Selenium and sediment loads in storm runoff in Panoche Creek, California, February 1998
Five to nine samples were collected per storm throughout the hydrograph of four storms in February 1998 from Panoche Creek at Interstate 5, California. The rainfall total of 10.40 inches for the month was greater than any other month during 1957 to 2000, and peak streamflows on February 3 and 7 exceeded the previous peak streamflow of record. Concentrations of suspended sediment, which were highly variable during the storms, ranged from 32,900 to 251,000 mg/L (milligram per liter) with a median of 126,000 mg/L. Dissolved selenium concentrations in the storm samples ranged from 16 to 60 μg/L (microgram per liter), with a median of 30 μg/L. These concentrations were considerably higher during the first storm than during subsequent storms. Total selenium concentrations in the storm samples ranged from 57 to 320 μg/L, with a median of 115 μg/L. Total selenium in four replicate and four rerun samples was highly variable (relative percent differences ranged from 0 to 57 percent), probably due to the extremely high concentrations of suspended sediment in the samples and possible interferences from other compounds. The calculated concentration of selenium attached to suspended sediment was less variable than suspended sediment or total selenium concentrations during storm runoff; concentrations ranged from 0.50 to 2.1 μg/g (microgram per gram) with a median of 0.89 μg/g. The logarithms of suspended sediment and total selenium concentrations were closely correlated to the logarithm of streamflow (R2 = 0.82, R2 = 0.71, respectively for all storm data). These relations for the first storm were significantly different from the later three storms. Because of these correlations, the logarithm of suspended sediment loading rate and the logarithm of total selenium loading rate were closely correlated to the logarithm of streamflow (R2 = 0.982 for both for all storm data). Loads of suspended sediment and total selenium were calculated for each of the four storms in three ways, including simple linear regression with streamflow for all storms, simple linear regressions for the first storm and the combined later three storms, and integration under the instantaneous load curves. The resulting suspended sediment loads for all four storms ranged from 1,793,000 to 2,555,000 tons; total selenium ranged from 4,909 to 5,830 lb (pound). Dissolved selenium concentrations correlated significantly with both the logarithm of streamflow and specific conductance. Simple linear regression with the logarithm of streamflow had an R2 of 0.45, and a multiple linear regression with the logarithm of streamflow and specific conductance had an adjusted R2 of 0.48 for all storm data. The relation between streamflow and specific conductance for the first storm and for the later three storms were significantly different. As for suspended sediment and total selenium, dissolved selenium loads were calculated three ways. The resulting loads for all four storms ranged from 773 to 1,007 lb. No significant storm occurred during the remainder of the study period (water years 1998 to 2000), and thus, no additional storm sampling took place. Assuming that future sediment and selenium transport is similar to that of 1998, a reasonable estimate of loading rates can be calculated from the streamflow record at the Interstate 5 gage. Additional storm sampling would improve the estimates and possibly allow for separate equations for the rising and falling limbs of storm hydrographs.