Spatial and Temporal Variability of Streambed Hydraulic Conductivity in West Bear Creek, NC
No Thumbnail Available
Files
Date
2007-04-11
Authors
Journal Title
Series/Report No.
Journal ISSN
Volume Title
Publisher
Abstract
Spatial and temporal variability of vertical hydraulic conductivity (KV) in a sandy streambed were evaluated from 487 field permeameter KV measurements taken during a 1-year study from December 2005 to December 2006. Bimonthly KV measurements were made at 46 locations (38 in December 2005) in a 262.5 m reach (the "large reach") of West Bear Creek in eastern North Carolina. More closely-spaced measurements were also made in two smaller (62.5 m) reaches within the large reach (Small Reach 1, measured in July 2006, and Small Reach 2, measured in August 2006). Vertical conductivity was calculated from field permeameter tests with data analysis based on Hvorslev (1951, case E, page 44).
Results show significant spatial (transverse and longitudinal) and temporal variability in KV. Calculated arithmetic mean KV values for the large reach ranged from 3.85 to 21.33 m⁄day and the mean of the 7 arithmetic mean values (one for each bimonthly run) averaged 15.44 m⁄day. Overall, the range of streambed KV was 0.01 to 66.21 m⁄day. Variance in lnKV (s2lnKv) ranged from 1.88 to 4.18 for the large reach measurement runs and s2lnKv values measured from Small Reach 1 and Small Reach 2 were 1.88 and 6.04, respectively. Smaller s2lnKv values were calculated in the center of the stream channel, compared to values of s2lnKv for measurements taken from the left or right sides of the stream channel. The opposite is true for average KV values, where the highest averages come from the center measurement points. While measurement spacing in the small reach runs was about 4-5 times closer than in the large reach runs, results from the two types of reaches had similar results with respect to range of KV, s2lnKv, and distribution of KV in the stream channel. Correlation lengths calculated from exponential model curves fit to the experimental lnKV semi-variograms were 1.4 and 8.2 for Small Reach 1 and Small Reach 2, respectively. Correlation lengths from the small reaches bracket those calculated from the large reaches. Temporal changes in K are probably due to deposition and erosion of stream sediment, and possibly time-varying behavior of biofilms and/or gas bubbles in sediment, but not to changes in temperature.
Contour maps of lnKV were created using three different interpolation techniques. Two of the three interpolation methods (radial basis function and kriging with a linear semi-variogram model) produced similar and fairly realistic looking maps. Large reach run maps showed a decrease in lnKV from the right bank to the left bank in some areas. However, overall the data show higher lnKV in the center of the channel. Small reach contour maps show higher lnKV in the center.
Average uncertainty in KV is equal to about 16.2%. This was calculated by adding, in root mean square fashion, the uncertainty due to lack of knowledge of the transformation ratio "m" ( 5%) to each uncertainty value arising from propagation of uncertainty in the measurands, and then averaging these 430 uncertainty values. Percent uncertainty in KV is larger for smaller KV values since the slope of the linear regression needed to calculate KV is less accurately constrained from the field measurements of head with time when KV is low.
Description
Keywords
field permeameter tests, spatial and temporal variations in hydraulic cond, streambed hydraulic conductivity
Citation
Degree
MS
Discipline
Marine, Earth and Atmospheric Sciences
