1. Deeba,Storrs, Buckley 1
Analyzing how reintroducing hyporheic flow in an urban stream effects conductivity and
chloride concentrations pre and post road salting
Emily Deeba, Erik Storrs, Camille Buckley
Introduction:
Urbanization negatively impacts surface water quality and reduces stream connection to
groundwater (Paul and Meyer, 2001). Increased urbanization alters land use and stream
characteristics creating highly variable runoff. River controls are commonly used in urban
settings as a way to mitigate runoff and to protect property. Hydrologic alteration occurs due to
these conventional river controls. Increased urbanization alters land use and stream
characteristics, creating highly variable runoff. River controls are commonly used in urban
settings as a way to mitigate this runoff, often with the consequence of hydrologic alteration.
One such river control is the straightening and lining of a river channel with an impervious
surface, often concrete. This lining disrupts the hyporheic zone, the region of streambed
sediments in which stream water and local groundwater mix (Hasenmueller, 2016). Groundwater
traditionally feeds gaining streams and establishes consistent flow and relatively stable
geochemistry (Ryan 2014). By encasing a stream with cement, the groundwater supply is
inhibited. The behavior of a channelized, impervious stream, thus, would exhibit characteristics
to bedrock channels as there would be insignificant hyporheic flow (Hasenmueller et al 2016).
Groundwater is an important buffering component of many streams. Increased urban land use
increases stream water conductivity; specifically, chloride was shown to increase logarithmically
to increasing impervious surface of contributing area (Kaushal et al. 2005). Channelized urban
streams are dominantly surface-runoff fed, as common river controls alter baseflow. Studies
indicate that precipitation fed streams gain salts from urban surfaces (Apel and Hudak 2001).

2. Deeba,Storrs, Buckley 2
This has implications for the geochemistry of urban streams, as runoff ratios are high due to
increased impervious surfaces in urban settings. Elevated sodium and chloride levels are toxic to
aquatic plants and fish. Additionally, salt would not be transformed by stream biota; it will only
be flushed and diluted over time ( Cooper et al. 2013).
Groundwater is hypothesized as a long-term reservoir for accumulating road salts and
could explain the persistence of salts in urban streams during summer months (Cooper et al.
2013). We study if, and how, urban stream conductivity, specifically chloride, changes as water
flows from a channelized, groundwater disconnected reach to a ground water connected reach.
Hasenmueller et al found larger, significant geochemical variation in lined portions of River des
Peres and less significant variation downstream in an unlined portion (2016).We questioned if
groundwater would serve as a buffer and decrease stream variability in conductivity and
chloride, or if the groundwater serves as a reservoir that contributes to increased variability in
conductivity and chloride. We also assessed how stream-groundwater connectivity affects
conductivity and Chloride by evaluating spatial and temporal changes in water chemistry in an
urban stream in St. Louis, Missouri pre and post-road salting conditions.
Study site: River des Peres is a first-order urban stream that is highly degraded and
extends about 30 km before discharging into the Mississippi River. The watershed has >90%
urban land coverage (Hasenmueller, 2016). Multiple combined sewer overflows (CSO’s) are
present upstream of the sample site. The watershed has abundant monthly rainfall (i.e., 97 cm/yr;
NOAA, 2015). The dominant geologic character of the catchment is dominantly Mississippian
limestone (which has buffering implications) (Hasenmueller et al. 2016). Spatial and temporal
changes were monitored in the Upper River des Peres, which is surrounded suburban areas and
roads. The section of the river is highly channelized and degraded. Samples were taken in a

3. Deeba,Storrs, Buckley 3
channelized, cement-encased portion of the stream and along the stream into Ruth Park, a more
natural, and unlined streambed.
Methods:
Measurements were taken to monitor impact of hyporheic flow on conductivity and
chloride. Samples were taken along a .75 kilometer stretch of River des Peres through Ruth park,
taking measurements every 2-3 meters from the channelized, GW disconnected portion of River
des Peres into Ruth Park; measurement increments were increased to about every 10-20 meters
further downstream in the GW-fed portion of the stream. A handheld ProPlus was used to take
water quality point measurements; temperature, dissolved oxygen, conductivity, pH and chloride
were all noted in the spatial survey of the stream. In some cases, conductivity was used as an
indicator of the salts in solutions, as salinity is directly correlated to total dissolved solids
(Copper et al. 2013). Our study was done to detect groundwater inputs and how contributions
changed geochemistry of River des Peres as it transitions from its disconnected portion into the
more natural setting with hyporheic flow (Haria, 2012). These findings will then be compared to
existing literature to understand how groundwater disconnected streams react to road salting as
hyporheic flow is reestablished.
Sampling began on December 20th, 2015, (before winter road salting) and continued
until March 26, 2016 (post winter road salting). Data collected from December 20th served as a
baseline, as road salt had not been applied since the previous winter. Two surveys of River des
Peres were done in February, following multiple road salting events.
Results:

4. Deeba,Storrs, Buckley 4
As seen in Figure 1.A. and 1.B., there are large spikes in both conductivity and chloride
for samples 13 and 14 for dates 2/20/2016 and 3/25/2016 respectively. There is also a large
spike in conductivity on 12/20/2015 at sample 15. These spikes represent samples collected at a
stagnant pool caused by a tree dam in River des Peres. Since the readings from this sample site
don’t reflect the overall data as a whole, the data collected from that site has been omitted from
our data. On 2/20/2016, the chloride sensor on the meter used to gather the data was maxed out,
so it was excluded from the data set as well.
Figures 2.A and 2.B show the conductivity and chloride data with the peaks from the tree dam
site removed.

