Groundwater Movement

The direction of horizontal groundwater movement can be inferred from maps of water-level altitude contours. Groundwater flow generally is from areas of recharge to areas of discharge, in the direction of decreasing water-level altitudes and perpendicular to the water-level altitude contours. Water-level altitude maps and regression analysis were used to determine shallow groundwater-flow directions and gradients in the Skagit River Delta area.

Groundwater-Flow Directions

Groundwater flow in the alluvial and recessional outwash aquifer generally moves in a southwestward direction away from the Skagit River and towards the Swinomish Channel and Skagit Bay (figs. 5–8). Local groundwater flow toward the river was inferred during February 2008 in an area west of Mount Vernon near well 24K01, and in an area southwest of Mount Vernon near wells 35R01 and 35R02. Water-level altitude contours in these areas indicate a south to south eastward groundwater-flow direction (fig. 7). Contours in these areas do not indicate groundwater flow to the river during the other quarterly measurement periods, and flow to the river in these areas likely is seasonal or transient in nature. Water-level altitudes in well 35R02 also were greater than nearby estimates of river stage during November 2007 and May 2008 (figs. 6 and 8; however, water-level altitude contours near this well indicate a westward groundwater-flow direction away from the river. Water-level altitudes varied seasonally, however, generally ranged from 1 to 2 ft (August 2007, fig. 7) in the west to about 15 ft (May 2008, fig. 8) in the east. Concave and convex bends in the water-level altitude contours indicate the presence of convergent and divergent groundwater-flow patterns on a local scale. These flow patterns likely reflect variations in hydraulic conditions within the aquifer resulting from spatial variations in aquifer material properties (clay, silt, and sand). Local groundwater withdrawals and agricultural drainage systems also may influence the groundwater-flow patterns.

The fitted regression coefficients and calculated groundwater gradient, flow direction, and water level at the centroid are presented in table 2. An estimate of goodness of fit (statistical parameter R²) for the regression also is given for each seasonal estimate. The high R² values indicate that the planar representations of the groundwater potentiometric surfaces are in good agreement with the quarterly water-level measurements. The large negative a₀ (intercept) coefficients are due to the origin of the State Plane System being far to the southwest and well away from the project area.

Contours of equal values of z (time-averaged groundwater altitude for the four quarterly measurements) were calculated algebraically from the annual a₀, a₁, and a₂ coefficients (fig. 9). Estimates of the overall (time averaged) groundwater-flow direction (8.5° south of west) and gradient (2.67 ft/mi) also were computed (table 2).

The overall groundwater-flow direction derived from the regression analysis is similar to those depicted on the water-level maps; shallow groundwater flows slightly south of west from the mainstem of the Skagit River out to the Swinomish Channel and Skagit Bay. Regression analysis of the four quarterly measurements generally agrees with the annual analysis, with only slight variations in gradient or direction (table 2).

Time-averaged water levels in individual wells are always higher or lower than the fitted planar surface (fig. 9). Where measured water levels in a subarea are higher on average than the fitted plane, the potentiometric surface is diverging from the uniform planar flow direction; conversely where water levels in a subarea are lower than the plane the flow is converging. Differences in measured water levels and the plane indicate a pattern in which groundwater converges from the river meander, diverges into two preferential channel areas, converges to flow between the till outcrop uplands, and finally diverges to flow separately toward Skagit and Padilla Bays. Taking into account this final divergence, the sea level (z = 0) contour would coincide fairly well with sea level in the bays and the Swinomish Channel.

To compare the overall slope of the ground surface in the area with that of the groundwater surface, a similar regression calculation was performed on ground-surface elevation at the quarterly monitoring well locations used in the construction of the water-level maps. Ground surface slopes were determined to be slightly steeper (3.82 ft/mi) and slightly more to the southwest (W 24.0° S) than the groundwater potentiometric surface.

The potential for vertical groundwater movement within the alluvial and recessional outwash aquifer was evaluated in areas where a comparison of water-level altitudes could be made in closely spaced wells completed above and below clay layers within the aquifer (figs. 5–8). Wells completed below clay layers are widely spaced and along the margins of the study area, however, comparisons were possible at three locations (19F01/19F02, 09L01/09L02, and 10A01/10B01). Water-level altitude differences at the first two well-pair locations indicate the potential for year-round downward flow, with higher altitudes in wells completed above clay layers for all quarterly measurements. A more complex relation was observed at the last well-pair location in which higher altitudes in wells completed above clay layers were observed during August 2007 and May 2008 measurements indicating the potential for downward flow, and lower altitudes in wells completed above clay layers during the November 2007 and February 2008 measurements indicating the potential for upward groundwater flow.

Groundwater-Level Fluctuations

Groundwater levels fluctuate over time, both seasonally and long term, in response to changing rates of recharge to and discharge from the groundwater system. When recharge exceeds discharge, the amount of water stored in an aquifer increases and water levels rise; when discharge exceeds recharge, groundwater storage decreases and water levels decline. Groundwater levels also may respond to changes in nearby stream stage. When stream stage exceeds nearby groundwater levels, streamflow may recharge the aquifer, causing a rise in groundwater levels; when groundwater levels exceed nearby stream stage, discharge from the aquifer to the stream may occur, resulting in a decline in groundwater levels.

Seasonal changes in groundwater levels in many of the wells in the Skagit River Delta follow a typical pattern for shallow wells in western Washington. Water levels rise in autumn and winter, when precipitation is high, and decline during spring and summer, when precipitation is low (fig. 10). Groundwater levels in wells along the eastern margin of the study area also are likely influenced by stage of the Skagit River (fig. 11). Water levels in these wells remained elevated through April, likely due to groundwater recharge from the river, and did not seem to begin to decline until the end of May in response to declining river stage (fig. 12).

Results of the regression analysis of quarterly groundwater levels, indicated by the calculated water level at the centroid (table 2), are in general agreement with the typical pattern for shallow wells in western Washington. Regression analysis also indicates the groundwater gradient was steepest in May, likely due to elevated groundwater levels associated with Skagit River stage, and that by August the water levels along the river declined in response to declining river stage, and the gradient was reduced (table 2). The groundwater gradient increases through the winter probably in response to groundwater recharge from precipitation.

Groundwater levels also may respond to changes in nearby ocean tides. Cyclical water-level fluctuations in a well equipped with a continuous water level recorder exhibited a periodicity that is characteristic of ocean tides (for example, component wave lengths of about 24 hours, and 12 hours and 25 minutes). The well (34N/03E-19F01) is north of the town of La Conner and less than a mile east of the Swinomish Channel, a tidally influenced surface-water body that connects Skagit Bay to the south to Padilla Bay to the north (fig. 3). Water-level fluctuations in the well correspond closely to predicted tidal extremes obtained from the National Oceanic and Atmospheric Administration tide gage near La Conner (fig. 13). Through a trial-and-error correlation analysis, it was noted that water levels in the well showed similar relative high and low points, with a delay of approximately 28 hours after the tidal high or low, and an attenuation factor of about 3 percent. The peak-to-peak fluctuation in the well was about 0.3 ft compared to the tidal fluctuation of about 10 ft at the La Conner gage.