MODFLOW MODELING FOR A PROPOSED ARTIFICIAL SALMON SPAWNING CHANNEL EXTENSION NEAR HYDER, ALASKA

Marx Creek is an artificial salmon spawning channel located approximately 4-1/2 miles north of Hyder in southeastern Alaska. Marx Creek was constructed in the late 1980s and was so successful that it was extended in the early 1990s. However, the extension was not nearly as successful as the initial channel because the extension was constructed immediately east of the flood-control dike separating Marx Creek from the glacially fed Salmon River, unlike the initial channel which had been constructed approximately 500 feet east of the dike. This allowed silt from the Salmon River to infiltrate into the extension through the dike. Consequently, a second extension has been proposed parallel to the first but further east of the dike. To this end, a MODLFOW model has been created to determine whether the ground water conditions will support a second channel extension, and how the construction of a second channel will affect flow in the first one.

Twenty monitor wells were installed at the site of the proposed channel extension during the summer of 2006. Slug and pumping tests were performed to measure the hydraulic conductivity at each well location, with values ranging from 0.5 to 1,500 feet per day. Water-levels were recorded twice daily in each well using Solinst Leveloggers for a period of 117 days (July to November 2006). Water-level fluctuations varied from 1.46 to 6.25 feet, and averaged 3.22 feet during this period. Discharge measurements were also made at each of 16 weirs along the upper portion of Marx Creek. The discharge at Wier 1 at the headwaters of the creek was 31,086 cubic feet per day, and the discharge at Weir 16 was 972,000 cubic feet per day.

These data have been used to create a MODFLOW ground-water flow model. The model was calibrated to the water-level data for the 117-day period from July to November 2006. The maximum difference between the measured and model-calculated heads was 4.32 feet, and the average difference was only 2.57 feet. After calibration the proposed channel extension was added to the model and it was again run for 117 days, at which time steady-state conditions had been reached. According to the model results, the discharge of the proposed extension should be approximately 29% greater than the discharge in the existing channel, and the extension should not significantly affect the discharge of the existing channel.