5. Deeba,Storrs, Buckley 5
Figure 3.A shows the conductivity for all the days normalized to zero. For 12/20/2015 (the only
pre-road salting day), we see what could be a drop in conductivity as the water flows
downstream. The other days don’t appear to show a significant drop in conductivity as water
flows downstream.
Figure 4.B shows the conductivity data for 12/20/2016 for samples 20-26. Samples 20-26 were
chosen because sample 20 has a lower conductivity than sample 1, the highest point upstream
where measurements were taken, and is where the conductivity levels begin to decline in earnest.
A linear model was then fitted to the data, the regression line for this model is shown in red in
Figure 4.B. When a t-test was applied to this trend, we found the p-value to be .00123, so we
reject the null hypothesis that there was no change in conductivity and accept the alternate

6. Deeba,Storrs, Buckley 6
hypothesis that there was a decrease in conductivity as water in the River des Peres flows
downstream. These results, in part, mimic previous work that found a 200 uS/cm decrease in
conductivity as stream water flowed from a channelized a groundwater fed portion of the stream
(Hasenmueller, 2016).
Figures 5.A and 5.B show differences in conductivity and chloride for the days that
samples were taken. Pairwise comparisons using t tests with pooled standard deviation and
bonferroni correction were taken between each of the days for both conductivity and Cl-. There
was a significant difference in conductivity between 12/20/2015 and 2/20/2016 (p-value of 2e-
16) and between 12/20/2015 (p-value of 6.8e-05). There was also a significant difference in
chloride levels between 12/20/2015 and 2/28/2016 (p-value of 7.7e-13) and between 12/20/2015
and 3/25/2016 (p-value of 1.0e-15). These findings echo those of Ledford et al, who found that
in winter stream chloride levels were 2-3 times higher during winter than during non-winter
months.
The figures also show an increase in variation from pre-salting conditions (12/20/2015) to
post-salting conditions(2/20/2016, 2/28/2016, and 3/25/2016). This is also consistent with the
findings of Ledford et al (2014).

7. Deeba,Storrs, Buckley 7
Discussion:
Reintroduced hyporheic interactions in the River des Peres displayed significant trends
regarding conductivity and chloride concentrations. As the distance in the unlined portion
increased, a gradual decrease in conductivity and chloride was expected. Hyporheic flow has
been shown to lower geochemical variation (Hasenmueller et al 2016). Spikes in conductivity
and chloride were expected to be seen shortly after the December 28th snowfall, which were
correlated to characteristic first flushes (Hasenmueller 2016). Spikes were expected to occur in
pools as stagnant water accumulates constituents and from decreased volume due to evaporation.
We questioned if the hyporheic zone acted as reservoir or buffer for conductivity and chloride.
Our four sample points showed conductivity spikes between the sampling dates. Low
chloride and conductivity are shown before winter storms that took place December 28, 2016.
The snow washed road salt into the RDP leading to a spike in conductivity on February 20, 2016.
Following February 20, 2016, a gradual decrease was shown in the last sample dates, which can
be contributed to small amounts of precipitation that occurred around those dates. Figure 6
shows gauge height, correlating to discharge, of the RDP at University City, about 2.5 km
downstream from where we sampled. Kelly’s peak and decreasing chloride trends in Illinois’
rivers were congruent with our findings (2013) (Figure 5.B). These similarities demonstrate
expected concentrations trends in the stream, thus allowing us to assume that road salting was

8. Deeba,Storrs, Buckley 8
not responsible for anomalies in our data.
Figure 6: Gauge height data of the RDP during winter 2015-2016 (USGS 07010022, 2016)
The data collected from December 20, 2015, showed similar conductivity in the lined and
unlined portion of the river. This trend remained constant until approximately 186 meters from
the first data sample. At this point chloride and conductivity decreased, alluding to dilution in the
river through either precipitation or groundwater interaction. As there was no significant runoff
or contributing precipitation during the span of each individual sample date, the decreased can be
correlated to an increase
groundwater interaction.
In this case, the
groundwater filtrates
and purifies surface
water by diluting
chloride concentrations and conductivity (Zhou, et al. 2014). Our pre-road
Figure 7: Conductivity data taken in November and March on the RDP with increasing
distance (Hasenmueller and Robinson, 2016)

