FIELD AND MODELING STUDY OF THE EFFECTS OF STREAM DEPTH AND GROUND WATER DISCHARGE ON HYDROGEOPHYSICAL

Abstract

Valley Creek, an urbanized stream in Southeastern Pennsylvania, has undergone changes typical of streams in urbanized areas, such as bank erosion, channel redirection, and habitat disruption. One area of disruption that has been little studied is the hyporheic zone, the top layer of the streambed where stream water exchanges with subsurface water and chemical transformations occur. The hyporheic zone of an 18 m reach of Valley Creek in Ecology Park was characterized using a tracer test coupled with a hydrogeophysical survey. Nested wells screened at depths of 20, 35, 50, and 65 cm were placed at four locations along the center of the stream to monitor the passage of the salt tracer through the hyporheic zone. Results from well sampling were compared with time-lapse Electrical Resistivity Tomography (ERT) monitoring of the stream tracer. The streambed was also characterized using temperature probes to calculate the stream water-groundwater flux and freeze core samples to characterize heterogeneities in streambed sediment. Models were created using MODFLOW, MATLAB, and EARTH IMAGER 2-D to understand differences between Ecology Park and Crabby Creek, a tributary within the Valley Creek watershed, where similar studies were performed in 2009 and 2010. Hyporheic exchange and ERT applicability differed between the two study sites. At Ecology Park, tracer was detected only in the 20 cm wells at nests 2 and 4 during the injection period. Noise in the falling limbs of the tracer test breakthrough curves made it difficult to determine whether tracer lingered in the hyporheic zone using well data. ERT surveys were unable to detect tracer lingering after the injection period. At Crabby Creek, tracer was present in all shallow wells, and lingering tracer was detected in the hyporheic zone using ERT during the post-injection period. ERT surveys at Ecology Park were less effective than at Crabby Creek for two reasons: the presence of groundwater discharge (which inhibited hyporheic exchange) and increased stream water depth at Ecology Park. Temperature modeling of heat flux data revealed groundwater discharge at three locations. MODFLOW models predicted that this discharge would diminish the length and residence time of subsurface flow paths. Groundwater discharge likely increased along the contact between the hydraulically conductive Elbrook Formation and the less conductive Ledger Formation. Models created with MATLAB and Earth-Imager 2-D showed ERT sensitivity to tracer in the hyporheic zone depended on stream thickness. With increased water depth, more current propagated through the stream, which reduced sensitivity to changes in the hyporheic zone. A sensitivity analysis showed that the resistivity change in the hyporheic zone at Ecology Park (average water depth 0.36 m) would have to exceed 30% to be detectable, which was greater than the induced change during the tracer test. Deeper water also amplified the confounding effect of changes in the background conductivity of the stream water, though time-lapse ERT detected no lingering tracer even after correcting for this drift. Studies performed at Crabby Creek were able to map lingering tracer in the hyporheic zone because the site had a thin water layer (0.1 m), a large percentage increase of conductivity during the tracer test, and no groundwater discharge. Conversely, at Ecology Park groundwater discharge inhibited hyporheic exchange, and imaging sensitivity was reduced by the thicker water layer, demonstrating the limitations of ERT for hyporheic zone characterization. The modified inversion routines used here demonstrated that, with accurate stream conductivity and depth measurements, ERT can be used in some streams as a method for hyporheic characterization by incorporating site-specific conditions.