Abstract
To study the hyporheic exchange driven by a single peak flood-induced
water level fluctuation (i.e. flood wave), a method combining numerical
simulation with theoretical derivation was proposed based on the Inbuk
Stream, Korea, where flooding occurs frequently. The hyporheic exchanges
induced by different flood waves were investigated by varying amplitude
(A ), duration (T ), wave type parameter (r ), and
rising duration (t p), which were adopted from the
real-time stream stage fluctuations. Additionally, the idea of constant
upstream flood volume (CUFV) condition for flood waves was put forward,
and the effects of “Botan” (T /A ) and peak number
(N ) on hyporheic exchange were studied. The results showed that
the hyporheic exchange flux (q ) was controlled by the water levelh (sine-type) and its change rate v (cosine-type), and was
proportional to the polynomial of themq \(\mathrm{\propto}\)(ω \(\bullet\)h +v ),
where ω is the angular frequency of the flood wave. Based on this
mechanism, the influence principles on hyporheic exchanges of the
typical flood wave parameters (A , T , r andt p) as well as T /A and Nunder CUFV condition were clarified. The main characteristic variables
of hyporheic exchange, which were maximum aquifer storage and residence
time, were positively correlated. They also had positive relations to
the integral of the flood wave over time, which increased when the wave
became higher, wider, rounder and less skewed. However, when CUFV
condition was imposed, the residence time was positively correlated withT /A , whereas the maximum aquifer storage was negatively
correlated with T /A . With the increase in N , water
exchanged more frequently and some water returned to the stream early,
leading to the slight decrease in maximum aquifer storage and residence
time. These findings enriched the theory of hyporheic exchange driven by
surface water fluctuation and be of great significance to enhance
pollutant degradation in the hyporheic zone downstream of reservoirs.
Keywords: hyporheic exchange; flood wave; numerical simulation;
theoretical derivation
Introduction
The hyporheic zone is a critical area where mixing of shallow
groundwater and surface water occurs in the fluvial system (Hester &
Gooseff, 2010; White, 1993). The important role it plays in maintaining
the ecological health of rivers has been repeatedly demonstrated and
gradually accepted by the academic community (Findlay, 1995; Marzadri et
al., 2013; Storey et al., 1999). There are many factors (Lewandowski et
al., 2019) that promote hyporheic exchange (HE). Flood events are
important driving forces due to mountain torrents, snow melt, storm
events and reservoir operations which commonly occur in natural rivers.
Flood-induced water level fluctuation (i.e. flood wave) and associated
lateral propagation can change the hydrological situation in the
hyporheic zone (Curry et al., 1994; Friesz, 1996), and subsequently
promote the transport and transformation of pollutants (Harvey et al.,
2013; Kolbjørn Jensen et al., 2017; Roley et al., 2012; Trauth et al.,
2018), the reproduction of benthic organisms (Boulton et al., 1998;
Holomuzki & Biggs, 2000; Olsen & Townsend, 2005) and the growth of
vegetation in the riparian zone (Harner & Stanford, 2003; Mouw et al.,
2009). However, large-scale human activities (e.g. damming, river
channelization, hard-type bank protection, artificial flow regulation)
(Hancock, 2002) can weaken the flood pulse of natural rivers to varying
degrees, and may have a substantial impact on the ecological function of
the hyporheic zone (Arias et al., 2013; Keizer et al., 2014; Nilsson &
Berggren, 2000). Hence, a better understanding of HE driven by flood
wave is key for the healthy river development and restoration of damaged
or degraded rivers.
Piezometers were appropriately arranged in a riverbed or riparian zone
and some chemical indexes were collected (e.g. Hanrahan, 2008; Jeon et
al., 2015) in previous field investigations. The studies showed that the
river-aquifer hydraulic gradient responds rapidly to the transient water
level fluctuations (Arntzen et al., 2006), and there was a certain
quantitative relationship between them (Fritz & Arntzen, 2007). The HE
flux (q ) between surface water and groundwater was found to be
linearly proportional to the river water level when there is a
significant hydrological connection (Cardenas, 2010). However, this is
inconsistent with the nonlinear relationship described by Chen et al.
(2013). Crosbie et al. (2014) proposed a two square function
relationship between q and the river water level. Based on field
tests, some studies have quantitatively analyzed the extent of water
infiltrating into the aquifer, the corresponding residence time and the
spatiotemporal distribution. Their results showed that the magnitude of
lateral exchange decreases as the distance from the river increases
(e.g. Graham et al., 2015; Liu et al., 2018; Sawyer et al., 2009);
however, in the streambed, both the vertical infiltration distance and
residence time decrease with proximity to the middle of the river
(Gerecht et al., 2011).
Previous numerical simulations indicated that river water fluctuation
increase the bank storage and infiltration extent (Chen & Chen, 2003;
Doble et al., 2012; Gu et al., 2012; Maier & Howard, 2011; Welch et
al., 2013; Welch et al., 2015), and enabled the retention of water
particles in the aquifer for a long period (Diem et al., 2014; McCallum
& Shanafield, 2016). Siergieiev et al. (2015) analyzed the bank storage
and residence time under different flood wave scenarios by constructing
a 2-Dimensinal variably saturated model of the bank cross section, and
indicated that bank storage increases with wave amplitude and duration,
while the ratio of return time to infiltration time is positively
related to amplitude and negatively related to duration. Trauth and
Fleckenstein (2017) set up a reactive transport groundwater model with a
wide range of flood scenarios. Their results showed that the longer the
flood duration and the higher the peak discharge, the greater the
magnitude of HE and the higher the reactive efficiency for aerobic
respiration as well as denitrification in the hyporheic zone below a
natural in-stream gravel bar.
The above studies have highlighted the quantitative relationship betweenq and the water level. However, few studies have quantified the
effect on q of the rate of
water level change. In addition,
previous numerical simulations focus mainly on the effects of flood wave
amplitude (A ) and duration (T ) on HE (e.g. Chen & Chen,
2003). However, some important wave parameters such as the wave type
(r ) and rising duration (t p), which
characterize the roundness and peak position of a wave, have been
ignored. Furthermore, there have been few studies about the influences
of flood waves on HE under constant upstream flood volume (CUFV)
condition, although it is of great value to the ecological operation of
a reservoir.
In this study, HEs driven by flood waves was systematically studied. We
aim to: (1) reveal the mechanism of superimposed influences from
flood-induced water level and its change rate on HE; (2) further
investigate the influences of typical flood wave parameters (A ,T , r and t p) on HE; (3) gain
insight to the influences of “Botan” (T /A ) and peak
number (N ) on HE under CUFV condition.
Methods