9. Deeba,Storrs, Buckley 9
salting conductivity data mimics Hasenmueller and Robinson's' findings on November, 2015
(Figure 7) and serves as a pre-road salting baseline for this study. We used the December 20,
2016, data sample as our reference/null point to determine trends for our other three points.
Data collected from February 20, 2016 and February 28, 2016 showed significant
increases in conductivity. The first and second sampling dates were determined to be significant
with regards to conductivity. The final sampling date data was not significant with respect to the
baseline, possibly alluding to a return to baseline conductivity.
Chloride measurements showed significance on February 28, 201616 and March 25,
2016. Chloride concentrations were found to be significantly different with respect to the
baseline measurements for both days. Chloride concentrations for March 25, 2016 provided an
anomaly in our findings; although chloride concentrations were significant, conductivity were
not. This discrepancy may have been caused by inaccuracies in the ProPlus Cl sensor or possible
contaminants due to CSO’s.
Conductivity trends observed for our null data point was somewhat seen in samples
February 28, 2016 and March 25, 2016. In figure 3.A you can see that both samples start with
dilute waters inside the end of the channelized portion, which quickly jump to higher
concentrations within the first 2-4 sample in the unlined portion. Once hyporheic flow is
reestablished, conductivity rises for a period, then begins to decrease. This decrease begins at
approximately the same point as where the baseline conductivity began to decrease. This might
allude to dilution through interaction of the hyporheic zone and/or entrapment or sedimentation
of constituents dammed in stagnant water. The February 20, 2016 conductivity data did not
resemble any trends expected seen within the other sets of data. It contained many spikes all
throughout the sample area and did not contain the drastic jump from diluted water to

10. Deeba,Storrs, Buckley 10
concentrated at the beginning. The river runs parallel to road where salt could directly flush into
the stream, explaining “random” spikes. The trends for chloride show the similarities to
conductivity.
Our results demonstrate groundwater and surface water interaction along the reach of an
urban stream. During the first portion of the River des Peres, the hyporheic zone seems to acts as
a reservoir for chloride, however, further down the stream groundwater begins to dilute the
surface water. Once transitioned into the unlined portion, chloride concentrations decrease as
groundwater inputs increase with length.
Conclusion:
Hyporheic zone interactions can cause variations in geochemistry based on location,
climate, and anthropogenic events such as road salting. Using a handheld ProPlus we recorded
and analyzed conductivity and chloride from a channelized, groundwater disconnected portion of
the River Des Peres to a more natural, groundwater connected portion. Significant differences
were found in chloride and conductivity pre and post road salt application. Unexpected increases
in concentration were found along the unlined portion. Increased sampling distance, however,
showed decrease in conductivity and chloride. Our findings suggests decreasing trends in
chloride and conductivity with increased hyporheic flow pre road salting and highly variable
chloride and conductivity concentrations post-road salting that decrease with time.
References
Apel, PL. Hudak, PF. Automated sampling of stormwater runoff in an urban watershed,
North-Central Texas. J Environmental Sci Heal 2001; A36(6): 897-907.
Cooper, C.A. Mayer, P.M. Faulkner, B.R. Effects of road salts on groundwater and surface water

11. Deeba,Storrs, Buckley 11
dynamics of sodium and chloride in an urban restored stream. Biogeochemistry
2014;121:149-166.
Hasenmueller, E. A., Robinson, H. K., 2016. Hyporheic Zone Flow Disruption from Channel
Linings: Implications for the Hydrology and Geochemistry of an Urban Stream, St.
Louis, Missouri, USA. Journal of Earth Science. 27, 1-13.
Kaushal SS, Groffman PM, Likens GE, Belt KT, Stack WP, Kelly VR, et al. Increased
Salinization of freshwater in the northeastern United States. Proc Natl Acad Sci U S A
2005;102(38):1317–20.
Kelly, W.T., Panno, S.V., Hackley, K., 2012. The Sources, Distribution, and Trends of Chloride
in the Waters of Illinois. Illinois State Water Survey.
<http://www.isws.illinois.edu/pubdoc/b/iswsb-74.pdf> 13 Apr 2016
Haria, A.H. Shand, P. Soulsby, C. Noorduijn, S. Spatial delineation of groundwater–surface
water interactions through intensive in-stream profiling. Hydrological Processes. 27,
628–634 (2013). Published online 5 November 2012 in Wiley Online Library
(wileyonlinelibrary.com). DOI: 10.1002/hyp.9551
Paul MJ, Meyer JL. Streams in the urban landscape. Annu Rev Ecol Syst 2001;32:333–65.
USGS 07010022 River Des Peres near University City, MO. USGS. <
http://waterdata.usgs.gov/nwis/uv?site_no=07010022> 10 Apr 2016
Ryan, P. C., 2014. Aqueous Systems - Controls on Water Chemistry. Environmental and Low
Temperature Geochemistry. 1st ed. 108, 11.
Zhou, S., Yuan, X., Pend, S., Yue, J., Wang, X., Liu, H., Williams, D.D., 2014.
Groundwater-surface water interactions in the hyporheic zone under climate change
scenarios. Environmental Science and Pollution Research. 24, 13943-13955