Water Supply Forecast Summary

Going into last winter, wet soil conditions existed in the southern and central mountains and foothills. Winter snow accumulations varied from below to much below average in the upper Highwood, Red Deer and North Saskatchewan River basins, and below average to average in the Oldman River basin and remainder of the Bow River basin. Winter precipitation was below normal to normal in the Milk River basin. Initial natural runoff volume forecasts ranged from below average to average.

Precipitation was below normal in most forecast basins in spring and early summer, which led to decreases in forecasts in many areas over that time. Inflows to the Bighorn and Spray Reservoirs however were higher than expected probably since there was more snow in the mountains of these basins than measurements indicated, and so forecasts rose for these locations. Heavy precipitation in early June in the Bow and Oldman River basins led to forecast increases for several locations. Natural runoff volumes for the March-June period were generally average in most of these two basins, and below average elsewhere.

From mid-June through July though, precipitation was much below normal in all forecast basins. July natural runoff volumes were among the 25 lowest on record, most notably 2nd lowest at Brazeau, 3rd at Lake Minnewanka, 4th at Edmonton, and 5th at Banff. Final forecasts for the year (on August 1) dropped significantly. August was again very dry in most areas, with August natural runoff volumes being among the fifteen lowest on record. Record lows for August volumes were set at Banff, Spray Lakes, Calgary, and the Belly River, while Edmonton was second lowest.

September precipitation totals were much improved; below normal to normal in the Milk and Oldman River basins, above normal in much of the Bow and Red Deer River basins, and much above normal in the North Saskatchewan River basin.

Total recorded natural runoff volumes for the March to September 2006 period in the province were generally much below average in the North Saskatchewan River basin, below to much below average in the Bow and Red Deer River basins, below average in the Milk River basin, and below average to average in the Oldman River basin. Brazeau reservoir inflows were 2nd lowest on record, and natural volumes were 12th lowest at Edmonton and 13th lowest at Banff. Most other areas ranked between 20th and 40th lowest in up to 91 years of record (Table 1). March-September 2006 volumes were lower or much lower than during the same period last year in most areas due to flooding last year, but were higher this year in the Waterton, Belly, St.Mary, and Milk River basins as flooding was not as prevalent last year in these areas.

Currently, soil moisture is average and flows range from below average to above average, as we head into winter.

On average, monthly forecasts were very accurate in the Red Deer and Oldman River basins. In the Red Deer River basin, fourteen forecasts produced during the year were on average within 5.8% of the recorded March-September totals, while in the Oldman River basin, a total of 35 forecasts were on average within 5.0%. Twenty-one forecasts in the North Saskatchewan River basin were within 10.4%, on average, of the recorded March to September volumes. Twenty-four forecasts for the Milk River basin were on average within 13.9% of the recorded March to September volumes, and 49 forecasts in the Bow River basin were within 8.9% on average. Monthly forecasts assumed normal precipitation for the remainder of the forecast season and so the difference between the forecasts and the recorded values is often attributed to the fact actual precipitation was not always normal. As a result of minimal precipitation during July and August, the recorded natural runoff volume tended closer to the probable lower range forecast than the probable forecast value, in several areas, especially those which switched rapidly from high runoff conditions before July to low volumes after July. Corrections to flow data by Water Survey of Canada can also affect forecast accuracy.