USGS Home
Contact USGS
Search USGS

U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2007–5258

Back to Table of Contents

Trend Analysis

Ground-Water Level Trends

Temporal trends in ground-water levels for three wells in the Wood River Valley with more than 50 years of measurements varying in frequency from weekly to biannually were evaluated using techniques described in section, “Ground-Water Level Trend Analysis.” All three wells, 432042114163801, 432143114114301, and 432657114144801, showed statistically significant (the level of significance, or p-value, was less than 0.005) declining trends in ground-water levels. Well 432042114163801 is completed in the confined aquifer near the intersection of U.S. Highway 20 and State Highway 75. The remaining two wells are completed in the upper unconfined aquifer: well 432143114114301 near Gannett and well 432657114144801 south of Bellevue.

To examine the response of water levels to precipitation, normalized plots of mean annual depth to water for the three wells and total annual precipitation at the Ketchum Ranger Station were constructed as described in section, “Ground-Water Level Trend Analysis” (fig. 5). Although there is broad agreement of trends and many individual years between the water level and precipitation plots, a large number of years do not match or show water-level declines in response to periods of increased precipitation. A simple linear regression between water levels and precipitation confirms a poor correlation between the two: all R² (coefficient of determination) values are less than 0.5. This lack of correlation is further confirmed in that water levels in these three wells show statistically significant downward trends, while, as previously mentioned, total annual precipitation at the Ketchum Ranger Station shows a statistically weak (p-value of 0.258) increasing trend.

Water levels in the three wells unquestionably respond to recharge from precipitation. However, other factors such as withdrawals, infiltration from surface-water irrigation (including canal leakage), and land-use changes probably have an effect. Water levels in these wells also have fluctuated over their period of record. Table 6 shows the extremes and annual variation of ground-water levels in these three wells.

The statistically significant downward trend in ground-water levels helps to explain the statistically significant downward trends in low-flow values for the Big Wood River near Bellevue and the Silver Creek at Sportsman Access gaging stations. Because both streams are fed largely by ground water in these reaches, any lowering of ground-water levels will reduce streamflow.

Streamflow Trends

During analysis of streamflow at the Big Wood River at Hailey gaging station, it was observed that flows prior to 1945 had periodically decreased to zero due to upstream diversions at the Justus Ditch for hydroelectric production and seasonal irrigation. An unknown quantity of water was then returned to the river downstream of the gaging station. Use of the ditch was discontinued sometime in the late 1970s to early 1980s. The combined records of the Big Wood Slough at Hailey and the Big Wood River at Hailey were used during this period for trend calculations.

No statistically significant trends were found in the mean annual streamflow for the Big Wood River at Hailey gaging stations (table 5). A statistically significant trend exists for the mean monthly flow for March, for which there was a 25-percent increase in mean monthly base flow from 1951-2005 (p-value of 0.005). A weak statistical increasing trend also exists for January (p-value of 0.123). The March increase in streamflow may be a function of earlier snowpack runoff, seen in basins throughout the Western United States (Stewart and others, 2005).

Analysis of the annual 7- and 30-day low-flow calculations at the Big Wood River at Hailey gaging stations indicate an increase in base flow (water years 1915–2005). Weak statistical trends for the 7-day (p-value of 0.117) and 30-day (p-value of 0.097) low-flow calculations suggest an increase of 20 ft³/s since 1915 (fig. 6A). As mentioned in the section, “Streamflow Trend Analysis,” the confidence interval for surface-water analysis in this report was restricted to 90 percent; while the trends for 7-day and 30-day low-flow calculations have confidence intervals of 88 and 90 percent, respectively, they may suggest a developing trend. If these trends exist, a possible explanation may be a statistically weak increasing trend (p-value of 0.258) of total annual precipitation at the Ketchum Ranger Station.

The Big Wood River near Bellevue gaging station (13141000) did not have a statistically significant annual trend in discharge or a BFI trend (table 6). However, mean monthly discharge values show statistically significant decreasing trends for December, January, and February (p-values of 0.014, 0.006, and 0.095, respectively, fig. 7). Statistically significant decreasing trends also exist for both the 7-day and 30-day low-flow analyses (p-values of 0.084 and 0.076, respectively), showing a mean decrease of about 15 ft³/s since the 1940s (fig. 6B).

To examine whether or not ground-water inflow has changed, data from the Silver Creek at Sportsman Access near Picabo gaging station (13150430) was examined for trends. Because Silver Creek and its tributaries are fed primarily by ground water discharging through seeps and springs, and because irrigation diversions and returns to and from Silver Creek are minimal, seasonal fluctuations in streamflow are directly related to water levels in the Wood River Valley aquifer system (Moreland, 1977). Therefore, streamflow in Silver Creek can be used as a proxy for trends in ground-water discharge to the Big Wood River on the Bellevue fan. Silver Creek at Sportsman Access has a statistically significant annual discharge decreasing trend (p-value of 0.018) for the period of 1975—2005 and mean monthly discharge values at this gaging station also show statistically significant decreasing trends for July through February and April (p-values all less than 0.052), table 5. Statistically significant decreasing trends also exist for both the 7-day and 30-day low-flow analyses (p-values of 0.074 and 0.101, respectively). The Silver Creek at Sportsman Access gage data did not have a statistically significant trend for the BFI analysis.

The decreasing trends in the Big Wood River near Bellevue since the 1940s, and in Silver Creek since 1975, indicate declining ground-water levels. There have been wet and dry cycles during this time. However, population has increased since the 1960s, and therefore the declines in ground-water levels could be due to increased consumptive use of water in the valley from increased population. Increased consumptive use may result in lowered ground-water levels, which reduces spring flows to the Big Wood River and Silver Creek. These reduced flows are most evident during the winter months for the Big Wood River near Bellevue gage since flow is primarily derived from ground-water inputs to the river during this time. Because Silver Creek is spring fed year round the effect of declining ground-water levels is evident for most of the year. More information is needed on water-use patterns to test this hypothesis.

Back to Table of Contents