven monthly (February to August) natural volume forecasts were produced for Lake Minnewanka for the March to September period. As illustrated in Figure 10, the forecasted values (brown triangles) were much lower than the recorded March to September 2003 natural volume (dark blue line) for the first four months of forecasts and higher for the last two months. The recorded volume was within the probable range for four of the seven forecasts (all but March, July and August) and was greater than the reasonable minimum for five of the forecasts (all but July and August).

Initial forecasts were low due to the lack of snow accumulation and dry antecedent conditions. The forecast increased significantly in May as a result of a major snowstorm in April greatly improving the snowpack. High snowmelt runoff volumes in May and June caused further increases in the forecast in June and July. The August forecast was lowered significantly as summer precipitation was much-below-normal in the basin. The July through September recorded volume was seventh lowest on record (1912-95 data) and August through September was even drier as third lowest on record, however both were much wetter than during those same periods in 2001. There was a very sharp decline in runoff this year, from the 20th highest volume in 85 years of record measured in the first half (March through May), to the seventh lowest on record in the second half (July through September).

Forecasts ranged within 1.6% to 26.2% of the recorded volumes, with June's being the most accurate and the only one of the forecasts within five percent of the recorded value. The forecasts were on average, within 15.4% of the recorded March to September volume.

Spray River near Banff

The March to September 2003 recorded natural volume for the Spray River near Banff was much-below-average at 308,000 dam³ or 84% of average and ranks 17th lowest on record (based on the 1912-95 data) (Table 4).

Seven monthly (February to August) natural volume forecasts were produced for the Spray River for the March to September period. As illustrated in Figure 11, the forecasted values (brown triangles) did not deviate much from the recorded March to September 2003 natural volume for the Spray River (dark blue line). The recorded volume was within the forecasted probable range and greater than the reasonable minimum for all forecasts but August's.

Initial forecasts were low due to the lack of snow accumulation. The forecast increased in April as a result of above-normal to much-above-normal snowfall in March, however the snowpack remained below-average. Rapid snowmelt in May and June caused another increase in the forecast in July. Summer precipitation was much-below-normal in the basin and as a result the July through September recorded volume was similar to the volume recorded during the same time period in 2001. However, water diverted to the Spray Reservoir has not been accounted for, as there was no gauge available this year, and so the actual natural volume to the Spray Reservoir would be slightly lower.

Forecasts ranged within 0.3% to 9.0% of the recorded volume, with May's being the most accurate and six of the seven forecasts within six percent of the recorded value. The forecasts were on average, within 4.3% of the recorded March to September volume.

Kananaskis River

The March to September 2003 recorded natural volume for the Kananaskis River was much-below-average at 330,000 dam³ or 80% of average and ranks 12th lowest on record (based on the 1912-95 data) (Table 4).

Seven monthly (February to August) natural volume forecasts were produced for the Kananaskis River for the March to September period. Forecasts generally overestimated volume slightly, which is evident in Figure 12, where the recorded March to September 2003 natural volume for the Kananaskis River (dark blue line) was below the forecasted values (brown triangles). All seven forecasts were within the forecasted probable range and greater than the reasonable minimum.

Initial forecasts were low due to the lack of snow accumulation. The forecast increased in April as a result of above-normal to much-above-normal snowfall in March, however the snowpack remained below-average. Rapid snowmelt in May and June caused another slight increase in the forecast in July. The forecast was lowered significantly in August as summer precipitation was much-below-normal in the basin. July through September recorded volume was second lowest on record (based on the 1912-95 data), and the August through September volume was the lowest, although both were similar to the volume recorded during the same time period in 2001. However, water diverted to the Spray Reservoir has not been measured, as there was no gauge available this year, and so the actual volume in the Kananaskis River is probably slightly higher.

Forecasts ranged within 1.1% to 10.8% of the recorded volume, with March's being the most accurate and four of the forecasts within five percent of the recorded values. Monthly forecasts were, on average, within 4.7% of the recorded March to September volume.

Bow River at Calgary

The March to September 2003 recorded natural volume for the Bow River at Calgary was much-below-average at 1,987,000 dam³ or 81% of average and ranks 17th lowest on record (based on the 1912-95 data) (Table 4).

Seven monthly (February to August) natural volume forecasts were produced for the Bow River at Calgary for the March to September period. As illustrated in Figure 13, the forecasted values (brown triangles) did not deviate much from the recorded March to September 2002 natural volume at Calgary (dark blue line) for the first three months (February-April), but then were higher for the last four months (May to August). The recorded volume was within the probable range for six of seven forecasts (all but July) and was greater than the reasonable minimum for all seven forecasts.

Initial forecasts were low due to the lack of snow accumulation. The forecast increased in May as a result of above-normal to much-above-normal snowfall in March and April, however the snowpack remained below-average. Rapid snowmelt in May and June caused another slight increase in the forecast in July. The forecast was lowered significantly in August as summer precipitation was generally much-below-normal in the basin. July's lower quartile forecast accurately predicted the recorded volume. July through September recorded volume was fourth lowest on record (based on the 1912-95 data), and the August through September volume was second lowest, however both were slightly higher than the volume recorded during the same time period in 2001.

Forecasts ranged within 1.0% to 12.9% of the recorded volumes, with April's being the most accurate and five of the seven forecasts within six percent of the recorded value. The forecasts were on average, within 5.2% of the recorded March to September volume.

Elbow River

The March to September 2003 recorded natural volume for the Elbow River was below-average at 163,000 dam³ or 74% of average and ranks 24th lowest on record (based on the 1912-95 data) (Table 4).

Seven monthly (February to August) natural volume forecasts were produced for the Elbow River for the March to September period. As illustrated in Figure 14, the recorded March to September 2003 natural volume for the Elbow River (dark blue line) was above the forecasted values (brown triangles) for the first two forecasts (February and March) and below the last three forecasts (June - August). The recorded volume was within the probable range for five of the seven forecasts (all but July and August) and was greater than the reasonable minimum for six of seven forecasts (all but August).

Initial forecasts were low due to the lack of snow accumulation, despite wet antecedent conditions entering last winter. The forecast increased substantially in April as a result of above-normal to much-above-normal snowfall in February and March, however the snowpack remained below-average. High amounts of runoff in May and June due to rapid snowmelt caused further increases in the forecast in June and July. Summer precipitation was generally much-below-normal in the basin, and so July through September recorded volume was tenth lowest on record, slightly higher than the volume recorded during the same time period in 2001.

Forecasts ranged within 1.4% to 11.2% of the recorded volume, with April's being the most accurate and five of the seven forecasts within six percent of the recorded value. The forecasts were on average, within 5.4% of the recorded March to September volume.

Highwood River

The March to September 2003 recorded natural volume for the Highwood River was below-average at 458,000 dam³ or 72% of average and ranks 27th lowest on record (based on the 1912-95 data) (Table 4).

Seven monthly (February to August) natural volume forecasts were produced for the Highwood River for the March to September period. As illustrated in Figure 15, the recorded March to September 2003 natural volume for the Highwood River (dark blue line) was below the forecasted values (brown triangles) for the first four (February-May) and the final forecasts. The recorded volume was within the probable range and greater than the reasonable minimum for all seven forecasts.

Initial forecasts were low due to the lack of snow accumulation, despite wet antecedent conditions entering last winter. The forecast increased substantially in April as a result of above-normal to much-above-normal snowfall in February and March, however the snowpack remained below-average. High amounts of runoff in May due to rapid snowmelt caused another increase in the forecast in June. Summer precipitation was generally much-below-normal in the basin, and so July through September recorded volume was eighth lowest on record, much higher than the volume recorded during the same time period in 2001.

Forecasts ranged within 0.2% to 19.0% of the recorded volume, with July being the best forecast and four of the forecasts within six percent of the recorded value. The forecasts were on average, within 7.8% of the recorded March to September volume.

A correction of volume data resulted in 40,000 dam³ more runoff this year, most of which missing due to gauge problems in March. Had the volumes been accurate during our forecasts, forecasts would have been within 5.8%, on average, of the recorded March to September volume, with earlier forecasts becoming more accurate. Note that the gauge data from March has yet to be finalized, and may again change significantly.